10.1007/978-3-319-58292-4 Mutualism and Insect Conservation PDF

Tim R.
New
Mutualisms
and Insect
Conservation
Mutualisms and Insect Conservation
Tim R. New
Mutualisms and Insect

Conservation
Tim R. New
Department of Ecology, Environment
& Evolution
La Trobe University
Melbourne, VIC, Australia
ISBN 978-3-319-58291-7 ISBN 978-3-319-58292-4 (eBook)

DOI 10.1007/978-3-319-58292-4
Library of Congress Control Number: 2017940274
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors
or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims
in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The ecology, origin and evolution of ‘mutualisms’, in their broadest sense associa-
tions that incorporate cooperative interactions amongst different species, with the
general implication that many such associations are regular and common, and many
of them obligatory through being both specific and specialised, is perhaps one of the
most critical themes in conservation. Ecological interactions, many of them essen-
tially viewed as ‘antagonistic’, whereby one species exploits another that suffers
accordingly, are of universal importance. However, the importance of mutualisms is
founded in the participants each benefiting from the association, so that the interact-
ing species in a functional interaction each, in some way, attain increased ‘fitness’
and, conversely, are disadvantaged should that association be disrupted.
A mutualism has been described as an interaction between species in which the
inclusive fitness of each party is increased by the action of its partner. In functional
terms, mutualisms are thus entities incorporating and sustaining the wellbeing and
functional roles of two or more species, which are often interdependent partnerships
and thus fundamental and necessary units for practical conservation attention.
Conservation driven by the needs of one participant or partner species must include
needs of the other(s). Examples and contexts abound, but their intricacies are com-
monly difficult to elucidate, and the consequences of disturbance or breakdown
highly uncertain and unpredictable in detail.
Every insect species interacts with other taxa, and almost all the insect species
signalled as of individual conservation concern as ‘threatened’ are in some way
‘ecological specialists’. Most have a very restricted suite of food resources in local-
ised environments, for example, and many are susceptible to declines or losses
through lacking the flexibility to exploit other resources when those they critically
need are lost and critical ecological associations disappear as those environments
succumb to change. Many such species are intrinsically ‘rare’, with features such as
narrow natural distributions and small populations enhancing the vulnerability that
can flow from extreme specialisation.
This book is a broad survey of some of these associations, emphasising the needs
to understand mutualisms and the roles they can play in wider insect conservation.
Whilst much attention to insect conservation has traditionally focused on single
v
vi Preface
species, this account pursues the broader theme advocated strongly for island biota
by Kaiser-Bunbury et al. (2009) that the real conservation challenge is ‘maintaining
and restoring the integrity of interactions between species’, amongst which mutual-
isms have a key and central place as involving widespread cascade effects if they are
disturbed. Many such associations involve insects.
Whether trophic interactions (herbivore-plant, predator-prey, host-parasitoid or
parasite, or other) or more specifically functional foci (such as pollinator-plant) are
emphasised, the gradient of continuity from species focus to ecosystem focus in
conservation embraces a multiplicity of mutualisms of fundamental significance for
their future wellbeing. The increasingly advocated, and logistically expedient, view
that individual species conservation must progressively give way to wider perspec-
tives poses some danger that individual specialised mutualisms (even if they are
noticed initially) may become submerged or overlooked but also that knowledge of
their intricacies, and any wider generalisations that may flow, are omitted from con-
servation agendas in which the needs of ecologically intricate and vulnerable flag-
ship species are not detected or selected for investigation. As noted by Fisher (1998),
‘Conservation management directed to ensuring the survival of functional flagship
species interactions will preserve other codependent and coexisting species and
interactions’. Historically, most studies on mutualisms involving insects have
focused on ‘single animal species – single plant species’ interactions, but the more
recent trend towards community-level studies has led to stimulating efforts to model
interaction network patterns to understand how these may persist (Fortuna and
Bascombe 2006).
The diversity of mutualisms and their ecological and evolutionary interests and
significance are amply displayed in the recent book edited by Bronstein (2015),
which brings together and synthesises examples from numerous taxonomic groups
and environments into an unsurpassed modern text which develops the legacy of its
major forerunner (Boucher 1985). It has been a stimulus for me to complete this
much more limited work, planned initially some years ago. My aim has been to
display to non-specialists and conservation managers the variety of intricate asso-
ciations in which insects participate, through recapitulating and discussing the natu-
ral history of some classic themes in mutualistic relationships and their relevance in
practical insect conservation. That relevance is, in part, based on the statement that
‘certain communities harbour mutualists in an amazing diversity, and their activity
is essential to the functioning of ecosystems’ (Prinzing et al. 2007). Whilst intui-
tively obvious that ecological interactions are fundamental in communities, appre-
ciation of the variety and importance of mutualisms raises a variety of questions on
how the roles of mutualists are determined and sustained, and how disturbances
may be predicted to affect the associations.
The book is divided into three parts, and includes selected information available
to me up to late 2016. The first part (Chaps. 1 and 2) discusses the meaning, variety
and scope of mutualisms, together with some of the processes underlying their
integrity and dynamics; the second (Chaps. 3, 4, 5, 6, and 7) is an overview of the
ecology and evolution of many of the classic cases of mutualism in which insects
participate, to illustrate their ecological and taxonomic variety and how they have
Preface vii
contributed to the wider understanding, not least through appreciating the nuances
of specific details superimposed on what may appear initially to be easily character-
ised associations; and the last (Chaps. 8, 9, and 10) emphasises their importance in
and for wider insect conservation issues. Collectively, they help to provide the infor-
mation on these intricate relationships that may facilitate their continuity, and that
of their participating species, into an ever-changing world.
Melbourne, Australia Tim R. New
References
Boucher DH (ed) (1985) The biology of mutualism. Croom Helm, London/Sydney

Bronstein JL (ed) (2015) Mutualism. Oxford University Press, Oxford
Fisher BL (1998) Insect behavior and ecology in conservation: preserving functional species inter-
actions. Ann Entomol Soc Am 91:155–158
Fortuna MA, Bascompte J (2006) Habitat loss and the structure of plant-animal mutualistic net-
works. Ecol Lett 9:281–286
Kaiser-Bunbury CN, Traveset A, Hansen DM (2009) Conservation and restoration of plant-animal
mutualisms on oceanic islands. Perspect Plant Ecol Evol Syst 12:131–143. doi:10.1016/j.
ppees.2009.10.002
Prinzing A, Dauber J, Hammer EC, Hammouti N, Bohning-Gaese K (2007) Perturbed partners:
opposite responses of plant and animal mutualist guilds to inundation disturbances. Oikos
116:1299–1310
Acknowledgements
The following publishers and organisations have graciously granted permission to

use or reproduce material under their copyright control. I should emphasise that
most figures have been redrawn to facilitate standardisation of labelling, and some
have been simplified by omissions of some details; each is acknowledged individu-
ally in the accompanying legends. Every effort has been made to obtain permission
to use previously published material. The publishers would welcome advice on any
inadvertent omissions or corrections that should be included in any future editions
or imprints. Thanks are conveyed to: The Association for Tropical Biology and
Conservation; Cambridge University Press, Melbourne; Elsevier, Oxford; John
Wiley and Sons Inc.; Naturalis Biodiversity Centre, Leiden; The New York
Entomological Society Inc.; The Royal Society, London; University of Chicago
Press, Chicago; University of Texas Press, Austin; Wiley-Blackwell Publishing,
Oxford.
It is with considerable gratitude that I acknowledge the magnificent support of
Mariska van der Stigchel, Assistant Editor at Springer. Her continued patience,
good humour, advice and practical help in dealing with many queries I inflicted on
her are appreciated deeply. Zuzana Bernhart enthusiastically accepted my proposal
for this book, and her friendship and advice have also made working with Springer
a rewarding process. The careful production of the book by Ms. Shobha Karuppiah
(Production Editor) and Mrs. Lakshmikanthan Sudantra Devi (Project Manager) is
also appreciated greatly.
ix
Contents
Part I The Meaning and Dynamics of Mutualisms

1 The Scope and Meaning of Mutualisms................................................. 3
1.1 Introduction..................................................................................... 3
References.................................................................................................. 13
2 Mechanisms and Maintenance................................................................ 15
2.1 Coextinction and Coevolution: Ecological Specialisation.............. 15
2.2 Stability and Integration of Mutualisms......................................... 20
2.3 Wider Community Stability............................................................ 26
References.................................................................................................. 30
Part II Lessons from Classic Mutualisms

3 Classic Themes: Pollination Mutualisms of Insects and Plants........... 37
3.1 Introduction: Pollination Systems................................................... 37
3.2 Obligate Pollination Mutualisms.................................................... 46
3.2.1 Figs and Fig Wasps............................................................. 47
3.2.2 Yuccas and Yucca Moths.................................................... 49
3.2.3 Other Examples.................................................................. 51
3.3 Nursery Pollination......................................................................... 55
3.4 Broader Perspective........................................................................ 56
References.................................................................................................. 59
4 Classic Themes: Ants, Plants and Fungi................................................ 63
4.1 Introduction: Ant-Plant Interactions............................................... 63
4.2 Extrafloral Nectaries....................................................................... 68
4.3 Food Bodies and Domatia.............................................................. 75
4.4 Ant-Epiphytes................................................................................. 84
4.5 Ant Gardens.................................................................................... 85
xi
xii Contents
4.6 Myrmecochory................................................................................ 87
4.7 Ants, Plants and Fungi.................................................................... 93
References.................................................................................................. 98
5 Classic Themes: Ants and Other Insects – Hemiptera......................... 105
5.1 Introduction: Trophobiotic Associations......................................... 105
5.2 Ant-Hemiptera Mutualisms............................................................ 106
References.................................................................................................. 116
6 Classic Themes: Ants and Other Insects – Lepidoptera....................... 119
6.1 Introduction: Ants and Lycaenid Butterflies................................... 119
6.2 Myrmecophily................................................................................. 120
References.................................................................................................. 139
7 Other Insect Mutualisms......................................................................... 143
7.1 Introduction: The Remaining Variety............................................. 143
7.2 Termites and Fungi......................................................................... 143
7.3 Beetles and Fungi........................................................................... 145
7.4 Mullerian Mimicry.......................................................................... 146
References.................................................................................................. 147
Part III Environmental Impacts and Insect Conservation:

Lessons from Mutualisms
8 Impacts of Alien Invasive Species........................................................... 153
8.1 Introduction..................................................................................... 153
8.2 Impacts on Mutualisms................................................................... 154
References.................................................................................................. 164
9 Exploiters of Mutualisms........................................................................ 167
9.1 Introduction: Roles of Additional Species...................................... 167
9.2 Exploiters and Cheaters.................................................................. 170
9.3 Protection from Over-Exploitation................................................. 175
9.4 Changed Communities.................................................................... 177
9.5 Nectar and Pollen Robbers............................................................. 182
9.6 Defences Against Exploiters........................................................... 185
References.................................................................................................. 189
10 Habitat Losses and Conservation of Mutualisms................................. 193
10.1 Introduction: Mutualisms in Practical Conservation...................... 193
10.2 Mutualisms and Disturbance.......................................................... 198
10.3 Community Contexts...................................................................... 203
10.4 Fragmentation and Loss of Habitats............................................... 204
10.5 Climate Change.............................................................................. 208
10.6 Reintroductions............................................................................... 216
10.7 Prospects......................................................................................... 218
References.................................................................................................. 220
Index.................................................................................................................. 225
Part I
The Meaning and Dynamics of Mutualisms
Mutualisms, essentially comprising some form of cooperation between species, and

through which each of the participating taxa benefit, have for long intrigued observ-
ers of the natural world. Over the last century or so, many of the observations of
early naturalists have been explored in increasing detail by both observation and
manipulation, and progressively applying ecological theory and modelling to eluci-
date the evolution, dynamics and ecological roles and impacts of those associations.
This first part gives some background to the main themes, to emphasise the many
and varied mutualistic associations in which insects participate, and demonstrates
the wide range of circumstances in which mutualisms may develop or be
disrupted.
As Leigh (2010) commented, ‘some of these mutualisms represent major evolu-
tionary innovations’. Their existence and wider associations within communities
render some of them ‘ecological hubs’, foci from which ecological effects ramify
widely, and from which disruptions or changes to local environments or the species
themselves may have severe consequences for participating taxa or local communi-
ties and ecosystems. Those species that are obligate mutualists, and cannot thrive
outside that association, or are highly specific exploiters of conditions created by a
mutualism (such as by being specialised and specific ‘parasites’ of the association)
may become ‘threatened’: some such insects are designated as conservation priori-
ties. Collectively, mutualisms embrace numerous insect taxa and the plants, other
insects and other animals with which they interact, and span a massive range of
ecological environments.
Reference
Leigh EG (2010) The evolution of mutualism. J Evol Biol 23:2507–2528

Chapter 1
The Scope and Meaning of Mutualisms
Keywords Antagonism • Biological trading • Interactions • Interdependence •

Interspecific associations • Symbiosis
1.1 Introduction
The functional variety of ecological interactions between species continues to chal-

lenge and resist any simple or unambiguous form of classification. Those associa-
tions are the core of biodiversity: simply, no species exists in isolation, and
categorising those various interactions in which each may participate incurs both
ecological and semantic difficulties. Perhaps the most familiar associations amongst
insects are direct trophic interactions, notably ‘herbivore-plant’, ‘predator-prey’,
and ‘parasitoid-host’, all of which are major processes in terrestrial and freshwater
communities, and which range from highly specific ‘one-to-one’ associations that
characterise linkages between ecologically specialised species to much more dif-
fuse generalised interactions, whereby polyphagous consumers exploit a variety of
other species for food, or any single food species is exploited by many different
consumers, so that the interactions are most commonly interpreted as ‘antagonistic’
or ‘one-sided’ with the consumer profiting at the expense of the consumed species.
In some cases, both participant species benefit from the association, as a mutualism
that may be either very specific or more diffuse, and either obligatory or facultative.
The numbers of participants and the levels of dependency or opportunism in such
associations can vary greatly. ‘Mutual benefit’ is the core of mutualism, and the
broad definition advocated by Pierce et al. (1991), as ‘an interaction in which the
genetic fitness of each participant is increased by the action of its partner’ incorpo-
rates a full range of variations from obligatory to highly facultative and diffuse, so
that it is difficult to derive any general model for mutualisms. Pierce et al. also sug-
gested that many non-obligatory mutualistic associations are ‘fragile’ in that alter-
native relationships and life styles may be both available and become more attractive
to pursue, should conditions change.
Two contrasting forms of diffuse mutualism, discussed by Gove et al. (2007),
were designated as (1) evenly diffuse, in which the various multiple partners are
© Springer International Publishing AG 2017 3

T.R. New, Mutualisms and Insect Conservation,
DOI 10.1007/978-3-319-58292-4_1
4 1 The Scope and Meaning of Mutualisms
evenly represented and similar in terms of frequency and the outcomes of their pres-
ence, and (2) unevenly diffuse, in which one or few species are especially important
and the others less so – some of them perhaps even being functionally redundant.
The latter condition is associated with the higher importance species being regarded
as ‘keystone species’, with the fate of these then far more important for persistence
of the interaction than if losses of subservient species occur. The consequences for
conservation clearly differ, and their distinction may have practical relevance. Many
myrmecochorous associations (Chap. 4), for example, are commonly presumed to
be reasonably evenly diffuse, but several studies have queried this. Gorb and Gorb
(1999) suggested that seed dispersal by ants is most effective when a few dispersing
species predominate: if so, uneven diffusion becomes an advantage. This was vali-
dated by assessing the role of the purported mutualistic keystone genus Rhytidoponera
in Western Australia (Gove et al. 2007; p. 209). Transect and site surveys showed
that Rhytidoponera, although comprising only 2% of the ants observed, removed
72% of cafeteria-displayed seeds. Those seeds were also moved further and, in gen-
eral, seeds removed by other ant genera were either rare or dropped after short
distances.
Superimposed on each of these interactions is the universal possibility of inter-
specific competition if and when a resource species becomes limiting, either through
its own scarcity or because too many consumers seek it. All of these interactions are
viewed as, typically, antagonistic and involving exploitation of some species by oth-
ers. Many species at the lower trophic level (food!) have evolved strategies to avoid
being eaten, and many consumers have developed corresponding responses to over-
come or avoid those defences. Doing so is one avenue towards specialisation and
has historically been regarded largely as a scenario of ‘warfare’ with each partici-
pant successively regaining advantage over the other as counteradaptations prog-
ress, and leading to increasingly subtle associations.
Another major category of ecological associations also manifests ecological spe-
cialisations, but has rather different functional emphasis. Mutualisms, implying
(sometimes obligate) relationships between different species for their mutual bene-
fit and without which they could not persist or their wellbeing is diminished, are
both amongst the most intricate and most ecologically specialised interactions, and
have been studied intensively from many points of view. Their origins, evolution,
and maintenance continue to intrigue, and the circumscribed nature of many such
associations renders them ‘tangible’ as models, modules, or microcosms for experi-
mental study and manipulation and investigations of their functional roles in wider
communities, and the clarification of mechanisms of interactions between species.
The last can have much wider ecological relevance.
Many species interactions can vary considerably at both ecological and evolu-
tionary levels within the context of different communities. As in some ecological
contexts, ‘specialisation’ in mutualisms equates to a reduced number of partners
with which a species interacts, the extreme being a single obligatory mutualist part-
ner species. Many mutualisms involve multiple partners, with that variety assuring
a range of possible costs and benefits within the associations, and potential for evo-
lution of increasing levels of specialisation. The ‘quality’ of each potential partner
1.1 Introduction 5
may influence that evolution, in conjunction with partner abundance in space and
time, but reflect a gradient from lower quality to higher quality partners amongst the
potential spectrum of candidates. As Fraser et al. (2001) demonstrated from com-
parisons of different ant species attending larvae of the lycaenid butterfly
Glaucopsyche lygdamus (p. 136), the levels of protection from natural enemies,
persistence, and size achieved relative to untended larvae can differ markedly
between different ant partners. In that study, Formica podzolica provided clear ben-
efits to the butterfly larvae, two other ant taxa had essentially neutral impacts, and a
fourth species (Formica obscuripes) appeared to be parasitic, because its presence
increased larval disappearance and decreased their growth rates.
From a meta-analysis of the relevance of context across 247 published studies of
interspecific associations, Chamberlain et al. (2014) suggested that mutualisms may
be the most variable interactions, as they are the most context-dependent and reflect
greater complexity of resource exchanges as well as – in some cases – the context
including presence of other species that influence the cost/benefit ratio. The last
includes contexts such as varying positive values of a pollinator to a plant. Thus, the
roles of the moth Greya (Prodoxidae) as a pollinator of Lithophragma parviflorum
(Saxifragaceae) (p. 197), studied by Thompson and Cunningham (2002), vary with
the incidence of other pollinators that differ considerably in abundance across sites.
Such variation influences the dynamics of the three (or more) species rather than
that of the twin core partners alone. In some habitats, co-pollinators (which in this
case included several species of bees and bombyliid flies) potentially reduce the
dependence of Lithophragma on Greya for pollination, and the overall interaction
has the flexibility to be mutualistic, commensal (with pollination and oviposition
having no effects on floral development), or antagonistic, depending on the inci-
dence and abundance of effective co-pollinators in any given context.
However, mutualisms have intruded less frequently into insect conservation
practice than have antagonistic interactions. Presence of key natural enemies (pred-
ators or parasitoids) may be considered ‘threatening’ to focal insects and targeted
for suppression in campaigns to conserve these. However, knowledge of their inci-
dence and importance is needed in conserving both individual participant species
and some key ecological processes that depend on mutualisms.
Mutualisms are among the most intimate of all insect associations, and the con-
ditions under which they have developed, and persist, may reflect long co-
evolutionary heritages that are disrupted by recent environmental changes. The
intermingling of interactions that are mutualistic or non-mutualistic, and sometimes
hard to distinguish, can become rather subjective because of incomplete knowledge
of the benefits to each partner in what are clearly seen as regular and predictable
associations, and in which other taxa may become involved. Thus, herbivory by
leaf-cutting ants (Chap. 5) is a clear component of their mutualism with fungi, as
providing the food on which those fungi are nurtured, but is clearly more antagonis-
tic to the plants on which that herbivory takes place. The plants are exploited directly
by those ants in an antagonism far different from the numerous mutualistic cases in
which ants protect plants against herbivory.
The functional relationships between taxa in many interspecific associations that

occur frequently are very difficult to interpret as being predominantly ‘mutualism’
or ‘opportunism’. The tree weta (Hemideina spp., Anostostomatidae) and the New
Zealand tree fuchsia (Fuchsia excorticata, Onagraceae) illustrate some of the diffi-
culties that occur (Wyman et al. 2011). Because it seemed unlikely that any net
benefit to the tree could occur through weta feeding on the fruit, and weta were only
irregular dispersers of seed, the association was deemed opportunistic with the weta
feeding arising simply from them roosting on or near the trees. Many such associa-
tions, even when frequent and seemingly both regular and specific, and including
components that could be interpreted as mutually beneficial, parallel this example.
They may need careful investigation to confirm their correct functional status.
Classically, then, mutualisms are delimited as interactions between two species
that require each other for reproduction or survival and often involve strong coevo-
lution (or, at least, reciprocal adaptation) to develop and sustain the association.
Others are more variable, and much of the ambiguity of definition reflects that the
extent of reciprocal adaptation between partners can vary across their distributional
ranges – some interactions, as noted above, varying from mutualistic, through more
neutral associations (broadly ‘commensalisms’) to antagonistic as the intensity or
balance of the interaction varies over space and time. Indeed, mutualisms are some-
times viewed as balanced antagonistic interactions, as ‘mutual exploitation’, implic-
itly assuming that they are context-dependent and vary over both space and time. As
such, they are susceptible to invasions by cheaters and parasites (Chap. 8) (Male
et al. 2012) and are fundamentally unstable. Long-term persistence of associations
can thereby be accompanied by variation and instability reflected in local conditions
and change as those conditions alter. Very broadly, three major categories of interac-
tions between pairs of species can be thought of as ‘antagonisms’ (in which one
partner suffers from the attention of the other, which benefits), ‘facilitations’ (when
one partner gains from the other, which does not suffer), and ‘mutualisms’ (in which
both partners benefit from their interaction). The gradations in relative ‘fitness’ of
the participants between these may be very unclear and difficult to quantify. Shifts
in balance between costs and benefits to the participants over time can lead to break-
down or abandonment of a mutualism, or a return to parasitism – which is some-
times viewed as a possible general ancestral state of all mutualisms in which benefits
are traded (Edwards 2009). However, the pathways that move hosts and their para-
sites toward co-operation (and, hence, mutualism) are not wholly clear. ‘Tolerance’
might affect that transition, and Edwards (2009) proposed that the roles of tolerance
include (1) promoting the evolution of mutualism; (2) maintaining a mutualism; and
(3) promoting pathways toward mutualism breakdown. This variety was presented
by hypothetical examples in which the fitness of modern and ancestral resource
holders alone or with the partner are compared (Table 1.1).
Leigh (2010) recognised two basic types of mutualisms among different species.
One involves the members of different species converging in some way towards a
common appearance of behaviour mode that increases benefits to all participants, as
that convergence occurs in more individuals. The convergent appearance of differ-
ent distasteful diurnal butterflies in Mullerian mimicry (p. 146) is perhaps the best-
1.1 Introduction 7
Table 1.1 Hypothetical Reproduction of focal species

examples demonstrate Alone With partner
evolution of dependency and
mutualism, through tolerance, A
by comparison of Ancestral 2 0.5
reproduction of the focal Modern 1 1.5
species alone or with its B
partner. Figures are Ancestral 2 0
hypothetical fitnesses of
Modern 0 1.5
ancestral and modern
resource holders, with A–C C
representing different levels Ancestral 2 0.5
of intimacy between the Modern 2 3
parties After Edwards (2009)
A Evolved dependency, in which tolerance
allows resource holders to increase fitness in
presence of a partner, but at reduced fitness
when alone, B More extreme case of evolved
dependency, in which partners in ‘modern’
state reveal near-total loss of reproductive inde-
pendence, C Outright mutualism, when toler-
ance allows the resource holder to reduce cost
of interaction with antagonists, so that hosts
gain benefit from encounters with the partner
species compared to ancestral and modern fit-
ness when alone
known example. In this, the mortality ‘cost’ of educating predators is shared across
species that the predator cannot distinguish – and which each may benefit from the
shared predation. The second, far more widespread, type involves two (or more)
species exchanging ‘goods or services’ (most commonly involving aspects of food
and some form of protection from enemy attack) for their mutual benefit. Much of
the second part of this book deals with this theme – it is noted here simply because
many of the examples discussed in this introductory section fall into this category.
A second helpful dichotomy (or gradient), again discussed by Leigh (2010), is the
range from long-term relationships – often obligatory, such as mutualisms between
leaf-cutting ants and fungi (p. 95) or termites and gut fauna (p. 143), through more
limited-time obligate relationships involving only particular life stages (such as ants
with myrmecophilous butterfly larvae, p. 120) to much more transient ‘brief
exchange’ mutualisms such as many specific plant-pollinator associations (p. 46).
Three broad functional categories of mutualisms are sometimes distinguished to
reflect some key functional groups and display the collective variety of the interac-
tions. As summarised by Bronstein (2001a, b), these are (1) transportation mutual-
isms, in which the major benefit to one partner is movement of either itself or its
reproductive products, and provides some reward to the transporting partner: polli-
nation or seed dispersal with a food reward to the vector is, perhaps, the most famil-
iar context; (2) protection mutualisms, in which one partner protects the other from
negative impacts of its environment, exemplified by the many cases in which ants
(as an insect group participating in a great variety of mutualisms) are purported to

protect plants from attack by insect herbivores; and (3) nutrition mutualisms, in
which one partner obtains essential nutrients from its partner – so that it is often
viewed as a more one-sided interaction benefiting the feeder, with a corresponding
cost (that of providing the sacrificed food) to the other partner.
Very broadly, Connor (1995) recognised three categories of mechanisms whereby
one species or individual may gain benefits from another, as (1) by-products, when
one party passes benefits to another as a by-product of acts performed for its own
wellbeing and at no additional cost to itself (Chap. 10); (2) investment, when selec-
tion favours an individual performing a costly act that benefits another individual,
when ‘return on investment’ may take various forms; and (3) purloined benefits,
acquired through predation or parasitisation. Connor’s scope included both inter-
specific interactions and those between conspecific individuals, but the categories
allowed him to suggest various forms of ‘basal mutualisms’ that collectively dem-
onstrated the functional complexities involved and the difficulties of interpreting the
mechanisms involved, whether the associations were obligate or facultative, perma-
nent or transient, highly specific or involving more taxa, and so on. The idea of
mutualisms being ‘biological markets’ (see Bronstein 2001a, b) implies that species
offer their partners commodities that are relatively inexpensive for them to produce,
in exchange for others that are more expensive or that they are incapable of produc-
ing without that partner. The same principle occurs in Ollerton’s (2006) likening of
mutualisms to ‘biological barter’, when he also noted that the dichotomy between
the trade goods compartmentalised as ‘materials’ and ‘services’ is a great oversim-
plification of a complex continuum in which their relativity may be hard to define.
Ollerton cited plants that produce resins as rewards for pollinating bees, and which
they use as material for nest construction, so is a physical resource and an essential
service in protecting bee offspring. Some associations, such as some ant-Acacia
mutualisms (Chap. 4) combine service and resource, with the Acacia providing nest
sites for the ants (service) and food bodies (resources) for them.
The ‘trading’ of services between species in mutualism entails that the partners
must each respond to contributions of the other. If it is costly to provide any service,
a conflict of interest may arise unless such responses occur – so that the stability of
the relationship can depend on reactions to the other partner’s contributions (Axelrod
and Hamilton 1981). Conflicts of interest are potentially disruptive, and identifying
sources of conflict and the features that govern costs and benefits to each partner
remains an important central theme. Its scope was indicated by Herre et al. (1999),
who emphasised that a primary need is ‘a clear appreciation of the spatial, temporal
and taxonomic context in which these systems operate’, with attention to both eco-
logical and evolutionary scales. Understanding and quantifying those relative costs
and benefits for each party is a fundamental need in describing and interpreting
mutualisms. Herre et al. also noted that complete agreement of the mutualists’ inter-
ests is always unlikely, and suggested several factors that can help to align their
interests or, conversely, to hamper them (Table 1.2). Many cases appear to conform
reasonably well to these parameters, but exceptions are not unusual. Those can
reflect amongst other factors (1) local conditions, either physical or the influence of
1.1 Introduction 9
Table 1.2 Some factors that Passage of symbionts from parent

might help to align interests to offspring, as ‘vertical
of mutualists, as listed by transmission’.
Herre et al. (1999)
Genotypic uniformity of symbionts
within individual hosts.
Spatial structure of populations
leading to repeated interactions
between potential mutualists.
Restricted options for both partners
outside the relationship.
The converse conditions (horizontal
transmission of symbionts between
unrelated hosts, multiple symbiont
genotypes, varied alternative options)
were thought to hamper or ‘unravel’
mutualisms
other species from their presence and abundance; (2) variations in resource avail-
ability and the ‘fitness’ of the participating species; and (3) variations in numbers
and density that may affect patterns of mutualist transmission. Different evolution-
ary outcomes across related taxa – as for many of the contexts outlined in Chaps. 4,
5, 6, 7, and 8 – can also occur, as evident from the complexities displayed within,
for example, fig-fig wasp or ant-butterfly mutualisms.
The notion of ‘biological trading’ was pioneered through Noe and Hammerstein
(1994, 1995), whose accounts considered the constraints and variety of opportuni-
ties that can occur through ‘honest trades’ and ‘cheating’ (Chap. 8). It was founded
by the realisation that some form of trading is the basis of many mutually beneficial
interactions, and that a pairwise trade can become far more complex to interpret if
the number of different traders increases. Thus, mutualisms between ants and plants
(Chap. 5) become complicated by Hemiptera as additional participants, and the gra-
dations between mutualism and antagonism become blurred if the ants both tend
and eat those bugs. Estimating the costs and benefits is a key aspect of exploring
mutualisms, and such tripartite interactions are mutualisms if, for each party, the
costs of the interaction are exceeded by the benefits, so that the net benefits are
reciprocal (Gaume et al. 1998). The universal basic presumption is that ‘benefits’ to
one species may entail ‘costs’ for the other, with the balance between these deter-
mining the outcome of the interaction – whether the association is mutualistic or
antagonistic, and how that balance may change. Many different factors contribute to
any such changes, with the proviso that extreme specialisation may also increase
risk of extinction. Considering associations between aphids and ants (Chap. 6),
influential factors encompass ecological, physiological, and evolutionary compo-
nents (Stadler and Dixon 2005). Each factor affects the balance of the association
and whether it becomes mutualistic or antagonistic. Of the three categories, evolu-
tionary components are by far the least studied, but Stadler and Dixon suggested
that they were likely to be as influential as physiological or ecological components.
The relative importance of all these influences on ant-aphid relationships is difficult

to assess, and doing so necessitates long-term studies rather than the more usual
short-term investigations.
The costs involved can be very varied, and many are difficult to quantify or even
detect. Thus, in plant-pollinator associations the costs to a plant that can vary with
number of pollinator visits were listed by Morris et al. (2010), who noted the pau-
city of knowledge on those costs and how they (together with benefits to the plant)
change with visitor abundance in any typical pollination mutualism. The listed costs
were (1) the replenishment of nectar (or other reward) collected by visitors; (2)
damage to the flowers caused by visitors, or attempts to repair this; (3) removal of
pollen deposited by previous visitors on flower stigmas; (4) increased competition
among pollen tubes so that a lower proportion of ovules are fertilised; (5) infection
by pathogens carried by visitors; and (6) infection by yeasts from visitors, and
which reduce nectar quality and affect future visitors. A complementary theme also
noted by Morris et al. is that the costs and benefits to pollinators interacting with the
plants may also influence – and perhaps affect – their roles in serving the plant.
The functional status of, and correct descriptive term for, some such associations
will continue to be debated, but some unusual mutualisms reported suggest that they
could represent more widespread scenarios. Some rove beetles (Staphylinidae:
Amblyopini), for example, are known almost solely from the fur of small mammals,
suggesting a possible blood-feeding ectoparasite habit – an idea supported by pres-
ence of blood in their guts. However, Ashe and Timm (1987) investigated the beetle
Amblyopinus tiptoni on the mouse Peromyscus nudipes in Costa Rica and found no
evidence of them feeding on blood, body secretions, or epidermal debris. In labora-
tory trials, the beetles were predators on the flea and mite ectoparasites of the
rodents, and this habit inferred that earlier records of blood-feeding arose from sec-
ondary ingestion of these blood-feeding prey. This interaction was interpreted as a
mutualism through which the beetles benefited by a persistent association with the
host and a food supply, and the rodent benefited from significant reduction of its
ectoparasite load. That the hosts largely ignore the beetles rather than remove them
was suggested to imply that this benefit is recognised. As Ashe and Timm noted,
nutritional interactions between vertebrates and insects appear to be rare – or, at
least, to have been only rarely reported or investigated.
Confusions over the meaning of mutualism have arisen also from the varied uses
of the term ‘symbiosis’; in some contexts, it equated to mutualism but in others
broadened to include harmful or neutral impacts between the participants – so that
‘symbiosis’ then becomes broadly a synonym of ‘interaction’ with the partners in
some way living together. Thus, in their major review of associations between ants
and plants (Chap. 5), Davidson and McKey (1993) acknowledged the evolutionary
specialisations involved but the interactions were regarded as ‘symbioses’, because
‘For myrmecophytes overall, existing evidence is often too meagre for a convincing
assessment of the fitness consequences of particular associations’, so that less
restricting descriptors were preferred. Likewise, terms such as ‘myrmecophyte’ and
‘ant-plant’ were not taken by them, in contrast to some other authors, to imply func-
tional associations alone but to include also more sporadic or casual co-incidences.
1.1 Introduction 11
In short, many mutualistic environments or participating species are exploited or

shared by other species, for which the roles they play are largely unknown; however,
many insects are regularly associated with one or the other of the mutualist partners
and occur nowhere else. Thus, many specialised beetles are mimetic cohabitants of
termites or ants in their nests, as ‘termitophiles’ or ‘myrmecophiles’, but relatively
few of these form any true mutualistic relationship with their host(s). Some form of
passive co-existence or antagonism has been reported far more commonly, and
some claimed mutualisms may reflect transient encounters rather than any of more
regular benefit to the parties. Some staphylinid beetles living in termite nests pro-
duce exocrine secretions that are licked by termites which, in return, regurgitate gut
contents used by the beetles (Pasteels and Kistner 1971) – but whether either par-
ticipant depends in any way on that exchange is uncertain and, perhaps, unlikely,
notwithstanding that the interaction appears to be mutually positive. In contrast, and
perhaps applicable to the great majority of ‘other species’ found in social insect
nests, ‘interactions between the builder species and its termitophiles (added: or myr-
mecophiles) have been inferred more than explicitly tested’ (Rosa et al. 2008), not-
ing also that termite hosts may tolerate those species because (1) they cannot defeat
them; (2) the guest rewards the termites; or (3) the guests are simply not noticed.
Many such associates have no direct contact with the host species but integrate into
the host’s social life through appeasement (which may be chemical, morphological,
or behavioural) of any aggression or antagonism displayed. Specific interactions
between given termite and staphylinid beetle species, for example, may involve
close chemical or morphological mimicry (Cunha et al. 2015). Highly specific sym-
bioses amongst ecologically specialised species may give the impression of mutual-
ism, but not be so.
The extent of specialisation between partners in mutualism, in extreme cases
self-evident, continues to be debated in many others. Steffan-Dewenter et al. (2007),
for example, discussed the different scales present among plant-pollinator associa-
tions and used a diagram (Fig. 1.1) to depict four general scenarios to compare and
assess that variety. The first of these can lead to rapid extinction should either part-
ner be lost, whilst the two intermediate states represent high reliance on one or other
partner. These also may incur high extinction risk. It is also self-evident that many
so-called dependent species, such as specific parasitoids and predators in funda-
mentally antagonistic relationships, become vulnerable if their host or prey species
declines (Moir et al. 2012).
Cushman and Beattie (1991) showed that mutualisms may or may not be symbi-
otic, as the interacting species may not even contact each other in indirect mutual-
isms. Whatever working concept is adopted, the burden of demonstrating true
‘mutual benefit’ is, in principle, easy to postulate but much harder to prove convinc-
ingly. Most studies available to Cushman and Beattie for interpretation failed to
verify this fundamental condition. The term ‘symbiotic mutualism’ refers to persis-
tent associations between species that gain mutual benefit and ‘lie at the intersection
between the overlapping categories of mutualisms and symbioses’ (Douglas 2015).
Mutualisms are intrinsically and widely associated with co-evolutionary changes
among the partners, and presumed to represent ‘reciprocal specialisation’, whereby
One plant species One pollinator species
Several plant species One pollinator species
One plant species Several pollinator species
All plant species All pollinator species
Fig. 1.1 A generalised scheme to summarise scenarios for specialisations among plant-pollinator
interactions, reflecting capabilities of specialised and generalised species (Steffan-Dewenter et al.
2007)
co-adaptations may become increasingly intricate and specialised. However, ‘asym-

metric specialisation’ is also likely and, following Vazquez and Aizen (2004), may
be far more widespread than earlier supposed. In this context, either a specialist
interacts with a generalist, or the converse – possibly with only a fraction of special-
ists interacting with other specialists, and others with extreme or moderate general-
ists. Models suggested that this kind of interaction is common in plant-pollinator
interactions, and large assemblages with many pairs of interacting species tend to
have a higher proportion of asymmetric interactions than found in smaller assem-
blages with few interacting pairs. In that system, specialist plants do not always
interact with specialist pollinators, as also with generalist plants and generalist pol-
linators (Ashworth et al. 2004). In contrast to the reciprocal specialisation that char-
acterises obligate mutualisms, in particular, and which in principle could be related
to high risk of extinction, the wider interactions in which generalist species partici-
pate may both promote survival and help to assure the survival of specialist species
whose more specialist associates have, for some reason, succumbed. Many general-
ists are most likely to be influenced most by the small subset of possible associates
with which they interact most intensively – and which reflect the strength and fre-
quency of the interaction. For each participant in a mutualism, its partner can be
considered a resource, and additional species, exploiters, may then also exploit that
resource (Chap. 8).
‘Partner quality’ is relevant beyond simply maintaining the mutualism, in deter-
mining the scope and intensity of influences that occur. Plant defences by mutual-
ists, for example, should not exclude other mutualists, such as pollinators by their
activities – but it is almost inevitable that these twin aspects may come into conflict,
as aggressive defenders (such as ants) may also counter plant sexual reproduction.
Features of the plant may help to reduce such conflicts (Ness 2006). Ness explored
mechanisms by which ant protection of the cactus Ferocactus wislezeni (a species
with extrafloral nectaries: p. 68) affect pollinators in Arizona. Flowers occupied by
the most aggressive of the four associated ant species (Solenopsis xyloni) were vis-
ited less frequently by pollinating bees, visit times were shorter, and flowers pro-
References 13
duced fewer and lighter seeds than flowers tended by the other ants. Ness attributed
the lower pollinator attendance to avoid predation risks, noting that solitary cactus
bees – in particular – should be especially ‘danger-averse’ and avoid high risk habi-
tats. In this example, which perhaps represents a more widespread scenario, the
most effective protective ant has key features (high abundance and high aggression)
that increase ecological costs to the plant. Over time, the plant’s fitness could come
to depend more on those mutualists that defend it less ably.
Asymmetry may indeed be usual in mutualisms (Aizen et al. 2008), so that a
plant species that depends strongly on an animal, such as a pollinator, commonly
reflects the animal depending only weakly on the plant, and the converse. Mutualistic
webs may comprise a core of interactions between generalist partners, and accom-
panying specialists that predominantly interact with generalists rather than with
other specialists. Modification of such webs by invasive species, many of them
generalists as a preadaptation facilitating establishment and spread in a new area,
can come as those taxa are integrated into native pollination networks (p. 159).
References
Aizen MA, Morales CL, Morales JM (2008) Invasive mutualists erode native pollination webs.
PLoS Biol 6(2):e31. doi:10.1371/journal.pbio.00600431
Ashe JS, Timm RM (1987) Probable mutualistic association between staphylinid beetles
(Amblyopinus) and their rodent hosts. J Trop Ecol 3:177–181
Ashworth L, Aguilar R, Galetto L, Aizen MA (2004) Why do pollination generalist and specialist
plant species show similar reproductive susceptibility to habitat fragmentation? J Ecol
92:717–719
Axelrod R, Hamilton WD (1981) The evolution of cooperation. Science 211:1390–1396
Bronstein JL (2001a) The exploitation of mutualisms. Ecol Lett 4:277–287
Bronstein JL (2001b) The costs of mutualism. Am Zool 41:825–839
Chamberlain SA, Bronstein JL, Rudgers JA (2014) How context dependent are species interac-
tions? Ecol Lett 17:881–890
Connor RC (1995) The benefits of mutualism: a conceptual framework. Biol Rev 70:427–457
Cunha HF, Lima JS, Dsouza LF, Santos LGA, Nabout JC (2015) No morphometric distinction
between the host Constrictotermes cyphergaster (Silvestri) (Isoptera: Termitidae:
Nasutiterminae) and its obligatory termitophile Corotoca melantho Schiodte (Coleoptera:
Staphylinidae). Sociobiology 62:65–69
Cushman JH, Beattie AJ (1991) Mutualisms: assessing the benefits to hosts and visitors. Trends
Ecol Evol 6:193–195
Davidson DW, McKey D (1993) The evolutionary ecology of symbiotic ant-plant relationships.
J Hymenopt Res 2:13–83
Douglas AE (2015) Conflicts, cheats and the persistence of symbioses. New Phytol 177:849–858
Edwards DP (2009) The roles of tolerance in the evolution, maintenance and breakdown of mutu-
alism. Naturwissenschaften 96:1137–1145
Fraser AM, Axen AH, Pierce NE (2001) Assessing the quality of different ant species as partners
of a myrmecophilous butterfly. Oecologia 129:452–460
Gaume L, McKey D, Terrin S (1998) Ant-plant-homopteran mutualism: how a third partner affects
the interaction between a plant-specialist ant and its myrmecophyte host. Proc R Soc Lond B
265:569–575
Gorb SN, Gorb EV (1999) Dropping rates of elaiosome-bearing seed during transport by ants
(Formica polyctena Foerst.): implications for distance dispersal. Acta Oecol 20:509–518
Gove AD, Majer JD, Dunn RR (2007) A keystone ant species promotes seed dispersal in a ‘diffuse’
mutualism. Oecologia 153:687–697
Herre EA, Knowlton N, Mueller UG, Rehner SA (1999) The evolution of mutualisms: exploring
the paths between conflict and cooperation. Trends Ecol Evol 14:49–53
Male P-JG, Leroy C, DeJean A, Quilichini A, Orivel J (2012) An ant symbiont directly and indi-
rectly limits host plant’s reproductive success. Evol Ecol 6:55–63
Moir M, Vesk PA, Brennan KEC, Poulin R, Hughes L, Keith DA, McCarthy MA, Coates DJ (2012)
Considering extinction of dependent species during translocation, ex situ conservation, and
assisted migration of threatened hosts. Conserv Biol 26:199–207
Morris WF, Vazquez DP, Chacoff NP (2010) Benefit and cost curves for typical pollination mutual-
isms. Ecology 91:1276–1285
Ness JH (2006) A mutualism’s indirect costs: the most aggressive plant bodyguards also deter pol-
linators. Oikos 113:506–514
Noe R, Hammerstein P (1994) Biological markets – supply-and-demand determine the effect of
partner choice in cooperation, mutualism and mating. Behav Ecol Sociobiol 35:1–11
Noe R, Hammerstein P (1995) Biological markets. Trends Ecol Evol 10:336–339
Ollerton J (2006) “Biological barter”: patterns of specialization compared across different mutual-
isms. In: Waser NM, Ollerton J (eds) Plant-pollinator interactions from specialization to gener-
alization. University of Chicago Press, Chicago/London, pp 411–435
Pasteels J, Kistner D (1971) Revision of the termitophilous subfamily Trichopseniinae (Coleoptera:
Staphylinidae). II. The remainder of the genera with a representational study of the gland sys-
tems and a discussion of their relationships. Misc Publ Entomol Soc Am 7:351–399
Pierce NE, Nash DR, Baylis M, Carper ER (1991) Variation in the attractiveness of lycaenid but-
terfly larvae to ants. In: Huxley CR, Cutler DF (eds) Ant-plant interactions. Oxford University
Press, Oxford, pp 131–142
Rosa CS, Marins A, DeSouza O (2008) Interactions between beetle larvae and their termite hosts
(Coleoptera; Isoptera, Nasutiterminae). Sociobiology 51:1–7
Stadler B, Dixon AFG (2005) Ecology and evolution of aphid-ant interactions. Annu Rev Ecol
Evol Syst 36:345–372
Steffan-Dewenter I, Klein A-M, Gaebele V, Alfert T, Tscharntke T (2007) Bee diversity and plant-
pollinator interactions in fragmented landscapes. In: Waser NM, Ollerton J (eds) Plant-
pollinator interactions from specialization to generalization. University of Chicago Press,
Chicago/London, pp 387–407
Thompson JN, Cunningham BM (2002) Geographic structure and dynamics of coevolutionary
selection. Nature 417:735–738
Vazquez DP, Aizen MA (2004) Asymmetric specialization: a pervasive feature of plant-pollinator
interactions. Ecology 85:1251–1257
Wyman TE, Trewick SA, Morgan-Richards M, Noble ADL (2011) Mutualism or opportunism?
Tree fuchsia (Fuchsia excorticata) and tree weta (Hemideina) interactions. Austral Ecol
36:261–268
Chapter 2
Mechanisms and Maintenance
Keywords Ants • Coevolution • Coextinction • Mutualism • Partner choice •

Pollinators
2.1 Coextinction and Coevolution: Ecological Specialisation
Ever since Darwin’s (1862) well-known prediction that a Madagascan orchid with
an unusually long nectary spur could be pollinated only by a moth with an equiva-
lently long proboscis, such highly specific associations between species have stimu-
lated much thought on mutualisms and coevolution, and the corollary of coextinction.
Darwin had received specimens of the orchid Angraecum sesquipedale, and mar-
velled at the length of the flower as ‘nearly a foot long’. He suggested that the then-
unknown pollinator must be a large moth, but the sole known pollinator, the African
sphingid Xanthopan morganii (earlier suggested by Wallace 1867, as a likely candi-
date), was not confirmed to be present in Madagascar until 1903. It was later con-
firmed as pollinating the orchid, using its unusually long proboscis to probe for
nectar and gaining pollen on its head for transfer to the next orchid visited. Such
highly specific and obligatory mutualisms, with little – if any – possibility for alter-
native participants, are believed widely to engender coextinctions, because loss of
one participant must inevitably affect the other(s). However, proving the reality of
that interdependence extends well beyond correlating morphological features and
species’ coincidence. The above example has frequently been advanced as a ‘text-
book case’ of coevolution embedding a mutualism, but there is little evidence to
support this. Indeed, Anderson and Johnson (2008) commented that the orchid may
actually have been too rare to exert much selective pressure on moth proboscis
length. The story of this intriguing case has recently been summarised and revised
(Arditti et al. 2012), to confirm from video footage that the moth indeed removes
pollinia from the orchid. Nevertheless, that pollination was observed first only in
1992—130 years after Darwin’s initial speculation. Parallels of elusive or rare
insect or plant species supposedly depending on equally intricate associations are
not uncommon.
Paralleling the above case, the Neotropical riodinid Eurybia lycisca is highly
unusual among butterflies in having a proboscis about twice the length of its body

DOI 10.1007/978-3-319-58292-4_2
16 2 Mechanisms and Maintenance
and by which it obtains nectar from flowers of Calathea crotalifera (Marantaceae)

(Bauder et al. 2011), appearing to be finely adapted for this. However, it does not
pollinate the flowers so that the association – despite superficial appearance to the
contrary – is not a mutualism but one of exploitation by the butterfly. Calathea flow-
ers are pollinated by long-tongued euglossine bees. Eurybia was regarded as a ‘nec-
tar thief’ (p. 67) that has gained access to nectar resources of flowers pollinated by
other long-tongued insects. The evolutionary costs of adopting this strategy are bal-
anced by access to a resource that few other taxa can exploit. However, although
nectar thieves (or ‘nectar robbers’) have historically been regarded as antagonistic
to the plants they exploit, their roles can be far more complex, with impacts on pol-
linators, pollination, and wider fitness of the target plant (Maloof and Inouye 2000).
Long-tongued skipper butterflies (Hesperiidae) also steal nectar from Calathea as
among their preferred host plants in Costa Rica but do not contribute to pollination.
In Bauder et al.’s (2015) words, such butterflies ‘ransack Calathea nectar stores,
leaving less nectar for pollinating Euryglossini’. Long-tongued hesperiids are abun-
dant visitors to Calathea, where each skipper can empty about half of a flower’s
nectar supply in each visit. They might thus have serious impacts on the nectar sup-
ply available for those legitimate pollinators.
The theme of coextinction was revisited by Dunn et al. (2009), who pointed out
the anomaly that some theoretical models predict that a mass coextinction event,
involving thousands of species, must be occurring, but very few cases have actually
been documented. The postulated gradient between hypotheses that (1) coextinc-
tions are indeed common but most have not been observed and (2) coextinctions are
actually rare because mutualists and their partners are far more adaptable than gen-
erally supposed allows for numerous intermediate conditions and hypotheses over
true relationships. Whether any generalities occur remains highly uncertain. The
first extreme above reflects that most mutualists and partners are small, often local-
ised, and poorly documented, with the levels of specificity largely unknown. The
last is a key determinant of their fate, so that there is need to counter possible ‘trad-
eoffs’ between traits that facilitate survival with alternative partners. For parasites or
insect herbivores, it is common to find that the true host or host plant spectrum is far
wider than initially apparent, and ecological flexibility thus far greater than putative
extreme specificity might imply. Many supposedly monophagous herbivores prove
on investigation to feed on a wider array of plants, and the testing regimes devel-
oped for evaluating the specificity of insect herbivores as classical biological control
agents for weeds have been fertile stimuli for discoveries in this area. Likewise,
many predators or parasitoids famed for apparently specific attack patterns can, if
opportunities are presented, attack other prey or host species. Most such interac-
tions are not mutualisms, but the ecological flexibility displayed is a salutary cau-
tion for interpreting any interspecific associations.
Understanding the principles and mechanisms of specific ecological interactions,
and how they are honed and sustained, is central to conservation, together with the
(far harder to assess and most commonly ignored) reality that specificity amongst
the interacting species is dynamic and may differ markedly under different condi-
tions, and change with both evolutionary and ecological pressures and opportunity.
2.1 Coextinction and Coevolution: Ecological Specialisation 17
Dunn et al. (2009) noted two such contexts as (1) some individuals in a population
may be less host specific than others, resulting in a range of host specificities with
varying frequencies, perhaps helping to buffer the population against coextinction,
and (2) alternate strategies may occur – some plants that can be pollinated only by
a specific pollinator, for example, may survive also by other means such as longev-
ity, clonal reproduction, or facultative self-pollination. Many mutualists are thus
likely to be far more resistant to loss of their associates than is initially obvious. The
gradations between more intricate mutualisms and wider aspects of dependency and
symbioses are thereby complex.
Mutualistic interactions often help to align the interests of the partners, so that
cooperation increasingly benefits both. The intricacies of the interactions and their
possible disruptions (such as by ‘cheaters’: Chap. 9) have led to suggestions
(Frederickson 2013) that mutualisms may persist because the host species sanctions
cheaters, those ‘sanctions’ being specific adaptations to accommodate cheaters, so
that cheater and host may also coevolve. In some classic pollination mutualisms
(notably figs-fig wasps, and yuccas-yucca moths, Chap. 3), however, Frederickson
believed this form of coevolution is unlikely because cheaters are too rare to do so.
Stability in those associations appeared to depend, not on coevolution of cheating
and sanctions, but more on host preadaptations than the cheaters.
Coevolution between partners in a mutualism has for long been considered a
major factor begetting species diversity, and some of the gaps in suggestions of
evolutionary timing and the specific pathways involved for any such association can
now be addressed more effectively. Modern ‘molecular clocks’ and allied tech-
niques provide an arsenal of new tools that can be used to assess relationships and
divergence rates, and studies continue to appear that elucidate earlier hypotheses on
such themes. As one relevant example, some Macaranga ant-plants in south-east
Asia are myrmecophytic and harbour mutualistic Crematogaster (subgenus
Decarema) ants (p. 94), with those ants also tending scale insects (Coccus sp.) as
trophobionts. This tripartite association, in turn, is parasitised by myrmecophilous
lycaenid butterflies (Arhopala spp.). Larvae of the butterflies eat Macaranga foli-
age, evading the ants by appeasing them with nectar-like secretions (p. 120), and
each lycaenid species occurs on only one or two Macaranga species. Two hypoth-
eses (or ‘models’) have been advanced to elucidate the function of these associa-
tions, namely, (1) ‘codiversification’, in which the Arhopala-Macaranga association
began at about the same time as the Crematogaster-Macaranga mutualism, as a
consequence of continuous community interactions; and (2) ‘latecomer’, where the
lycaenids associated with the pre-occurring Crematogaster-Macaranga-Coccus
system, and adaptations of the latter fostered lycaenid diversification. Ueda et al.
(2012a, b) used molecular techniques to support the latter model, implying that the
minimum age of the lycaenids (2.05 million years) was only about a tenth of the age
of the plant-ant association with which the scale insects became involved about
9-7 million years ago. The initial plant-ant association is far older, and codiversified
for the past 20–16 million years. The development pathway of these intricate asso-
ciations thus became clearer, but with the confirmation that the participants have a
long history of possible interactions.
By whatever means the associations evolved, the sentiment that ‘mutualisms can
bind multiple species to a common fate’ (Kiers et al. 2010) is an almost universal
principle. As Moir et al. (2010, 2011) emphasised, ‘coextinction’ is still poorly
understood. The term was defined by them as ‘the loss of dependent species due to
change in their host populations, such as reduced host abundance or removal of
individual hosts from the wild’, and this definition has been widely paralleled else-
where – such as by Colwell et al. (2012) as ‘extinction of a species as a consequence
of its dependence on another that has gone extinct or declined below some threshold
abundance’. In complex associations, extinction cascades (multiple species extinc-
tions of interdependent taxa following loss of a pivotal species or guild) are likely
to occur. The level of host specificity and mutual dependence is central – simply,
truly ‘dependent species’, most familiar as the numerous insect herbivores with
near-monophagous feeding habits, can occur only with their host(s). Most such
associations are not conventionally seen as mutualisms but rather as one-sided
exploitative interactions in which the consumer benefits at the expense of the con-
sumed. They are, however, an important pathway through which mutualisms can
arise.
The widespread lack of evidence from empirical studies presents a paradox. This
was addressed recently by Powell (2011), in the perspective of coextinctions in
nature being resisted more strongly than suggested, and being genuinely rare events,
and interpretation hampered by implications of inaccurate (or incomplete) aware-
ness of host ranges. Each of these themes poses practical problems in assessment.
Thus, undersampling is very common in insect studies, and can lead to overesti-
mates of host specificity and the form of interspecies associations, and many coex-
tinction events amongst cryptic and poorly known insects may be underestimated.
Powell’s study, of some Australian psyllids (Hemiptera) on native acacias, explored
a further theme – that coextinction may falsely appear to be rare if some participant
species are lost well before their declining extant partners, so that temporal disrup-
tion obscures the functional coextinction. The more general prediction of her study
was that fewer specialist insects will be associated with an increasingly rare plant,
so that a plant species recognised as ‘threatened’ may support few specialist associ-
ates. Acacia-frequenting psyllids can have high levels of host specificity, and
although Powell’s endorsement of the above idea was not strong, early loss of spe-
cies from declining plant species was considered a plausible explanation for the
‘coextinction paradox’, with the additional implication that fewer generalist psyl-
lids were also associated with declining Acacia spp., possibly reflecting low encoun-
ter rates and lack of familiarity with increasingly rare host plants.
Difficulty of documenting coextinction also includes the incomplete knowledge
of systematics and species boundaries of the associated taxa, and limitations of
historical collections (Colwell et al. 2012), so that species’ identities may be con-
fused, historical misidentifications or taxonomic uncertainty abound, and long-term
trends in incidence are obscure. The above psyllids exemplify this situation, which
appears widespread among radiations of plant-feeding insects. Many endemic psyl-
lid species occur on Acacia (Yen 1980), but most have not been described or diag-
nosed fully and their host plant ranges have not been defined fully, so that their
2.1 Coextinction and Coevolution: Ecological Specialisation 19
160
140
No. extinct butterfly species

120
100
Observed no. extinct species
80
Estimated no. extinct species
60
40
20
200 400 600 800 1000
No. extinct plant species
Fig. 2.1 Plot of mean butterfly species extinctions against total host plant extinctions in Singapore.
This simulation by logistic regression reveals that the mean number of extinct butterflies increases
exponentially with the total number of extinct host plants. The number of extinct butterfly species
predicted is lower than that actually observed (Simplified from Koh et al. 2004)
vulnerability through host obligations is almost impossible to confirm. Nevertheless,

it remains true that extinction of any single participant in an association rarely
occurs in isolation. Losses of butterfly host plants in Singapore indicated possible
coextinctions of the butterflies themselves (Koh et al. 2004). Simulation exercises,
involving logistic regression analyses, suggested that the number of extinct butterfly
species increased exponentially with the total number of extinct host plants
(Fig. 2.1). Butterflies might go extinct sooner than their host plants when declines
(rather than extinctions) of some host species reduce the butterflies to less than
minimum viable population levels. However, the many ‘dependent species’ are not
necessarily mutualists – most insect herbivores, for example, are more obviously
antagonistic, exploiting their food plants without any documented benefit to the lat-
ter but clearly dependent on some suitable food species being available.
At any trophic levels, susceptibility of dependent species to the fate of an obli-
gate host or food species is well established, and a central plank of much conserva-
tion management is to sustain such associations and supply. Losses of such partners
can have wide ramifications, but many remain largely unknown. Oestrid flies, for
example, are well known as causing myiases in domestic stock and, in that role,
arouse considerable attention to their control. Many are also highly host specific,
and Colwell et al. (2009) drew attention to their inevitable demise should their
endangered mammal hosts become extinct. Those hosts include three species of
rhinoceros (collectively hosting four species of Gyrostigma) and the African ele-
phant (with five oestrid species across four genera). No other hosts are known for
any of those flies. In common with many other insects, their loss could easily pass
unnoticed – but that of their hosts, either globally or regionally, arouses wide
attention.
2.2 Stability and Integration of Mutualisms
Breakdown or loss of mutualisms has been discussed extensively. Three major tra-
jectories are often projected (Sachs and Simms 2006): (1) mutualisms are vulnera-
ble to extinction, as one or the other partner population in an obligate mutualism
does not persist above the minimum threshold level needed for continuing interac-
tions; (2) mutualisms can shift to parasitisms, as parasitic individuals supplant
cooperative ones and disrupt the previously harmonious balance; and (3) changes in
the cost:benefit balance can shift to favour a return to autonomy, so that the mutual-
ism is abandoned.
If extinction is to be avoided, some form of buffering of fluctuations or distur-
bance is necessary. This links directly with conditions of how mutualisms persist
and the extent of their stability. Stability thus implies some form of maintenance of
the balance between the partner species and may involve intricate behavioural
responses to relevant chemical and physical cues which are themselves sensitive to
wider environmental changes.
Stability within mutualisms may depend on the prevention of both disruption and
overexploitation by each partner, a theme studied by Goto et al. (2010) for a moth
pollinator/seed predator mutualism that represents a possibly far more widespread
situation. If a moth, or other mutualist pollinator, lays too many eggs on or in a
single flower, the higher proportion of seeds eaten by its larvae may hamper plant
reproduction and lead to destruction of the association. Some yuccas (p. 50) selec-
tively abort flowers with a high moth egg load, and such a mechanism might help to
stabilise mutualism in two ways (Huth and Pellmyr 1999), as non-random abortion
of flowers (1) preventing moths from laying heavy egg loads and simultaneously (2)
helping to regulate moth population sizes and so reduce chances of overexploitation
in the future.
For the Epicephala-Glochidion mutualism (p. 53), the proportion of aborted
flowers increased with egg load and with higher ovule damage, and deposition of
moth eggs singly helped to avoid harmful consequences. The female moths are able
to detect flowers already containing eggs and avoid them. These complex systems,
purportedly maintained by extreme specificity, are difficult to understand. One
mechanism, investigated for the Epicephala-Breynia vitis-idaea mutualism closely
allied to the above, indicated the subtleties of specific chemical attractants to the
pollinators (Svensson et al. 2010), by identifying the floral compounds implicated
as attractants. Olfactory-based attractants to hosts were earlier confirmed for fig
wasps, and Epicephala moths, and unique chemical compounds are well known as
pollinator attractants for orchids. However, rather than the supposedly restricted and
highly specialised ‘private signal’ guidance, the active signals in this case were from
quite widespread compounds, possibly components of a broader ‘floral filter’, with
the scent only a component of this. This contrasted with the system for the dioecious
fig Ficus semicordata, in which a single compound (4-methylanisole) was the main
active attractant for the single obligate pollinator wasp, Ceratosolen gravelyi (Chen
et al. 2009), at the host stage suitable for pollination and oviposition. That com-
2.2 Stability and Integration of Mutualisms 21
pound plays a central role in how the wasp detects the host at the most suitable stage
of development. This suggested three key hypotheses relevant to understanding how
such interactions function: (1) specificity between the partners is moderated by
scent alone; (2) there is a stage-dependent attraction of the specific pollinator; and
(3) attraction occurs across male and female figs, preventing the pollinators from
distinguishing between the sexes.
The dynamic balances between ‘mutualism’ and ‘antagonism’ in interactions are
potentially a source of instability in mutualisms. The ‘natural’ (if simplistically
interpreted) tendency to select for increased fitness at the expense of a mutualist
must be countered by each partner in some way, in order to prevent overexploitation
and transformation of the interaction to effective antagonism. In examples such as
that above, the balance between successful pollination during moth oviposition, the
numbers of eggs laid and larvae developing, and the extent of seed predation by
larvae can provide a variety of outcomes. Each may favour one or the other partner
and change the balance between mutualism and antagonism. Simplistically, in the
context of this example, pollination is necessary but is also costly because a propor-
tion of the seeds is lost. A ‘cost blow-out’ could prove problematic and lead to loss
of the obligate interaction – a theme noted by Pellmyr (1997) as one possible reason
for the scarcity of obligate mutualisms. Enduring specific mutualisms are far fewer
than those in which some flexibility among partners is usual.
The stability of mutualisms may be influenced strongly by exploiters (Chap. 8),
and any mutualistic association may become the focus of attention by predators or
others, leading to changes in balance. Thus, the mutualism might be stabilised by
predators attacking other non-mutualists or destabilised by them attacking the
mutualists – so that one or both mutualists may be selected to either (1) facilitate
predator attention or (2) deter or impede predators (Schatz et al. 2008). According
to some authorities, mutualisms are predicted to be inherently unstable, because
each partner is under selection pressure to benefit from the interaction without ‘pay-
ing the costs involved’. Essentially, each partner is under pressure to take more and
pay less, with possible consequences, including transformation of one or the other
partner from mutualist to parasite, or the extinction of the mutualism.
Using ant-plant mutualisms as their models, Orona-Tamayo and Heil (2013)
reviewed the factors involved in stabilising mutualisms and pointed out the logistic
advantage of this system that, despite a given mutualism being obligate, the partners
can in practice be separated and their functions evaluated separately. Impacts of the
formation and loss of the mutualism can be simulated, and their consequences for
the ‘performance’ of each partner studied independently. A common outcome is
simply the recognition that observed changes attributed to partner influences must
be interpreted with great caution, perhaps especially to distinguish coevolved traits
from those that reflect ecological ‘fitting’ within the partnership. A further advan-
tage of studying ant-plants (Chap. 4) is that the plants are commonly closely related
to sympatric non-ant-plants, so that the two conditions of ‘myrmecophytes’ and
‘non-myrmecophytes’ can be compared to assess whether ant-plants and non-ant-
plants, and plant-ants and non-plant-ants, show adaptations or behaviours associ-
ated strictly with the mutualism but are intrinsic to the species themselves.
Orona-Tamayo and Heil pondered, for example, whether plant-associated ants are
especially adapted to defend their specific host, or whether they more simply expand
the traits used for their own protection under a range of circumstances to include the
host as well. Such questions have wide interest, and studies of ant-plant mutualisms
offer opportunities to assess evolutionary and ecological processes in these associa-
tions. Among epiphytic ant-plants (p. 84), concealed floral nectar may hinder
exploitation by generalist opportunist ants and help to stabilise the obligate mutual-
isms by restricting access to those key specialists. Several species of Rubiaceae in
Fiji offer the novel reward of sugary sap concealed in flowers. That resource is
accessible only to the specialist obligate mutualist for six of the nine species of
Squamellaria. Workers of this specialist ant species, Philidris nagasau, bite through
the thick covering epidermis, and the ants forage by tracking sucrose levels in the
post-anthetic flowers (Chomicki et al. 2016). The nectar reward was thus accessible
only to visitors capable of chewing (and not pollinating), so filters out opportunist
nectar foragers that are unable to gain that concealed sugar reward. However, unlike
some other systems, the ‘filtering’ is physical rather than based on chemical speci-
ficity, and the concealed sugar reward may have been an important component of
the evolutionary transition from facultative to obligate mutualism, by preventing
partner exploitation whilst increasing benefit trading.
Five main groups of strategies that may stabilise mutualisms against exploiters
were summarised by Orona-Tamayo and Heil (2013), with the caveat that different
authors have historically used rather different assumptions and terminology, leading
to some confusion and debate. Those strategies are (1) exclusive rewards to a spe-
cific partner such as a specialist pollinator that becomes morphologically or behav-
iourally adapted to overcome a physical barrier or to exploit a specific plant – the
Madagascan orchid Anagraecum sesquipedale and its hawk moth pollinator (p. 15)
is a commonly cited example; (2) partner choice, better known in non-insect mutu-
alisms but exemplified by ant-fungus mutualisms (p. 96) involving the selection of
suitable future partners before the association is established, so allowing host indi-
viduals to reward cooperative and non-cooperative partners differently in advance
of more regular association; (3) host sanctions, in which the host monitors activities
of the symbiont to punish exploiters, for example, by reducing the reward; (4) part-
ner fidelity feedback, inducing reduced fitness of cheaters (Weyl et al. 2010), and
operating through the very basis of a mutualism, that each partner benefits from the
interaction – so that benefits from ‘partner 1’ to ‘partner 2’ automatically feed back
to partner 1 when the amount/quality of provisions to partner 1 depends on the per-
formance of partner 2 (Connor 1986); and (5) competitive screening, proposed
more recently (Archetti et al. 2011) to entail that the host imposes some form of
‘contract’ of the appropriate costs and benefits, leading to a ‘decision’ by a prospec-
tive mutualist of whether to invest and so, in essence, causing the potential partners
to screen themselves for quality as mutualists.
Slightly differently, Chomicki et al. (2016) nominated three types of mechanisms
considered in stabilising mutualisms: (1) by-product mutualisms, (2) partner fidelity
feedback, and (3) partner choice. The first of these is when mutualistic behaviour is
‘cost-free’, in involving the by-products of other traits rather than individual adap-
2.2 Stability and Integration of Mutualisms 23
tive developments. Evolution of mutualism is thus promoted by each partner devel-

oping high ‘fidelity’, in many cases specialisation, so that feedback over extended
periods can occur and benefits accrue to both partner species. Using comparative
models, Foster and Wenseleers (2006) evaluated three aspects of benefit from such
feedbacks and suggested that partner fidelity and partner choice may be more
important than the development of genetic correlation (cooperative association),
with the last predicted to be the least important mechanism. Such evolutionary ques-
tions have received considerable attention – as examples, see essays by Hoeksema
and Bruna (2000), Borges (2015), and several contributors to Bronstein (2015), but
much of the theory is beyond the scope of this book.
‘Choosing’ a partner in a mutualism commonly involves what Archetti et al.
(2011) referred to as ‘hidden characteristics’ in which a range of potential partners
vary in their qualities, but the selecting partner has no way of appraising these in
advance and selecting ‘the best’. Partner ‘quality’ is a central issue in mutualisms
(Heath and Stinchcombe 2013), but uncertainties over how mutualisms evolve and
how they are stabilised to prevent (or select against) low-quality partners continue.
Many ant-plants, for example, associate with several ant species that differ in their
abilities to patrol the plant and protect it from herbivory, and the potential mutualist
may be committed when the queens initially colonise (Yu and Davidson 1997), after
which other ants are subsequently deterred or excluded, with ‘screening’ processes
by the plant probably influencing that mutualist. Archetti et al. (2011) discussed two
such examples: (1) Cecropia pungara and Pachycondyla luteola and (2) Acacia spp.
and Pseudomyrmex spp., both well-documented mutualisms (p. 64). Their sugges-
tions for determining these relationships both devolve on competitive advantage of
the mutualist ants over others. A very high proportion of C. pungara are inhabited
by P. luteola, and this ant establishes only with this host, so specificity is well
defined, and tree saplings can bias competition towards P. luteola by providing food
bodies early in development. The queen ants, unlike those of the other species pres-
ent, gather food during colony founding, so this may accelerate worker ant produc-
tion and enable superior competitive ability – with corresponding advantage to P.
luteola. Ant-acacias also appear to ‘choose’ partners, with those tree species offer-
ing high rewards gaining high-quality mutualists, and species with less food and
housing available inhabited by low-quality, non-protecting, species: this pattern was
discussed further by Heil et al. (2009).
Many facultative mutualisms involve arrays (‘guilds’) of interacting species,
with uneven extents and potency of influences on each potential partner. Variations
in the effectiveness of a mutualism, in consequence, may flow from different pos-
sible partners – with interactions with less beneficial partners also reducing the
frequency and opportunity for associations, so that the more beneficial species then
comprise the superior mutualists. As Miller (2007) noted, this situation poses the
problem of how generalised facultative mutualisms can persist. The costly traits
needed (e.g. by a plant in relation to mutualistic pollinators) to attract partners may
lead to diminished returns as more partners are recruited and as variation in partner
species quality increases. Studying two species of mutualist ants associated with the
cactus Opuntia imbricata in New Mexico, Miller implied that competitive interac-
tion between the ants could facilitate maintenance of the facultative mutualism or,
conversely, accelerate its decline, through varying effects on the benefits and cost-
benefit balance for the nectary-bearing host plant. In some cases, having multiple
partners may reduce the benefits of facultative mutualisms, but despite competition
within a mutualist guild possibly being important (as predicted by Stanton 2003),
more experimental studies are needed to explore whether this is a general scenario.
The balance of benefits in a mutualism can often be far more complex than ini-
tially apparent. The myrmecophyte Cordia nodosa (Boraginaceae) in Peru hosts the
ant parasite Allomerus octoarticulatus as the most widespread of several ants that
can co-occur with this plant. Other potential mutualists on the same plant are Azteca
spp. and Myrmelachista schumanni (Frederickson and Gordon 2009), and all tend
scale insects within domatia. The association is based on a combination of provid-
ing domatia and food bodies (Chap. 4) and receiving protection from marauding
herbivores, with honeydew from scale insects (p. 106) providing additional food.
Manipulative experiments (Frederickson et al. 2012) showed that ants increased
plant growth when herbivores were present, but decreased it when herbivores were
absent – so that the habit of hosting ants could be deemed costly to the plants – but
no cost to the ant colonies was detected from their defending the plants against her-
bivores. Over a year, when herbivores were absent, plants with ants were shorter (by
18%) and had fewer leaves (36% fewer) and domatia (40% fewer) than plants with-
out ants. Plants without ants, or with only the parasitic Allomerus, reached a ‘size
threshold’ and then tended to become smaller. The alternative patterns, suggested
by ants occupying new domatia as fast as these became available, were (1) ant col-
ony growth keeps pace with plant growth, or (2) the number of worker ants in each
domatium declines as the plant grows, and the system on Cordia was paralleled by
that on Duroia hirsuta (Rubiaceae) (Frederickson and Gordon 2009). The ants for-
age only on the host, with the food resources available there being (1) the food bod-
ies formed on new foliage and shoots; (2) honeydew from scale insects tended
within domatia, and preying directly on the scale insects; and (3) the arthropods
visiting C. nodosa. As summarised in Fig. 2.2, these components have different
emphases; ‘1’ is a direct cost to the plant, ‘2’ is an ecological cost, and ‘3’ is a ben-
efit. Frederickson et al. (2012) interpreted the slower growth of Cordia with ants
than without them as possibly due to direct costs of providing food bodies for the
ants or from the ecological cost of losing resources to the scale insects. It is possible
that the plant defence might here be simply a by-product from Allomerus foraging
for insect prey on the plants.
Several reviewers have assembled the very diverse information published on dif-
ferent mutualistic interactions to try to indicate the costs and benefits and the eco-
logical importance of those associations – many of which are outlined and discussed
in the following chapters. An informative synthesis appraising the mutualistic inter-
actions of ants with aphids, scale insects, tree hoppers, and lycaenid butterflies led
to the spectrum summarised in Table 2.1 (Stadler et al. 2001), which displays the
costs or benefits to the participating ants. Not all effects had then been demonstrated
for all the mutualistic groups, but it seems likely that the guiding principles
involved – and often shared across these groups – may be widespread. Those con-
measured
plant performance
domatia plant
leaves
height
+/_ _? + measured/manipulated
ants
_
_ +?
folivores measured
+
scale insects
Fig. 2.2 Interactions to indicate the principal cost and benefit pathways between colonies of the
myrmecophytic ant Allomerus octoarticulatus and plants of Cordia nodosa, and showing the
manipulations undertaken and variables measured. Cordia harbours ants (which reduce herbivory
but facilitate scale insects) and scale insects in domatia. Direct costs to plants include production
of reward (domatia, food bodies) for ants; ecological costs include losses of resources to ant-
tended scale insects. Solid lines are direct interactions, broken lines are indirect interactions (Based
on Frederickson et al. 2012)
Table 2.1 Major benefits and costs from mutualistic interactions between ants and selected other
insect groups (Stadler et al. 2001)
Major costs for ant partners
Physiological costs (direct costs)
Longer developmental time/lower larval weight
Lower pupal weight
Higher mortality risks due to specialised predators/parasitoids in the presence of ants (direct
costs)
Chemical camouflage, protection, slow movements of natural enemy
High rates of parasitisation possible
Costs in the absence of ants (indirect costs)
Obligate myrmecophiles unable to survive without ants
Higher risk of attack by natural enemies due to aggregation or conspicuousness
Major benefits for ant partners
Protection against natural enemies (direct benefit)
Effects on development (direct benefits)
Increased food uptake, higher rates of excretion/turnover
Higher pupal weight
Increased fecundity
Larger population size, as indirect evidence for faster development and higher fecundity
Distribution is uneven across the four taxa assessed (Aphididae, Coccidae, Membracidae,
Lycaenidae) as discussed further in Chaps. 5 and 6, but listing the topics here demonstrates the
considerable variety of factors involved
High
Fitness of myrmecophiles
(benefits>costs)
With ants
Without ants
Low
(costs>benefits)
Low High
Mortality risk of myrmecophiles
Fig. 2.3 Conceptual framework that incorporates some of the major costs and benefits for myr-
mecophiles of associating with ants. ‘Fitness’ and ‘mortality risk’ are schematically shown (From
Stadler et al. 2001)
texts led to a framework (Fig. 2.3) in which fitness consequences and mortality risks
are depicted, and shows the transitions between entering a mutualistic association
with ants and remaining unencumbered. If the mortality risks to myrmecophiles
from natural enemies, weather, or disease are high, association with ants increases
fitness. Conversely, if risks are low, greater fitness may accrue without a
mutualism.
2.3 Wider Community Stability
Several of the preceding examples of mutualisms are based on the exchange of a

‘service’ (such as pollination, dispersal, protection from attack) for which a ‘reward’
(such as nectar, other nutrition, a place to live) is given. Any such exchange may
have a long history of evolution and also occur within the context of numerous other
interactions from a variety of species, whose incidence and impacts vary continu-
ously within the local community. The central importance of competition in under-
standing the ecology of mutualisms has been emphasised repeatedly, so that
competition ‘lies at the heart of selection to attract, retain, and benefit from mutual-
isms’ (Jones et al. 2012). Competition for mutualists is complex and may occur
between conspecific partners and amongst the variety of mutualist and exploiter
(Chap. 9) partners, in either case associated with hierarchies among partners that
affect the outcomes of the mutualism within the local context and spectrum of taxa
involved. The scenarios summarised in Fig. 2.4 illustrate the principles involved,
and developed further by Jones et al. (2012) to highlight the importance of both
intraspecific and interspecific competitions in the dynamics of mutualisms.
As several of the foregoing cases show, opposing views on stability within mutu-
alisms are voiced for various plant-pollinator networks (Burkle et al. 2013).
2.3 Wider Community Stability 27
a
+
M1 M2
+
b
R1
M1 M2
R2
M1a R1
M2
M1b R2
d
M1a R1
M2
M1b R2
Fig. 2.4 Aspects of competition for resources among mutualists: (a) the net reciprocal benefit (+)
between two partners (M1, M2), comprising the basic presumption of a mutualism; (b) a resource-
based expression of this, where M1, M2 produce resources (R1, R2) respectively and each resource
is consumed by the non-producing species; arrows show both resource production and (broken
lines) consumption; (c) mutualism in which lower-quality mutualists have a competitive advan-
tage; M1 is now treated as two individuals or species from a mutualist guild (M1a, M1b). M1a is
the better mutualist as it offers a large amount of resources (thick arrow to R1); however, M1b is
the better competitor for resources produced by M2 (shown by a thick arrow from R2-M1b), with
advantage coming from interference competition (dashed arrow from M1b to M1a’s consumption
arrow from R2); (d) mutualism when high-quality mutualists have a competitive advantage: as in
‘c’, M1 is the better mutualist in producing large amounts of resource R1, but is now also a better
competitor for R2 (thick arrow from R2-M1a), resulting from partner control by M2 (dashed inhi-
bition arrow from M2 to M1b’s consumption arrow) (as discussed and later expanded, by Jones
et al. 2012)
Opinions from digests of published studies range from (1) their being especially
susceptible to human-induced changes because of their sensitivity to seasonal varia-
tions, behaviour and specificity, and relative abundance of multiple species
(Tylianakis et al. 2008) to, conversely, (2) their being especially robust and resistant
to disturbances because of lack of redundancy present (Bascompte et al. 2003).
Mutualisms may benefit communities directly by increasing stability and persis-
tence (Ringel et al. 1996), but direct field investigations to test that hypothesis are
both few and difficult to undertake. One such is a trial undertaken to determine
whether the mutualism between trophobiont insects and ants (Chap. 6) influences
the species richness and abundance of the associated local arthropod community (de
Freitas and Rossi 2015). The hemiparasitic mistletoe Psittacanthus robustus
(Loranthaceae) in Brazil hosts several trophobiont Hemiptera that feed on sap and
egest honeydew, which is sought by two tending mutualistic ant species (Camponotus
rufipes, Cephalotus pusillus). Experimental treatments using the host tree (Vochysia
thyrsoidea, Vochysiaceae) in an area of about 30 ha comprised trees (1) from which
ants and trophobionts were excluded; (2) from which ants only were excluded; and
(3) on which both ants and trophobionts were present, using the mutualism as a
‘control’. Exclusion treatments were achieved through using a commercial barrier
resin around the tree trunks and direct manual removal of all arthropods afterwards,
so restricting access to only winged insects – which were removed at intervals dur-
ing the 3-month survey. On five occasions during that period, all arthropods were
collected from the trees, by direct aspiration followed by beating. The local abun-
dance and richness of the community (the sampling yielded representatives of nine
orders, 31 families, and 39 species) (Fig. 2.5) was not affected in any major way, but
analysis by implied feeding roles showed that the mutualism had a negative impact
on predators, although not on herbivores, omnivores, or parasitoids. The most
diverse predators (Coccinellidae, five species) were all classed as generalists, and
their lower abundance on plants with the mutualism was attributed to the protection
of trophobionts by the ants, so augmenting the instability of the system. Conservation
of mutualistic insects may be important in maintaining high diversity in enveloping
landscapes.
Designating the roles of additional species that adopt an existing mutualism can
become complex. The gall midge Rabdophaga salicisbrassicoides (Cecidomyiidae)
produces galls, sometimes abundantly, on shoots of Salix exigua in North America.
Those same stems are commonly occupied also by ant-tended aphids which may
also feed on the gall tissue (Savage and Peterson 2007). Effects of the presence of
ants and aphids on gall numbers, assessed by manipulative experiments and direct
observation, confirmed some of the complexities involved. Densities of ants, aphids,
and midge galls were all positively correlated with each other, and aphid popula-
tions were highest when both ants and midges were also present. Similarly, gall
density and numbers of midge larvae/stem were the highest on stems with both ants
and aphids present. Mechanisms for this were not wholly clear, but this study
showed that herbivores feeding on the same host may indirectly facilitate each oth-
er’s well-being through positive feedback involving mutualisms.
2.3 Wider Community Stability 29
a
6
Species richness
5
4
3
2
1
40
Total abundance
b
30
20
10
5
c
Relative abundance
4
3
2
1
Mutualism Only ants Total

removed exclusion
Treatment
Fig. 2.5 Mutualism between ants and trophobiont insects: a summary of some wider community
effects based on the study of associations between the hemiparasitic mistletoe Psittacanthus robus-
tus, its trophobiont hemipterans, and attending Camponotus ants in Brazil. Arthropods from five
censuses after start of treatments are compared by (a) species richness, (b) total abundance, and (c)
relative abundance compared with the surrounding arthropod community (After de Freitas and
Rossi 2015)
A major problem in predicting outcomes of mutualisms, emphasised by Ringel

et al. (1996), is that the two-species systems to which almost all investigations have
been largely confined are a severe simplification from ‘the real world’. A four-
species model of a pollination mutualism involving two plants (among which intra-
specific competition for resources occurs in both taxa), a single pollinator (a
generalist regulated by predation), and a single top predator (that has other sources
of food) was used to more closely emulate natural conditions. The consequences
were analysed for three community structures: (1) mutualisms between the pollina-
tor and both plants; (2) one with the plants lacking a mutualism, so that the pollina-
tor is a nectar thief; and (3) one with both pollination and nectar theft. Even the
apparently slight difference in possible interactions by increasing the number of
participating species from two to four had considerable impacts on predictions from
the simpler scenario, but implied strongly that mutualisms may be stabilising and so
enhance chances of community persistence. Richer communities may indeed be

more stable, with their mutualisms more stabilising than in a simple two-species
assembly.
More broadly, recognition that positive interactions such as mutualisms are
important in studying and understanding ecological communities (Stachowicz
2001) has led a transition from considering them largely as curiosities to appreciat-
ing their roles in affecting distribution and abundance of numerous species and
enhancing community stability. Without understanding the mechanisms that form
and sustain positive interactions, ‘it is unlikely that we will be able to develop effec-
tive strategies for ecosystem management to meet the challenges posed by global
change’ (Stachowicz 2001).
References
Anderson B, Johnson SD (2008) The geographical mosaic of coevolution in a plant-pollinator

mutualism. Evolution 62:220–225
Archetti M, Ubeda F, Fudenberg D, Green J, Pierce NE, Yu DW (2011) Let the right one in: a
microeconomic approach to partner choice in mutualisms. Am Nat 177:75–85
Arditti J, Elliott J, Kitching IJ, Wasserthal LT (2012) ‘Good heavens what insect can suck it’ –
Charles Darwin, Angraecum sesquipedale and Xanthopan morganii praedicta. Bot J Linn Soc
169:403–432
Bascompte J, Jordano P, Melian CJ, Olesen JM (2003) The nested assembly of plant-animal mutu-
alistic networks. Proc Natl Acad Sci 100:9383–9387
Bauder JA-S, Lieskonig NR, Krenn HW (2011) The extremely long-tongued Neotropical butterfly
Eurybia lycisca (Riodinidae): proboscis morphology and flower handling. Arthropod Struct
Develop 40:122–127
Bauder JA-S, Warren AD, Krenn HW (2015) The ecological role of extremely long-proboscid
Neotropical butterflies (Lepidoptera: Hesperiidae) in plant-pollinator networks. Arthropod-
Plant Interact 9:415–424
Borges RM (2015) How mutualisms between plants and insects are stabilized. Curr Sci
108:1862–1868
Bronstein JL (ed) (2015) Mutualism. Oxford University Press, Oxford
Burkle LA, Marlin JC, Knight TM (2013) Plant-pollinator interactions over 120 years: loss of spe-
cies, co-occurrence, and function. Science 339:1611–1615
Chen C, Song Q, Proffit M, Bessiere J-M, Li Z, Hossaert-McKey M (2009) Private channel: a
single unusual compound assures specific pollinator attraction in Ficus semicordata. Funct
Ecol 232:941–950
Chomicki G, Staedler YM, Schonenberger J, Renner SS (2016) Partner choice through concealed
floral sugar rewards evolved with the specialization of ant-plant mutualisms. New Phytol
211(4):1358–1370. doi:10.1111/nph.13990
Colwell DD, Otranto D, Stevens JR (2009) Oestrid flies: eradication and extinction versus biodi-
versity. Trends Parasitol 25:500–504
Colwell RK, Dunn RR, Harris NC (2012) Coextinction and persistence of dependent species in a
changing world. Annu Rev Ecol Evol Syst 43:183–203
Connor RC (1986) Pseudo-reciprocity: investing in mutualism. Anim Behav 34:1562–1584
Darwin CR (1862) The various contrivances by which British and foreign orchids are fertilised by
insects, and on the good effects of intercrossing. John Murray, London. (later printing
consulted)
References 31
de Freitas JD, Rossi MN (2015) Interaction between trophobiont insects and ants: the effect of
mutualism on the associated arthropod community. J Insect Conserv 19:627–638
Dunn RR, Harris NC, Colwell RK, Koh LP, Sodhi NS (2009) The sixth mass coextinction: are
most endangered species parasites and mutualists? Proc R Soc Lond B 276:3037–3045
Foster KR, Wenseleers T (2006) A general model for the evolution of mutualisms. J Europ Soc
Evol Biol 19:1283–1293. doi:10.1111/j.1420-9101.2005.01703.x
Frederickson ME (2013) Rethinking mutualism stability: cheaters and the evolution of sanctions.
Q Rev Biol 88:269–295
Frederickson ME, Gordon DM (2009) The intertwined population biology of two Amazonian myr-
mecophytes and their symbiotic ants. Ecology 90:1595–1607
Frederickson ME, Ravenscroft A, Miller GSA, Hernandez LMA, Booth G, Pierce NE (2012) The
direct and ecological costs of an ant-plant symbiosis. Am Nat 179:768–778
Goto R, Okamoto T, Kiers ET, Kawakita A, Kato M (2010) Selective flower abortion maintains
moth cooperation in a newly discovered pollination mutualism. Ecol Lett 13:321–329
Heath KD, Stinchcombe JR (2013) Explaining mutualism variation: a new evolutionary paradox?
Evolution 68:309–317
Heil M, Gonzalez-Teuber M, Clement LW, Kautz S, Verhaagh M, Bueno JCS (2009) Divergent
investment strategies of Acacia myrmecophytes and the coexistence of mutualists and exploit-
ers. Proc Natl Acad Sci 106:18091–18096
Hoeksema JD, Bruna EM (2000) Pursuing the big questions about interspecific mutualism: a
review of theoretical approaches. Oecologia 125:321–330
Huth CJ, Pellmyr O (1999) Yucca moth oviposition and pollination behavior is affected by past
flower visitors: evidence for a host-marking pheromone. Oecologia 119:593–599
Jones EI, Bronstein JL, Ferriere R (2012) The fundamental role of competition in the ecology and
evolution of mutualisms. Ann N Y Acad Sci 1256:66–88
Kiers ET, Palmer TM, Ives AR, Bruno JF, Bronstein JL (2010) Mutualisms in a changing world:
an evolutionary perspective. Ecol Lett 13:1459–1474
Koh LP, Sodhi NS, Brook BW (2004) Co-extinctions of tropical butterflies and their hostplants.
Biotropica 36:272–274
Maloof JE, Inouye DW (2000) Are nectar robbers cheaters or mutualists? Ecology 81:2651–2661
Miller TEX (2007) Does having multiple partners weaken the benefits of facultative mutualism? A
test with cacti and cacti-tending ants. Oikos 116:500–512
Moir ML, Vesk PA, Brennan KEC, Keith DA, Hughes L, McCarthy MA (2010) Current constraints
and future directions in estimating coextinction. Conserv Biol 24:682–690
Moir ML, Vesk PA, Brennan KEC, Keith DA, McCarthy MA, Hughes L (2011) Identifying and
managing threatened invertebrates through assessment of coextinction risk. Conserv Biol
25:787–796
Orona-Tamayo D, Heil M (2013) Stabilizing mutualisms threatened by exploiters: new insights
from ant-plant research. Biotropica 45:654–665
Pellmyr O (1997) Pollinating seed eaters: why is active pollination so rare? Ecology
78:1655–1660
Powell FA (2011) Can early loss of affiliates explain the coextinction paradox? An example from
Acacia-inhabiting psyllids (Hemiptera: Psylloidea). Biodivers Conserv 20:1533–1544
Ringel MS, Hu HH, Anderson G (1996) The stability and persistence of mutualisms embedded in
community interactions. Theoret Popul Biol 50:281–297
Sachs JL, Simms EL (2006) Pathways to mutualism breakdown. Trends Ecol Evol 21:585–592
Savage AM, Peterson MA (2007) Mutualism in a community context: the positive feedback
between an ant-aphid mutualism and a gall-making midge. Oecologia 151:280–291
Schatz B, Kjellberg F, Nyawa S, Hossaert-McKey M (2008) Fig wasps: a staple food for ants on
Ficus. Biotropica 40:190–195
Stachowicz JJ (2001) Mutualism, facilitation, and the structure of ecological communities.
Bioscience 51:235–246
Stadler B, Fiedler K, Kawecki TJ, Weisser WW (2001) Costs and benefits for phytophagous myr-
mecophiles: when ants are not always available. Oikos 92:467–478
Stanton ML (2003) Interacting guilds: moving beyond the pairwise perspective on mutualisms.
Am Nat 162:S10–S23
Svensson GP, Okamoto T, Kawakita A, Goto R, Kato M (2010) Chemical ecology of obligate pol-
lination mutualisms: testing the ‘private channel’ hypothesis in the Breynia-Epicephala asso-
ciation. New Phytol 186:995–1004
Tylianakis JM, Didham RK, Bascompte J, Wardle DA (2008) Global change and species interac-
tions in terrestrial ecosystems. Ecol Lett 11:1351–1363
Ueda S, Okubo T, Itioka T, Shimizu-kaya U, Yago M, Inui Y, Itino T (2012a) Timing of butterfly
parasitization of a plant-ant-scale symbiosis. Ecol Res 27:437–443
Ueda S, Quek S-P, Itioka T, Onamori K, Sato Y, Murase K, Itino T (2012b) An ancient tripartite
symbiosis of plants, ants and scale insects. Proc R Soc Lond B 275:2319–2326
Wallace AR (1867) Creation by law. Q J Sci 4:471–488
Weyl EG, Frederickson ME, Yu DW, Pierce NE (2010) Economic contract theory tests models of
mutualism. Proc Natl Acad Sci 107:15712–15716
Yen AL (1980) The taxonomy and comparative ecology of selected psyllids (Insecta: Hemiptera:
Psylloidea) on Acacia species (Mimosaceae). PhD thesis, La Trobe University, Melbourne
Yu DW, Davidson DW (1997) Experimental studies of species-specificity in Cecropia-ant relation-
ships. Ecol Monogr 67:273–294
Part II
Lessons from Classic Mutualisms
The structures and characteristics of many of the predominant and most intensively
studied mutualistic interactions that involve insects are summarised in this part.
They involve a variety of different taxa, but all are terrestrial, and detection or study
of mutualisms involving aquatic insects is scarce. Many regular associations involv-
ing, for example, specialised ways of living by insects inhabiting pitcher plants and
other organism-based phytotelmata are undoubtedly quite specific (Kitching 2000)
but their functional interdependence remains largely obscure. They have contrib-
uted relatively little to the wider understanding of mutualisms.
Indeed, the variety of mutualisms explored in detail predominantly involve only
a few groups of insects – these are categorised in various ways, however. Functional
emphasis, such as pollination (Chap. 4), encompasses a wide taxonomic range of
insects with varied roles as vectors of plant reproductive products and the benefits
they gain and confer. Stronger taxonomic focus (Chaps. 5, 6 and 7) emphasises the
dominant participation of ants and the complexity of their mutualisms and related
associations with plants, fungi and numerous other insects.
Understanding the biology of many ants, as well as of notable groups of
Hemiptera, Lepidoptera and others, whose taxonomic and ecological variety have
generated a considerable array of contexts that collectively aid development of
mutualistic associations, is a central need, to which numerous naturalists and others
have contributed observations and thoughts. In many cases, that understanding must
draw also on other interactions between the participants – competition, herbivory,
predation, parasitisation may each affect local evolution, developments and out-
comes of mutualisms, and interpreting their roles is also a key need. Comparisons
of intricate mutualisms with parallel interactions between taxonomic relatives can
indicate patterns of transition and significance of the gradations between indepen-
dence and dependence.
The three predominant and most comprehensively studied participant orders
(Hymenoptera, Hemiptera, Lepidoptera) each contain a spectrum of species illus-
trating such transitions, so that independent studies across a range of taxa affords
sets of individually derived changes for which comparison may lead to more general
34 II Lessons from Classic Mutualisms
conclusions (Stadler et al. 2001, p. 25). Other orders have also contributed – the
parallels between social termites and ants, and interactions between beetles and
fungi, extend the scopes of mutualisms beyond those evident from focus on ants
alone. Nevertheless, studies of ants as ecologically pervasive mutualists have facili-
tated understanding of many other ecological contexts, as well as mechanisms of
plant defences against herbivory, determinants of food web structures and the
dynamics of many interspecific interactions. Ants have been described as mutualists
‘par excellence’, and lend themselves well to manipulative experimental studies
(Ness et al. 2010). Using ants as an example, it may be possible to identify ecologi-
cal factors that are associated with mutualisms. Oliver et al. (2008) found, for exam-
ple, that those ants most likely to tend Hemiptera for honeydew as trophobionts
(Chap. 5) are polygynous, ecologically dominant within their local community,
tree-nesting and with large colonies, and also able to occupy disturbed habitats.
Ness et al. considered that ant mutualisms have several features that render them
‘model systems’ in elucidating the structure and roles of their various associations,
including involvements in plant defence. Three main attributes facilitate their prac-
tical use, as (1) their abundance and behaviour can be monitored easily, so that
changes in numbers or performance can be described and compared reliably: the
costs and benefits of ants within a mutualism, and changes with time, manipulation
and across sites or treatments, can be assessed; (2) ants can be excluded from
particular microsites, such as plants or plant parts, by physical barriers, and the
consequences of exclusion or removal assessed; and (3) many individual mutualism
partners (such as myrmecophytic plants or long-lived myrmecophilous butterfly lar-
vae) interact with the same ant colony over extended periods, of weeks to years –
indeed, Frederickson et al. (2005) implied that some ant mutualisms may have
functioned continually for centuries! Specifically, they noted that Myrmelachista
schumanni, a participant in the ‘devil’s gardens’ (p. 86) with the tree Duroia hirsuta,
colonies can live for 800 years, through killing plants of other species in the local
area. Extrapolating from growth rates of the gardens, one garden with 351 plants
was estimated to be 807 years old (95% confidence interval 446–4334 years). Up to
15,000 queen ants may be present in such a colony.
Any such sustained association contrasts with, for example, mutualisms that
involve more mobile partners such as many flower-visiting pollinators. Most insect-
plant associations, however, are reestablished more frequently, perhaps every gen-
eration as ‘horizontal transmission’, and this is so for both obligate and facultative
mutualisms. This frequency of reassociation provides opportunities for exploitation,
and may render the mutualisms vulnerable or unstable.
The major omission from this account is of the intricate mutualisms between
insects and protozoa and microbes, other than for a simple summary of some insect-
fungus interactions (Chap. 7). The widespread and obligate nature of mutualisms
between termites and cockroaches and highly specialised hind gut ‘fauna’ has for
long been a ‘textbook example’ of mutualism. Those gut-inhabiting organisms are
transmitted largely though faecal ingestion, in some cases by active trophallaxis,
and are essential in breaking down cellulose and ensuring adequate nutrition for the
insects. However, little or no attention has been paid to conservation of those asso-
II Lessons from Classic Mutualisms 35
ciations, other than occasional allusions to pesticide impacts. A parallel system, of

scarab beetle larvae with specialised microbial gut faunas, has been even less well
studied – but also considered a mutualism, with the gut organisms critical in facili-
tating digestion of decaying material and enabling the massive ecological variation
amongst these beetles. Dead organic material, most notably wood, can bring con-
sumers into contact with the numerous microorganisms that live there, many of
them adapted to digest decaying organic matter. As Price et al. (2011) surmised,
acquisition of such microorganisms by insects as mutualists might facilitate their
exploitation of these abundant food sources. Management of dead wood is a well-
defined theme in conservation of numerous saproxylic insects, including many
beetles of current conservation concern (New 2010). The roles of microbial mutual-
ists in those contexts are still to be explored constructively.
References
Frederickson ME, Greene MJ, Gordon DM (2005) ‘Devil’s gardens’ bedeviled by ants. Nature
437:495–496
Kitching RL (2000) Food webs and container habitats. The natural history and ecology of phyto-
telmata. Cambridge University Press, Cambridge
Ness J, Mooney K, Lach L (2010) Ants as mutualists. In: Lach L, Parr C, Abbott K (eds) Ant ecol-
ogy. Oxford University Press, Oxford, pp 97–114
New TR (2010) Beetles in conservation. Wiley-Blackwell, Oxford
Oliver TH, Leather SR, Cook JM (2008) Macroevolutionary patterns in the origin of mutualisms
involving ants. J Ecol Biol 21:1597–1608
Price PW, Denno RF, Eubanks MD, Finke DL, Kaplan I (2011) Insect ecology. Behavior, popula-
tions and communities. Cambridge University Press, Cambridge
Stadler B, Fiedler K, Kawecki TJ, Weisser WW (2001) Costs and benefits for phytophagous myr-
mecophiles: when ants are not always available. Oikos 92:467–478
Chapter 3
Classic Themes: Pollination Mutualisms
of Insects and Plants
Keywords Coadaptation • Ecological specialization • Evolution • Figs •

Fig-wasps • Flower visitors • Nursery pollination • Yuccas • Yucca moths • Diptera
• Hymenoptera • Lepidoptera
3.1 Introduction: Pollination Systems
Pollination involves interactions between two predominant groups of organisms:

the flowering plants and the vectors of their gametes – and amongst the great variety
of animal vectors, insects are the most numerous. Such interactions broadly com-
prise one of the most varied and widespread of all mutualistic relationships. Those
interactions, however, vary greatly in specificity and obligation. Interactions within
plant-pollinator associations have for long dominated studies on mutualisms, with
those studies including substantial contributions from field and mathematical mod-
elling studies leading to increased understanding of the evolution and dynamics of
mutualisms. The vast literature on pollination biology – building on some 250 years
of serious observations, documentation and interest – includes many essays and
reviews discussing mutualisms and wider associations between the vectors and
plants and the pathways by which such associations have developed and are sus-
tained. The interactions can often be compartmentalised only uncertainly (Waser
and Ollerton 2006).
The intricacies of coadaptations between flowering plants and their insect or
other pollen vectors reflect many highly specialised coevolutionary systems, with
the widespread underlying mutualistic benefits of nutritional reward (nectar, pollen
or other plant product) to the vector and pollen dispersal (increased reproduction
and possibly expanded distribution) to the plant. Many such associations are highly
specific – and many others more generalised – depending, respectively, on single
pollinator species, a few taxonomically or functionally related species or broader
groups such as ‘bees’, ‘butterflies’ or ‘flies’. Each major insect group is associated
with flowers that have particular suites of attractant features such as shape, colour,
scent and reward.
Pollination systems are thus very diverse and incorporate numerous evolutionary
and ecological strategies and ruses to enhance chances of success. Most, however,

DOI 10.1007/978-3-319-58292-4_3
38 3 Classic Themes: Pollination Mutualisms of Insects and Plants
involve transient visits by the pollinator with no more enduring interdependence

beyond the exchange of nutrients for pollination services. Lists of ‘floral visitors’,
compiled for numerous plants, are valuable aids to the initial assessment of pollina-
tor arrays and may indicate regular associations that could be, or foreshadow, mutu-
alisms. Cycads, for example, are commonly reported to be wind-pollinated, but
insect vectors also play a considerable part. Australian Macrozamia spp. are polli-
nated by weevils (Tranes sp.) and thrips (Cycadothrips chadwicki) (Terry et al.
2005), as passive pollinators in dependent mutualisms. Both these insects have spe-
cialised behaviour that facilitates transfer of pollen between the cycad cones. In
marked contrast to some of the associations discussed below, the range of possible
cycad pollinators is taxonomically broad, and the roles of many beetles and others
associated with the flowers are still very unclear.
Despite specialised pollination mutualisms implying coevolution linked with
interdependence and the partners adapting in parallel, the partners may indeed show
considerable independence. The trio of species comprising yellow loosestrife
(Lysimachia vulgaris, Myrsinaceae) and its two specialised oil-collecting bee pol-
linators (the sister species Macropis europaea and M. fulvipes, Melittidae) in Europe
demonstrates some of the complexities (Triponez et al. 2015). A study of their long-
term phylogeographic history showed that the two bees and plant had different
migratory capabilities, the two bees had different ecological roles, and spatial distri-
bution of their genetic variation differed in the three species. Their evolutionary
pathways had been largely independent. Different patterns of gene flow occur com-
monly amongst mutualists and are likely to influence coevolution, with the promo-
tion of local differences in mutualism strength through affecting preferences and
partner choice (Anderson et al. 2004).
Most pollination mutualisms are indeed very generalised, with (1) the plant spe-
cies pollinated by a variety of vectors, commonly of different insect taxa, and (2)
most pollinators able to gain rewards from a variety of plant species, as associations
that assume immense economic importance in crop production and other contexts in
which efficient pollination is necessary for human welfare. Many insects are thus
involved, many of them only casually or irregularly, as pollen vectors and as very
transient partners in mutualisms that have no regular or predictable basis. Three
large insect orders dominate implications and discussions of more regular mutualis-
tic interactions, some having well-defined and intensively studied roles in pollina-
tion. These ‘big three’ are Hymenoptera (notably bees), Diptera and Lepidoptera,
members of which demonstrate highly intricate and specific relationships as polli-
nators with particular flowering species, many with implications of mutualism.
Somewhat anomalously, ants are only rarely specialist pollinators, although they
participate in numerous other mutualisms with plants (Chap. 4).
The gradations between specialists and generalists within a continuous range of
pollination associations were illustrated schematically for pollinating bees and
flowering plants as in Fig. 3.1 (Minckley and Roulston 2006). There, bee species
1–5 are all specialists in pollinating a single plant species, but only species 1 has a
3.1 Introduction: Pollination Systems 39
Fig. 3.1 Scheme to

indicate the variety of
pollination options that can
occur amongst bees and
plants. Seven bee species
(numbers 1–7) and six
plant species (letters A–F)
are shown to illustrate the
variety of combinations
(linking lines). Species in
bold are specialised
participants; others are
generalists that interact
with more than one partner
(Minckley and Roulston
2006)
one-to-one interaction with a plant. Bee species 2–4 interact with each other and
may need to compete for the same plant, and bee 5 interacts with generalist bees on
the single plant it exploits. Bee species 6 and 7 are generalist pollinators that can
interact either with generalist or specialist plant species. Solitary bees show many
ecological specialisations, but host specificity for pollen (monolecty or oligolecty)
is the most frequent and widespread of these. That scheme was depicted in a slightly
different format that emphasised species richness in each cohort (Steffan-Dewenter
et al. 2006) and with the addition of an ‘all species’ interaction that should, except
in the most catastrophic circumstances, be resilient to losses of some important pol-
linators and some key host plants.
There is wide appreciation that the continued functioning of most terrestrial eco-
systems depends on the continuity and conservation of pollinator-plant associations
and wider networks. Recent declines of pollinators, posing formidable ecological
and economic problems in many parts of the world, have emphasised losses of bees,
but other insects are easily implicated as well. The great diversity of flies as flower
visitors (Larson et al. 2001) implies that their capability for pollination has arisen
independently many times within the order. Their variety of foraging strategies
(Inouye et al. 2015) and their responses to floral features and rewards (Woodcock
et al. 2014) imply similar arrays of behavioural responses and outcomes to other,
better known, pollinators.
Nevertheless, as for any of these insect groups that form regular or frequent asso-
ciations, some attended plants exhibit specialised features that strengthen or spe-
cialise particular linkages – examples discussed below include the Globe-flower
pollinating fly genus Trollius (Anthomyiidae) (p. 54) and long-tongued
Nemestrinidae (p. 58). Inouye et al. (2015) also noted the various sapromyophilous
flowers that mimic carrion or dung and attract particular groups of flies needing
those resources. These plants include one of the world’s largest flowers, Rafflesia,
from the Malaysian/Indonesian region (Beaman et al. 1988). Pollination of R. pricei
(endemic to part of Sabah) is carried out by carrion flies (Lucilia and Chrysomya,
both in the family Calliphoridae). Both genera were observed visiting a male flower,
but only Lucilia papuensis was seen on a female flower. As in other sapromyophi-
lous taxa, visual and olfactory attractants to the short-lived flowers are vital.
‘Deceitful flowers’ mimicking possible mates, oviposition substrates or other
resources needed by particular flies are also common – plants in 14 families have
such features (Woodcock et al. 2014), but pollinator specificity for most of these is
unknown.
Unlike many pollinating Hymenoptera, which seek resources to provision their
offspring away from the plants they visit, many flower-visiting flies also seek breed-
ing sites and may oviposit naturally on or in substrates provided by the plants. One
consequence, because the flies do not have to provision brood from their foraging
activities, is that they may be able to depend on less rewarding flowers than many
Hymenoptera. It has been claimed that Diptera may have influenced early angio-
sperm radiation through pollination activities (Ssymank et al. 2008), with modern
flies visiting numerous flowers as either generalists or more restricted specialists –
they cited, for example, references to flower flies (or hover flies, Syrphidae) visiting
flowers of 94 plant families. In much commercial crop production, flies are viewed
as second only to Hymenoptera in importance.
Many adult Lepidoptera feed on nectar, and diurnal butterflies are associated
commonly with colourful flowers – the burgeoning literature on butterfly gardening
as an increasingly common conservation practice to increase butterfly abundance
and richness in parks and other open spaces, many in urban areas, contains many
examples of ‘butterfly flowers’ that are especially attractive through colour or scent
characteristics and lists of different species that may be combined to extend sea-
sonal flowering to provide nectar for the combined flight season of local butterflies.
Some butterfly pollination is significant. The spectacular South African mountain
pride butterfly (Aeropetes tulbaghia, Nymphalidae) is regarded as a keystone spe-
cies, as the only known pollinator of about 20 plant species. Ball (2012) regarded its
diurnal associations in pollination of large red or orange flowers as an example of
convergent evolution.
In contrast, many moths are nocturnal and are collectively regarded as the most
significant group of nocturnal insect pollinators. Their activity largely avoids coin-
cidence with most day-active bees, flies and butterflies, and, rather than being
brightly coloured, many ‘moth flowers’ are white and rely on strong scents to attract
vectors. Hawk moths (Sphingidae), some of which are diurnal are sometimes
viewed as a separate functional category of pollinators. Their fast flight and hover-
ing behaviour contrasts with that of most other moths that must land on flowers in
order to feed and demands a supply of sugar-rich nectar. Sugar concentration alone
may affect a moth’s preference – as in the relative preferences of bee hawk moths
(Hemaris spp.) for different Asteraceae in New Jersey. The moths appeared to prefer
the thistle Cirsium arvense over invasive Centaurea spp., visiting significantly more
inflorescences of the former (Tartaglia and Handel 2014), which also had nectar
with significantly higher sugar concentration.
Plants pollinated by hawk moths are viewed commonly as specialised, through
features such as nocturnal opening and deep corolla tubes effectively preventing
access by many other insects. However, some such pollinators appear to be rare.
Observations on the highly localised canopy tree Oxyanthus pyriformis pyriformis
(Rubiaceae) near Durban, South Africa, for example, showed that they are wholly
dependent on vectors for pollination, but over four seasons of monitoring, only very
low hawk moth visitation rates were recorded – but with those moths the only
insects visiting the flowers (Johnson et al. 2004). Fewer than one of every thousand
flowers developed fruit between 2001 and 2003.
Nocturnal pollination networks, principally involving moths, can differ substan-
tially from the more easily observed daytime associations. Moth surveys in boreal
pine forest in Scotland confirmed the variety of taxa involved (Devoto et al. 2011)
as well as the considerable asymmetry within the associations. Most moths (around
90% of the 4162 individuals) carried no pollen, and the carriers represented only 25
of the total 103 species captured. It seemed unlikely that moths were key pollinators
in those forests.
The roles of many butterflies as pollinators continue to be debated, with some
authors regarding them primarily as opportunistic flower visitors that exploit nectar
as thieves. The wood white (Leptidea sinapis, Pieridae) in Europe is one example
enhancing this suggestion. L. sinapis shows strong preferences for particular flow-
ering species to gain nectar and was thus considered a likely pollinator of three
species, two species of Viola and Lathyrus montanus, in Sweden (Wiklund et al.
1979). More than 90% of butterfly visits were to these three species, in a community
of approximately 70 nectar-producing species. Examination of butterflies captured
after flower visits showed the presence of pollen to be very low – only seven rele-
vant pollen grains were found on 19 butterflies that had collectively visited 30 flow-
ers. L. sinapis seemed unable to function as an effective pollinator of its preferred
nectar flowers and was considered a parasite (thief) of these. The pollen levels on
the butterflies contrasted markedly with, for example, those from a survey of visits
by another pierid (Colias sp.) to Phlox in North America (Levin and Berube 1972),
where mean numbers of pollen grains from flower visits were 1000–2000.
Very few well-defined pollination mutualisms involving butterflies have been
documented. The classic study by Gilbert (1975) on Passiflora-feeding Neotropical
Heliconius butterflies (Nymphalidae) and cucurbit vines (Cucurbitaceae), which
flower throughout the year in tropical forests in Trinidad, remain one of the most
informative and far-reaching investigations of the tripartite relationships between
the pollen plant, butterfly and larval food plant, with components and influences
summarised in Fig. 3.2.
The European carnation Dianthus carthusianorum (Caryophyllaceae) is special-
ised for pollination by Lepidoptera, but in Switzerland depended on only two of the
Fig. 3.2 Summary of pathways and influences of various factors on a system of interacting spe-
cies at individual, population and community levels. The examples of Neotropical Heliconius but-
terflies and their mutualism with Anguria vines which they pollinate and are rewarded by pollen/
nectar and whose larvae feed on Passiflora species (Gilbert 1975, as a classic example that has
stimulated much more recent advance)
five recorded pollinator species (Bloch et al. 2006). Any decline in (or loss of) either
of those could seriously threaten the plant. Those butterflies, Melanargia galathea
and Satyrus fevula (Nymphalidae), have both declined substantially in the area and
have disappeared from many previously occupied sites. It seemed unlikely, should
these two be lost, that the remaining three species could adequately compensate for
the lost pollination, and Dianthus would be likely to suffer from increasing inbreed-
ing depression.
In short, frequent visitors to flowers and that are also effective pollinators may
interact more strongly with particular plant species than will rare visitors or fre-
quent visitors that are inefficient pollinators. Pollinators will also interact more
strongly with plants that provide a generous reward and are visited frequently than
with rare plants or plants visited frequently but proffering only poor rewards
(Vazquez and Aizen 2004).
The few highly specific integrated obligate mutualisms discussed below are by
far outnumbered by those briefer encounters. The wider relationships between
insects and flowers, with revelations of pollinator declines in both anthropogenic
and natural ecosystems engendering massive conservation concerns for both crop
and natural vegetation sustainability, are of critical interest and importance. In
accordance with the reality that coevolution between members of such partnerships
can lead to very specific associations, many commentators have referred broadly,
and perhaps misleadingly, to all reasonably specific pollinator-plant associations as
‘mutualisms’. Nevertheless, many such associations can be defined by the presence
of only very limited arrays of pollinators, imposing specialisations and dependence
of the plants on these. For many cases, any contributory roles by other species have
not been determined.
Reproduction of orchids, a diverse group of immense horticultural and ecologi-
cal interest, involves specialised mycorrhizal and pollinator mutualisms. Orchids
are a high priority in plant conservation in which the specialised pollinator associa-
tions ‘may pose considerable conservation challenge’ (Phillips et al. 2009). Those
challenges are increased by complex systems of pollinator attraction. Sexual mim-
icry commonly combines the twin strategies of (1) attracting male insects by mim-
icking the sex pheromone of the female of the pollinator species as a longer range
component and (2) visual and tactile mimicry of the female as a stimulus to pseudo-
copulation once encountered. Many orchids provide no reward to the pollinator (so,
despite their intricacy, may not be strictly defined as mutualisms), but others do so.
The large orchid genus Caladenia in Australia and nearby regions, for example,
includes some species that provide small amounts of nectar and may thus encourage
repeated visits from pollinators (Phillips et al. 2009). However, functional knowl-
edge of these relationships is still inadequate, notwithstanding that high levels of
orchid-pollinator specificity are often assumed. The thynnine wasp pollinators of
Australian Caladenia fall into the two groups of (1) ‘major responders’, putative or
confirmed pollinators that remove pollinia or pollen or exhibit strong sexual behav-
iour at the deceptor flowers and (2) ‘minor responders’, other species that seem to
be only rarely attracted to the flowers, or lack sexual behaviour after landing. In
some groups, pollinator sharing between closely related Caladenia species occurs,
but taxonomic complexities in both the orchids and the wasps leave some ambiguity
in confirming this.
Pollinator sharing amongst related orchids is apparently widespread. Thus, a

guild of 15 species of oil-secreting orchids in South Africa, all Coryciinae but oth-
erwise diverse, are all adapted for exclusive pollination by the oil-collecting bee
Rediviva peringueyi (Melittidae) but the generalist bee has only low dependence on
the orchids. It more commonly frequents flowers of the more abundant oil-secreting
Scrophulariaceae (Pauw 2006; Pauw and Bond 2011). Seed production by the
orchids, however, depends wholly on pollination by this bee, and co-occurring
orchid species avoid reproductive interference by placing pollinia on different parts
of the bee. Self-pollination does not occur.
The 15 orchid species occur in close associations and share key features that
constitute a ‘syndrome of floral features’ (Pauw 2006) as (1) pale yellow-green
flowers without extensive black markings; (2) secretion of floral oil as a reward for
pollinators; (3) a characteristic pungent scent; (4) a flowering period over mid-
August to late October, peaking in September; and (5) flower depth of 5–8 mm.
These similarities reflect adaptations for pollination by a shared pollinator, and
direct observations and inference led Pauw to designate the orchids as a ‘pollination
guild’. Individual bees were observed flying between inflorescences, and each bee
carried up to 27 pollinia of up to three orchid species. Specialisation was further
evident because the five other co-occurring species of Rediviva did not visit mem-
bers of this pollination guild – but, rather, attended other Scrophulariaceae with
brightly coloured flowers. It is also likely that the smaller bee species could not
reach the rewards offered by the guild orchids. The conservation scenario presented
is sobering: Pauw (2006) noted the severe (ca 80%) loss of the parent vegetation of
the orchids that had already occurred has led to retention of only tiny fragments of
their natural habitat, many less than a hectare in size. Rediviva was absent from
many of these fragments, with resultant pollination failure and increased functional
isolation of remaining viable orchid populations.
In this example, pollination success influenced the composition of the orchid
assemblage, acting as a ‘biotic filter’ whereby with decreasing pollination some
species were lost. These were successively the least clonal species, with the differ-
ent orchids differing considerably in their ability to reproduce through underground
bulbils and, hence, be less reliant on pollination for their persistence. Pauw and
Bond (2011) suggested that conservation of the non-clonal orchids necessitates con-
servation of bee pollination as a critical aspect of conservation planning. Some
conservation areas proclaimed at that time were believed to be too small to maintain
a population of Rediviva (Pauw 2007, above), so that specialist plant species that
rely on it for pollination cannot produce seeds. As one example, monitoring of the
endangered orchid Pterygodium cruciferum, thought to be pollinated by the bee,
had zero capsule set in three successive years on one small site, with other co-
occurring species there also failing. Dependence on a single pollinator species
equates to very low level of ecological redundancy, and the loss of the bee is likely
to flow to cause extinctions of the species within this pollination guild. Problems of
interpreting the extent of mutualisms in such contexts, in which a guild of related
taxa share a common specialisation limiting the taxa with which they can interact,
are shown well by other oil-collecting pollinating bees. Production of floral oil, and
the associations of the bees that collect that oil, illustrates the lability of those asso-
ciations. Close associations have been gained or lost many times during the history
of both groups of participants, but the oil-producing ‘eliaphores’ occur in approxi-
mately 2000 plant species in 11 families of angiosperms, and oils are collected by
about 450 bee species. The largest oil-collecting bee genus (Centris) includes about
230 species and, in references cited by Martins et al. (2013), the basis for the need
of oils is for female bees to surface the nest and nest walls and as a larval food (when
mixed with pollen). Fifteen of the 25 species of oil-collecting bees on Angelonia
(six species) and Basistemon (one species), both Plantaginaceae, were considered
effective pollinators, with specialisation amongst the species including the region of
the body on which pollen was carried for transfer. This specialised pollinator guild,
with its complex interactions with individual host species, may allow considerable
flexibility within the broader constraints of the defined mutualism.
The oil-collecting bees comprise only about 1.5% of the world’s bees, but several
bee species may visit a given plant species. Thus, observations on five populations
of Nierembergia linariifolia (Solanaceae) over 2 years in Argentina (Cosacov et al.
2008) showed the variety of differing balances and pollinator visitation rates that
create a mosaic of pollination mutualisms within the complex. Oil-collecting bees
were by far predominant, with >96% of floral visits (Fig. 3.3). How the pollination
association may be interpreted is relevant far more widely in pollinator guilds and
specificity: the number of pollinator species implies generality but, from a func-
tional viewpoint, the system is highly specialised because different oil-collecting
bees were predominant at each Nierembergia population and with differing levels of
specificity amongst those populations.
Fig. 3.3 The bees visiting the oil-producing Nierembergia linariifolia (Solanaceae) in surveys of
five populations over 2 years. Mean number of bee visits/flower/hour is shown in relation to func-
tional roles of the visiting species. Bee assemblages are shown as large-sized oil bees (two species,
black), small- and medium-sized oil bees (three species, dotted) and pollen-collecting bees (two
species, open); sites are (1) Pampa de Achala, (2) University, (3) Dean Funes, (4) Capilla del
Monte and (5) Charbonier (Simplified after Cosacov et al. 2008)
Such restricted pollination networks may have properties that render them more
resilient than larger networks. The Neotropical mutualisms between oil-producing
Malpighiaceae and specialised solitary oil bees (Centridini), analysed by Bezerra
et al. (2009), comprise highly discrete and characteristic modules. Field studies of
the 26 species (13 each of plants and bees) in Brazil confirmed that all the
Malpighiaceae interact only with Apidae, and 10 of the 13 bees were Centridini.
The other three bee species took only pollen and did not collect oil. This pollination
network was very resilient to extinction because any single species extinction was
readily compensated within the network and caused little functional change. In part,
this reflected a phylogenetically restricted suite of taxa in both roles, and links with
high nestedness through the core of highly committed species that interact mainly
with each other, and a subset of more specialist species that interact with those rela-
tive generalists (Bascompte et al. 2003; Krishna et al. 2008). The ‘key’ to nested-
ness, high also in some ant-plant mutualistic networks (Chamberlain et al. 2010), is
that such a generalist core is attended by a few specialists that add to the number and
variety of interactions.
3.2 Obligate Pollination Mutualisms
Although many angiosperms are pollinated by associates as vectors, relatively few

obligate pollination mutualisms are known, and those involving active pollination
are rare (Pellmyr 1997). ‘Active pollination’ refers to the conditions where specific
pollinators possess morphological and behavioural features for the purpose of
retrieving and transporting pollen and depositing it on stigmas.
The two most familiar obligate pollinator mutualisms are those between figs
(Ficus spp., Moraceae) and fig wasps (Agaonidae) and yuccas (Yucca spp.,
Agavaceae) and yucca moths (Tegeticula spp., Parategeticula spp., Prodoxidae),
both iconic systems that have been studied extensively over more than a century
(Addicott et al. 1990). They exemplify the close interactions in which pollination
mutualisms are associated clearly with the plants also providing breeding sites for
the pollinators. Those interactions, reflecting their specificity and intricacy, can be
divided into three main functional groups of pollination systems based on oviposi-
tion and larval food supplied to the pollinators (as ‘brood-site pollination mutual-
isms’: Sakai 2002). The most intricate of these groups are ovule parasites, the highly
specific associations typified by fig wasps and yucca moths and occurring only in
the few lineages outlined below. This association was regarded as rare, due to the
high costs of seed loss as a reward to pollinators – Sakai cited only five contexts
(figs-fig wasps, yuccas-yucca moths, Lithophragma-Prodoxidae, Trollius-
Chiastocheta [p. 53], Lophocereus-Upiga [p. 51]). The second group comprises
pollen parasites, in which larvae of pollinators, almost always thrips, feed on pollen
grains of fresh flowers on the plants as part of wider feeding on plant tissue. Many
thrips appear not to be host specific, and their importance as mutualists is often
uncertain. Sakai’s third category recognised pollinator larval development in
3.2 Obligate Pollination Mutualisms 47
decomposing flowers and inflorescences after pollination has occurred, a scenario in

which many small beetles and flies participate and which has garnered far less
detailed study than more specific or intricate mutualisms. Many of the insects
involved may be very casual or irregular pollinators.
The first category of mutualisms involves highly specific insect-plant associa-
tions in which the female insects actively pollinate by collecting and transferring
pollen and oviposit in the flower, whence developing larvae consume part of the
plant’s seed output. The associations thus involve both pollination and seed preda-
tion, so that the vectors are simultaneously mutualists and antagonists and are main-
tained by their life cycles of being synchronised with flowering phenology. No
co-pollinators are normally present in either case, and both these systems have con-
stituted intricate and manipulable microcosms for study.
The fundamental natural history of both these long associations between figs and
fig wasps and yuccas and yucca moths is well known, but subtleties within their
evolution continue to emerge and to stimulate continuing interest. The basic pattern
of each is summarised briefly below.
3.2.1 Figs and Fig Wasps
The genus Ficus (Moraceae) includes more than 750 species, collectively distrib-
uted over much of the tropics and subtropics. Figs, specialised inflorescences termed
‘syconia’, are pollinated obligatorily and exclusively by fig wasps that are also pri-
mary seed predators. Many appear to be highly host specific, and the widespread
presumption of a ‘one-to-one’ species relationship between fig and wasp has only
recently been revised to reveal far greater complexity in some groups. Several stud-
ies (reviewed by Machado et al. 2005) showed that breakdowns in the ‘one-to-one’
rule for fig-fig wasp pollination systems are common. In some cases, different
wasps are associated with the same fig in different parts of its range, and in some
others multiple pollinators occur together on the same host. A further complication
arises from hybridisation between figs. Genetic sampling (Machado et al. 2005)
also revealed that wasps which are indistinguishable genetically can pollinate dif-
ferent host species.
Specificity is influenced also by the ovipositor lengths of the various fig wasp
species, either facilitating or preventing access to suitable sites. Stability of some
fig-fig wasp mutualisms could be maintained if, for example, pollinators with short
ovipositors – so restricting oviposition to a particular subset of flowers – restricted
seed production from the inflorescence (Nefdt and Compton 1996). As also noted
earlier (Bronstein 1988), that some ovaries in the inflorescence of monoecious figs
have styles longer than the wasp ovipositor, and are thereby inaccessible, assures
that at least some seeds will mature, as well as that seed-eating wasp progeny will
be produced. The principle (Fig. 3.4, Anstett 2001) and the conflict (in which each
wasp larva develops at the cost of one fig ovary: see Janzen 1979) illustrates one of
three major hypotheses that have been discussed extensively in attempting to explain
Fig. 3.4 Schematic structure of the fig inflorescence (‘syconium’) to indicate organisation of the
ovaries, with differing lengths of styles and pedicels projecting towards the cavity from the con-
taining wall. Ovaries with long styles and short pedicels usually develop into seeds, whilst those
with short styles and long pedicels usually host a pollinator larva (Partly after Anstett 2001)
the ‘seed-wasp production ratio’. These, with several other hypotheses advanced
historically, are:
1. The ‘limited egg supply hypothesis’, that on average insufficient eggs are depos-
ited in any inflorescence to occupy all female flowers, so that pollination is
almost certain to occur
2. The ‘unbeatable seeds hypothesis’, that some flowers cannot be used by any pol-
linating or parasitoid wasp, so that seeds are ‘unbeatable’
3. The ‘ovipositor length optimisation hypothesis’, that costs of a longer ovipositor
render it too expensive to evolve an ovipositor that could reach all ovaries so that,
in this case, the number of wasp eggs is never limiting
These hypotheses are listed here simply to illustrate the complexities of inter-
preting evolutionary patterns amongst even these relatively well-documented mutu-
alisms (Anstett 2001; Kerdelhue and Rasplus 1996).
Wingless male fig wasps (which emerge before females, mate with them – some-
times before females leave the fig ovary within which they develop) cut exit holes
through the syconium wall. The winged females leave their natal syconium after
mating, carrying pollen and seek figs on another tree, which they then enter and
complete pollination and oviposition, the latter occurring many times within the
same syconium. Larvae develop within the fig ovaries.
As with many other tiny insects, the dispersal behaviour of female fig wasps is
intricate, with females of each generation of many species needing to move to other
trees to discover ‘receptive figs’ for oviposition. Dispersal is necessarily downwind
and incurs high mortality: for Elizabethiella baijanthi, a specific pollinator of the
South African Ficus burtt-davyi, Ware and Compton (1994a, b) estimated that only
5% of emerging female wasps successfully completed journeys to new fig trees.
Two major forms of pollination occur, with the wasps involved differing in struc-
ture and behaviour. Some are ‘active pollinators’ that remove pollen from storage in
pockets on their thorax using their forelegs and deposit it on the fig stigma, whilst
others are ‘passive pollinators’ that lack specialised thoracic structures and trans-
port pollen on their abdomen, whence it is transferred by direct contact with the fig
stigma. Fig species are either monoecious or dioecious, with rather different
implications for the fig-wasp mutualisms. In monoecious figs (about half the spe-
cies), each inflorescence produces seeds, wasps and pollen. The only way for the
trees to disperse pollen is through wasps breeding in the same inflorescence, and
female wasps must exploit an inflorescence in order to reproduce. The other, dioe-
cious, fig species have about half the trees female (producing only seeds) and the
other male (producing pollen and vector wasps). Gravid female wasps that enter
female figs and pollinate them cannot lay eggs and do not reproduce, dying without
doing so. The male trees allow wasps to reproduce. The mutualism between monoe-
cious figs and their specific agaonid pollinators is both intricate and depends on the
sequence of phases of fig development, within the general pattern shown in
Table 3.1. During flowering, only female florets are receptive at the time the female
wasps arrive, with larvae and intact seeds developing over the next several weeks.
The fig-fig wasp mutualism arose around 75 million years ago. The subsequent
co-diversification of the plants and pollinators is probably unique and reflects con-
siderable phylogenetic matching – in part linked to only limited host shifts having
occurred (Cruad et al. 2012). That extensive study, involving DNA surveys for more
than 750 interacting fig-fig wasp pairs, indicated that the intricate mutualisms rep-
resent an extreme example of long-term coadaptations and coordinated dispersals
that continue to influence the evolution of tropical ecosystems. Weiblen (2002) con-
cluded his account of fig wasp ecology by noting that the wasps ‘represent a most
extreme example of specialisation in the continuum of evolving plant-insect
interactions’.
3.2.2 Yuccas and Yucca Moths
The approximately 40 species of Yucca occur in Central and North America and are
pollinated only by specific members of two small (and closely related) moth genera.
The female moths possess unique maxillary ‘tentacles’ that collect and transport
Table 3.1 Major features of fig pollination-mutualism systems, in part after Anstett et al. (1997)
Monoecious species
Ancestral state
Each inflorescence produces seeds, pollen and wasps
Each female flower may produce either one seed or one fig wasp
Trees selected to breed wasps: only way to disperse pollen
Trees also selected to produce both seeds and pollinators
Female wasps selected to visit figs: only way to reproduce; also selected to produce as many
offspring as possible
So, selection on both partners to cooperate and may be ‘conflict’ between figs and wasps over
proportion of female flowers that will produce a seed
Dioecious (‘gynodioecious’) species
Probably originated at least twice from monoecious ancestors
Half the trees are female, producing only seeds
Half the trees are anatomically monoecious (inflorescences with male and female flowers),
but functionally male, producing pollen and pollen-transporting wasps
Male figs: wasps develop in female flowers that do not participate in reproduction, as only
very rarely produce seeds
Female figs: pollen-transporting wasps enter, but cannot lay eggs, and die without
reproducing but produce seeds through pollination
Both categories
Pollen of one fig dispersed only by female wasps mating within that fig, so production of
wasps is part of the male function of the plant
pollen, gathering a ball of pollen after they mate on the inflorescences. These
mouthparts facilitate precise handling of pollen and may be a key feature in the
mutualism’s development (Pellmyr and Krenn 2002). The moths then fly to another
inflorescence, where the pollen is deposited on the receptive stigma as the moth
oviposits in or on the pistil. Larvae feed on developing seeds and later leave the
plant and drop to the soil, where development is completed.
The yucca-yucca moth association was established at least 40 million years ago,
enabling much subsequent opportunity for modifying and refining the mutualisms
and trace aspects of the transition from antagonism to mutualism and reversals.
These two intricate systems are amongst the most thoroughly investigated polli-
nation mutualisms. Both have been described as ‘spectacular examples of coevolu-
tion’ (Westerbergh and Westerbergh 2001), but the initial stages of these intricate
mutualisms were noted in that account as ‘unknown and hypothetical’. Addicott
et al. (1990) noted the following contrasts between these three scenarios:
1. Adult yucca moths are active for only a short period each year, whilst adult fig
wasps can be active at any time.
2. Reproduction is synchronous amongst yucca inflorescences but asynchronous
amongst fig trees.
3. Fruit production in either may be complete, or partial, if the flowering is out of
synchrony with the respective pollinators.
The level of successful association thus depends very largely on the synchrony
of life cycle stages, notably that pollen-bearing female wasps or moths are active
whilst the host flowers are receptive for pollination and oviposition. Such ‘finely
tuned’ entwinings of life cycles give the appearance of full compatibility. Anstett
et al. (1997), however, considered that this apparently perfect association between
figs and fig wasps could also be interpreted as ‘reciprocal antagonism’, in that over
a long coevolutionary history each species has been selected to exploit its obligate
partner without being able to avoid being exploited itself. Conflicts within mutual-
isms are widespread, perhaps universal (Anstett et al. 1997), with many evolution-
ary conflicts noted amongst figs and fig wasps (Janzen 1979).
3.2.3 Other Examples
Figs-fig wasps and yuccas-yucca moth associations are the two most ‘iconic’ obli-
gate pollination mutualisms, and are by far the most intensively studied, with
increasingly sophisticated and detailed investigations revealing arrays of idiosyn-
crasies that demonstrate the difficulties of full understanding and the dangers of
interpretation from more superficial investigations. More recently several other
such cases of obligate mutualisms involving pollinators and seed predation have
been discovered and studied, as exemplified below.
The senita cactus (Lophocereus schottii, Pachycereae) in the Sonoran Desert has
a mutualistic relationship with the Senita moth (Upiga virescens, Pyralidae)
(Fleming and Holland 1998). The cactus is night-flowering and self-incompatible.
The moths, which rest on cactus spines during the day, visit flowers as soon as they
open around sunset and then:
1. Actively collect pollen by rubbing their abdomen (which has a posterior ‘brush’
of long slender scales) across dehisced anthers
2. Actively deposit the pollen on the erect flower stigma by rubbing the abdomen
over it
3. Lay an egg, often on the tip of a flower petal
4. Sometimes crawl into the corolla, presumably to obtain nectar.
Male moths lack the ‘pollen brush’ and are not involved in pollination. Larvae,
with a survival rate of only about 20%, become seed predators within the cactus’
ovary and later chew through the base of the fruit to reach a cactus branch within
which pupation occurs. Larvae destroy about 30% of fruits (and seeds) resulting
from pollination by the moth.
Female Upiga avoid laying on flowers that already have an egg. Fruits with an
exit hole are always abscised, so that all fruits that supported a single larva undergo
total seed mortality – even if only a proportion of seeds have been eaten. Flower
visitation by other insects at night was rare – only 1.2% of the 7660 individual noc-
turnal visitations recorded by Holland and Fleming (1999) were not U. virescens but
halictid bees pollinated during the day. U. virescens was thus regarded as an obligate
mutualist but L. schottii as at the interface between an obligate and a facultative

mutualist because pollination and seed set can sometimes occur from such co-
pollinators. Nevertheless, oviposition site selection by the moth is varied, with the
open cactus flowers presenting a range of substrates, such as petals, anthers and
corolla tubes (Holland et al. 2004). Most eggs, as inferred above, were laid on pet-
als, and far fewer on anthers and corolla tubes – although, allowing for relative areas
of the different substrates, more eggs were laid on anthers and corolla tubes than on
petals. Egg and first instar larval survival is higher from anthers and corolla tubes,
as hatchling larvae have less distance to travel to find fruit than those emerging from
petals, and wilting and sticky petals also contribute to high mortality.
This example, which has now been studied in considerable detail, augments per-
spective from the two major iconic mutualisms that have also served as models for
coevolutionary studies. It was used to demonstrate that reciprocal specialisation of
a pollinator and plant can be associated with the characteristics of both parties, as
discussed by Waser et al. (1996). Specialisation of a pollinator on a plant may be
predicted when the pollinator’s generation time is no longer than the duration of the
flowering season and the relevant stages of plant and pollinator life cycles synchro-
nise. Specialisation of a plant on a pollinator may be predicted when (1) the plant is
large and long-lived, (2) has many reproductive episodes and (3) has an effective
pollinator with predictable population dynamics. The strong association between
Upiga and Lophocereus should allow for effective synchronisation to develop
through coevolution.
The possible scenarios of evolution of moth pollination in Lophocereus were
discussed by Hartmann et al. (2002), as transitions from a bat-pollinated associa-
tion. Traits included (1) retention of nocturnal flowering but with reduced flower
size and nectar production; (2) an initial transition to hummingbird pollination from
bat pollination, followed by moth pollination preadapted by reduced flower size and
nectar production, but then with a reversal from diurnal to nocturnal flowering; and
(3) an ancestor with a more generalised pollination syndrome developing from bat
pollination and leading to the more specialised pollination system that can involve
insect pollinators. Such alternatives, suggested by molecular sequence data used to
determine the phylogeny of L. schottii within the wide array of North American
Pachycereae, supported that Lophocereus is a sister group to the hummingbird pol-
linated Pachycereus marginatus, with bat pollination the most likely ancestral phase
for both these later specialisations.
A further lepidopteran example, from Japan, involves a moth (Epicephala sp.,
Gracillariidae) and the monoecious tree Glochidion acuminatum (Phyllanthaceae),
studied by Goto et al. (2010). It shows many parallels with the Yucca scenario, in
that the female moths are the specific pollinators that use their specialised proboscis
to collect and transport pollen and oviposit into the flower pistil where the larvae
each consume some of the six seeds within each fruit. The habit appears to be wide-
spread within these groups (Kato et al. 2003). Svensson et al. (2010) and Kawakita
and Kato (2009) both suggested an estimate that about 500 species of Phyllanthaceae
are each pollinated by a host-specific species of Epicephala in the above manner,
the female moths laying in female flowers. A single female moth may visit about 20
female flowers sequentially during a single oviposition period.
An apparently unusual modification of the principle of ‘intermediate exploita-

tion’, whereby some seeds are free from predation in mutualisms between plants
and seed-feeding predators, occurs in the interaction between Epicephala moths
and two species of Phyllanthus from New Caledonia (Kawakita and Kato 2004). As
in other species, the flowers are pollinated exclusively by particular species of
Epicephala – but a single larva then consumes all six ovules of the developing fruit.
However, some fruits are left intact, presumed because moth eggs are absent from
some pollinated flowers. No evidence was found that this was because of egg mor-
tality, and Kawakita and Kato queried whether moths do not always lay in the flow-
ers they pollinate. This behaviour seemed paradoxical. One suggested corollary, in
need of verification, was that the behaviour might be an advantage because unim-
pacted fruits might force the braconid parasitoid of the moth larvae to spend exces-
sive time searching for hosts and so decrease the probability of them detecting and
attacking Epicephala larvae.
The taxonomic richness of the Epicephala-Glochidion mutualisms reflects rich
coevolution and speciation in the Old World tropics, with highly intricate
associations.
The scenario parallels many other plant-associated insect radiations for which
increasingly sophisticated molecular analyses have revealed large numbers of dis-
tinct species earlier masquerading as a few more polyphagous and morphologically
indistinguishable taxa. Thus, yucca moths are now known to comprise far more
species than previously suspected (Powell 1992). Pellmyr (2003) noted 13 described
species of Parategeticula rather than the three long-recognised taxa embraced until
then, with other species believed to also occur. Incidence of further cryptic species
seems increasingly likely to be confirmed as studies continue (Pellmyr et al. 2006).
Some moths are associated with two or more yucca species (Pellmyr and Segrave
2003). Likewise, proliferation of detected fig wasp species continues – with both of
these complex examples leading to changing perspectives on regional interactions
and challenging the formerly widely-held ‘one-to-one’ specificity in mutualistic
associations.
A non-lepidopteran pollinator mutualism elucidated in Finland and central
Europe involves the globeflower (Trollius europaeus, Ranunculaceae) and its seed-
eating pollinators, globeflower flies (Chiastocheta spp., Anthomyiidae), shown by
Pellmyr (1989) to be an obligate mutualism there. Unlike the above cases, this asso-
ciation involves both male and female flies as flower visitors, and the roles of ovi-
positing and non-ovipositing flies in passive pollination is still not wholly clear
(Jaeger et al. 2000). In this system, pollination and oviposition are independent,
without any direct causal relationship. Some populations observed showed very
high fly densities that could potentially lead to over-exploitation of the host plant –
but Jaeger et al. considered the association stable because T. europaeus could pre-
vent Chiastocheta from ovipositing and the fly larvae consume only a fraction of the
seeds present; they believed that the system could not go to extinction under those
conditions. That survey showed that as no significant pre-predation increase in seed
set occurred with increased numbers of fly eggs, and flowers without Chiastocheta
eggs had high seed set, most pollination may be achieved by non-laying fly visitors.
The outcomes that (1) additional eggs did not lead to higher pollination and
(2) oviposition was not a mutualistic behaviour implied a conflict between the part-
ners for the number of eggs laid.
That only one species of Trollius is solely pollinated by Chiastocheta flies may
imply that this obligate association evolved only recently (Despres et al. 2002). The
more numerous (about 30 species) Asian species of Trollius are more varied, some
pollinated by these flies but others entirely free of them, so that the association is
obligate in Europe but commonly facultative in Asia where the greatest diversity of
the plants occur. Further, six species of Chiastocheta (and two other, vicariant, spe-
cies) coexist as pollinators of T. europaeus in Europe, leading Despres et al. (2002)
to note the apparent paradox of fly diversification being higher in Europe than that
associated with greater host diversity in Asia. They suggested that this might be
linked to the carpel number of the flowers that compensated for seed losses due to
predation.
The aquatic perennial water soldier (Stratiotes aloides, Hydrocharitaceae) is
dioecious but largely reproduces vegetatively and is a valued indicator species for
high-biodiversity aquatic habitats in Europe. Populations in some countries have
declined in recent years, stimulating investigations of the roles of insect pollinators
in the plant’s seasonal life cycle. The fly Hydrellia tarsata (Ephydridae) was found
abundantly in both male and female Stratiotes flowers in Germany (Katzenberger
and Zacharias 2015). Pollen is actively transferred on the flies, which develop in S.
aloides leaves on the water surface, and flies are attracted to the flowers both visu-
ally and by their rotting meat smell. This association was regarded as mutualistic,
with Hydrellia the major pollinator – no other flower visitors were thought to be
sufficiently regular in incidence to undertake that role effectively.
In the above cases, the gains in seed production from successful pollination are
balanced against the reduction in plant fitness from seed consumption by larvae of
the pollinators and determine the balance of the interaction, in which the success of
one partner determines success of the other. As Song et al. (2014) noted, many of
these cases are driven in part by floral scents that influence contacts between the
partners, and this had been documented for fig wasps, yucca moths, Epicephala
moths and Chiastocheta flies, amongst others. They described a similar mutualism
in the Himalayan perennial herb Rheum nobile (Polygonaceae) and a fly, the fungus
gnat Bradysia sp. (Sciaridae), with female gnats the main pollinators and experi-
mental evidence that the flies are specifically attracted by an unusual volatile floral
compound (2-methyl butyric acid methyl ester). The flowers of R. nobile do not
produce nectar, so other visitors are rare, and most fruit set under natural conditions
was due to pollination by Bradysia – somewhat unexpectedly because of the
widespread belief that fungus gnats are generally ineffective pollinators. This mutu-
alism may not be as exclusive as some others, above, because R. nobile can set fruit
by self-pollination. Examination of the floral scents at different stages of flower
development confirmed that attraction of the gnats is very stage specific, with the
female flies attracted only at anthesis. Although the possibility that the yellow bracts
also attracted the gnats at that time could not be wholly excluded, the very charac-
teristic scent was considered the primary attractant (Song et al. 2014).
3.3 Nursery Pollination 55
3.3 Nursery Pollination
The broad term ‘nursery pollinators’ is sometimes used for mutualistic systems in
which pollinator offspring feed on the pollinated seeds or fruits. As the above cases
indicate, pollination mutualisms in which the plant provides breeding sites for pol-
linators are almost invariably highly specialised. Of the three categories recognised
by Sakai (2002) (p. 48), the first is clearly the most restrictive and necessitates great-
est specialisation, but related and, perhaps, even more unusual situations can occur.
Thus, Humboldtia brunonis (Fabaceae) is polymorphic for the presence of domatia
(p. 75), with some trees lacking these structures – so that, rather than being a true
‘ant-plant’ (Chap. 4), it is regarded as a ‘semi-myrmecophyte’ (Shenoy and Borges
2008). The pollinating bee Braunsapis puangensis in India breeds within the doma-
tia and is the predominant visitor to flowers of Humboldtia; however, across its
range at least 20 ant species occupy those domatia and the level of ant aggressive-
ness may influence whether the pollinating bee can become resident. Humboldtia
plants with domatia may benefit both from pollination by resident bees and protec-
tion by resident ants.
In the best-known nursery pollination mutualisms, the adult female insects are
the exclusive pollen vectors of their host plants (Dufay and Anstett 2003). The pos-
sible conflicts of interest that can arise between the mutualists are amenable to
investigation, because benefits to each partner can be expressed directly as numbers
of offspring, but Dufay and Anstett recognised three major evolutionary conflicts of
interest – a ‘conflict’ defined as occurring when a process has negative effects on
one partner and positive effects on the other. Those conflicts (Table 3.2) raise ques-
tions in any of the parallel but evolutionarily independently derived cases. In gen-
eral, it seemed that resolution of these conflicts occurred, not from true coevolution
but by some evolutionary changes in a feature of only one partner.
Table 3.2 The main evolutionary conflicts amongst nursery pollination mutualists
Effect on each
Phenology partner
Plant Pollinator Life cycle stage Plant Pollinator Question
Fruit Larvae Larval growth supported by − + 1
inflorescence tissues
Receptive Adults Pollinators visit either
flower Rewarding flower + +
Non-rewarding flower + − 2
Receptive Adults Some pollinators actively + −/0 3
flower pollinate
The three conflicts (each formulated here as a question, listed below) reflect (1) larvae developing
at the expense of some tissues of developing fruits, (2) conflict in dioecious species between non-
rewarding flowers and adult pollinators and (3) pollinators that benefit the flowers but entail costs
or reduced benefits for themselves in doing so (From Dufay and Anstett 2003)
1. Why do plants not kill larvae of their pollinators?
2. Why do pollinators pay the cost of visiting non-rewarding flowers?
3. Why do pollinators pollinate?
Although major obligate mutualisms such as those noted above are relatively
few, less specific – but still restricted – pollinator-plant systems are far more numer-
ous and often more antagonistic in their implications. One well-studied example is
of the moths Hadena bicruris (Noctuidae) and Perizoma sp. (Geometridae) as
mutualist pollinators of Silene and related Caryophyllaceae, which provide a range
of nursery pollinations interpreted across the gradient of antagonistic to potentially
mutualistic (Kephart et al. 2006). They are of interest in assessing the evolution of
mutualisms but currently appear to show little evidence of any true mutualisms, in
part because they lack strict specialisations and show considerable variability in
host specificity and the outcomes of the interactions recorded (Kephart 2006).
Hadena moths pollinate their host plants and oviposit on the ovaries within the
flowers; their larvae later feed inside the developing fruits. The numerous cases
reviewed by Kephart et al. suggested that the structure of Caryophyllaceae-moth
interactions represents a relatively unspecialised suite of nursery pollinations in
which regulation of seed damage by the moths may not occur, with the nature of the
interaction being both context and species dependent. Little evidence was found of
fulfilling the strict criteria that would characterise any obligate mutualism and
reflect close association of the partners in time and space, namely, (1) a moth lays
eggs on a single taxon of Caryophyllaceae; (2) a taxon of Caryophyllaceae has a
single moth taxon that is responsible for ovule and seed predation; and (3) that moth
is the plant’s sole or major pollinator. Although many Hadena and Silene species
exhibit specificity for no more than one or two partners, some plants associated with
up to nine species of Hadena. In general, more detailed field studies are needed to
examine the costs and benefits within the range of interactions and how these are
influenced by the particular community context in which they occur.
3.4 Broader Perspective
The mutualisms that involve nursery pollination systems clearly represent a high
taxonomic variety of both plants and insects with, despite independent and disparate
evolutionary developments, many features in common. One such feature is the
mediating role of volatile floral scents between the partners (Hossaert-McKey et al.
2010), and the ways by which these have been investigated. Many gaps in under-
standing remain, and the specific signals to which many pollinators respond are still
unclear – but with the specialised obligate mutualisms it seems that signals are also
highly specific and recognised only by the plant’s specific pollinator.
In contrast to the highly integrated fig and yucca associations, examples involv-
ing Hadena/Perizoma–Silene are often considered a more ‘primitive’ stage of a
potential mutualism that, depending on the influences of co-pollinators, may shift
from mutualism to parasitism (Gimenez-Benavides et al. 2007). These authors stud-
ied the association in Spain between Hadena consparcatoides and Silene ciliata,
with the latter the only known host plant for this univoltine moth. Pollination is
nocturnal, and flowers are largely closed during daylight hours – but this does not
3.4 Broader Perspective 57
Fig. 3.5 Seed production by Silene dioica in relation to the proportion of flowers pollinated by
insects other than the obligate seed predator moth Perizoma affinitatum, to indicate the subtle
transition between costs and benefits in relation to pollinator activity (Westerbergh 2004)
deter diurnal visitors such as Syrphidae (>80% of diurnal visitors) that can eat the
pollen of exserted anthers and hawk moths that can reach the nectar. Exclusion trials
showed that fruit set and seed set/fruit were affected differently by nocturnal and
diurnal pollination – the former showing low fruit set attributed to a shortage of
nocturnal pollinators, whilst generalist diurnal co-pollinators may compensate for
such shortage and is driven in part by a generalist flower scent and avoidance of any
specific olfactory cues for Hadena. In another study, Silene dioica benefitted from
the trade-off between pollination and seed predation through Perizoma affinitatum,
for which the balance is positive only if co-pollinators attended fewer than 60% of
the flowers (Westerbergh 2004). Under those conditions, Westerbergh’s model
(Fig. 3.5) showed that Perizoma produces more seeds than its larvae consume. The
persistence of both partners is thus likely, and the interaction is stabilised further by
the presence of S. dioica flowers with different stigma lengths, some of which ham-
per the moth from ovipositing in short-styled flowers. In isolated populations, where
co-pollinators are rare (and gene flow restricted), the interaction is essentially shift-
ing from parasitism to mutualism.
True pollination-related mutualisms reflect likely reciprocal specialisation or
coevolution, but exploiters may also become intricately adapted to coexist within
these systems (Chap. 8). More widely, the twin strands of traditional studies on
plant-pollinator interactions, respectively, emphasising either the ecological or the
evolutionary processes involved, have become integrated progressively over the last
half century or so (Mitchell et al. 2009) to recognise their diversity and the unifying
features that occur, and also to increasingly acknowledge the limits of ‘specialisa-
tion’ and ‘generalisation’ from the viewpoint of either plant or pollinator, and the
importance of defining the ‘pollination syndrome’ attending each plant. The impli-
cations of different flower species flowering together and sharing a suite of pollina-
tors include this leading to detrimental competition. Early history of this idea
(reviewed by Mitchell et al. 2009) revealed awareness of the various possible out-
comes, with facilitation, as well as competition, being possible. Facilitation could
occur, for example, if flowers of several species are assessed by pollinators as equiv-
alent, so that adding more flowers of another species could increase the total number
Fig. 3.6 A conceptual framework of mechanisms of competition for pollinators that incorporates
both pollinator quantity (number of visits) and pollinator quality. The major drivers, essentially
variable in time and space, are shown to the left of the diagram and flow to functions of male
(lower) and female (upper) sexual functions (Mitchell et al. 2009)
of pollinators visiting the community. Increased per capita visits to one or more
species may also be possible, but the scheme summarised in Fig. 3.6 displays the
variables involved and emphasises the distinctions between ‘quality’ and ‘quantity’
of pollinator visits.
Nemestrinid flies, Prosoeca spp. (p. 181), in southern Africa are the exclusive
pollinators of some species of Lapeirousia (Iridaceae) (Goldblatt et al. 1995), and
the genus Nivenia in the same family is also pollinated predominantly by these flies,
with secondary pollination from female anthophorid bees. Within the insect-
pollinated species of the subgenus Lapeirousia, three functional groups were recog-
nised by flower type: (1) pollinated exclusively by Prosoeca; (2) pollinated
exclusively by another nemestrinid (Moegistorhynchus longirostris) and a tachinid
fly (Philoliche gulosa), these two having the longest proboscis of all the pollinators
found on these plants; and (3) pollination by a combination of bees (predominantly
female anthophorids and native Apis mellifera) with lesser roles by some bombyliid
flies and Lepidoptera. A possible fourth syndrome can be distinguished if the spe-
References 59
cies pollinated by hawk moths (Sphingidae) are separated from the third, generalist,
group above. Whatever the predominant pollinators, most of the flowering species
secrete nectar which is rich in sucrose, considered a characteristic of flowers
pollinated by long-tongued flies. However, in this association, the plants are more
specific than the flies, which can also visit other taxa.
Plant-pollinator mutualisms form one of the major associations between these
taxa. The other that has commanded massive attention is the role of insects as ‘pro-
tectors’, aiding defence of the plants against herbivores and encroaching competi-
tive vegetation. The paramount group of insects involved, ants, form numerous
intricate associations with plants (Chap. 4). Many plants interact with different
guilds of mutualists, for example, with pollinators and protectors, simultaneously,
and the traits needed may differ accordingly. In that context, pollinating mutualists
are attracted by floral features, whilst protectors may link with extrafloral nectaries
and/or domatia (p. 75). These scenarios have only rarely been assessed together to
determine whether they may be correlated in some way. Using 37 species of
Gossypium (Malvaceae) in greenhouse conditions, Chamberlain and Rudgers
(2012) examined correlations between six extrafloral nectary traits and four floral
traits, to discover if the traits needed to maintain multiple mutualist guilds are cor-
related across related plant species. Positive interspecific correlations between traits
occurred widely, and the two categories were never negatively correlated – as find-
ings that may contribute to more general understanding of how such features sustain
multiple mutualisms.
References
Addicott JF, Bronstein JL, Kjellberg F (1990) Evolution of mutualistic life-cycles: yucca moths
and fig wasps. In: Gilbert F (ed) Genetics, evolution and coordination of insect life cycles.
Springer, Berlin, pp 143–161
Anderson B, Olivieri I, Lourmas M, Stewart BA (2004) Comparative population genetic structures
and local adaptation of two mutualists. Evolution 58:1730–1747
Anstett MC (2001) Unbeatable strategy, constraint and coevolution, or how to resolve evolutionary
conflicts: the case of the fig/wasp mutualism. Oikos 95:476–484
Anstett MC, Hossaert-McKey M, Kjellberg F (1997) Figs and fig pollinators: evolutionary con-
flicts in a coevolved mutualism. Trends Ecol Evol 12:94–99
Ball JB (2012) Lepidopterology in southern Africa: past, present and future. In: New TR (ed)
Insect conservation: past, present and prospects. Springer, Dordrecht, pp 279–300
Bascompte J, Jordano P, Melian CJ, Olesen JM (2003) The nested assembly of plant-animal mutu-
alistic networks. Proc Natl Acad Sci 100:9383–9387
Beaman RS, Decker PJ, Beaman JH (1988) Pollination of Rafflesia (Rafflesiaceae). Am J Bot
75:1148–1162
Bezerra ALS, Machado IC, Mello MAR (2009) Pollination networks of oil-flowers: a tiny world
within the smallest of all worlds. J Anim Ecol 78:1096–1101
Bloch D, Werdenberg N, Erhardt A (2006) Pollination crisis in the butterfly-pollinated wild carna-
tion Dianthus carthusianorum? New Phytol 169:69–76
Bronstein JL (1988) Mutualism, antagonism, and the fig-pollinator interaction. Ecology
69:1298–1302
Chamberlain SA, Rudgers JA (2012) How do plants balance multiple mutualists? Correlations
among traits for attracting protective bodyguards and pollinators in cotton (Gossypium). Evol
Ecol 26:65–77
Chamberlain SA, Kilpatrick JKI, Holland JN (2010) Do extrafloral nectar resources, abundances,
and body sizes contribute to the structure of ant-plant mutualistic networks? Oecologia
164:741–750
Cosacov A, Nattero J, Cocucci AA (2008) Variation of pollinator assemblages and pollen limita-
tion in a locally specialised system: the oil-producing Nierembergia linariifolia (Solanaceae).
Ann Bot 102:723–734
Cruad A, Ronsted N, Chantarasuwan B, Chou LS, Clement WL (and 24 other authors) (2012) An
extreme case of plant-insect codiversification: figs and fig-pollinating wasps. Syst Biol.
doi:10.1093/sysbio/sys068
Despres L, Pettex E, Plaisance V, Pompanon F (2002) Speciation in the globeflower fly Chiastocheta
spp. (Diptera: Anthomyiidae) in relation to host plant species, biogeography, and morphology.
Mol Phylogenet Evol 22:258–268
Devoto M, Bailey S, Memmott J (2011) The ‘night shift’: nocturnal pollen-transport networks in a
boreal pine forest. Ecol Entomol 36:25–35
Dufay M, Anstett M-C (2003) Conflicts between plants and pollinators that reproduce within inflo-
rescences; evolutionary conflicts on a theme. Oikos 100:3–14
Fleming TH, Holland JN (1998) The evolution of obligate pollination mutualisms: senita cactus
and senita moth. Oecologia 114:368–375
Gilbert LE (1975) Ecological consequences of a coevolved mutualism between butterflies and
plants. In: Gilbert LE, Raven PH (eds) Coevolution of animals and plants. University of Texas
Press, Austin/London, pp 210–240
Gimenez-Benavides L, Dotterl S, Jurgens A, Escudero A, Iiriondo JM (2007) Generalist diurnal
pollination provides greater fitness in a plant with nocturnal pollination syndrome: assessing
the effects of a Silene -Hadena interaction. Oikos 116:1461–1472
Goldblatt P, Manning JC, Bernhardt P (1995) Pollination biology of Lapeirousia subgenus
Lapeirousia (Iridaceae) in southern Africa: floral divergence and adaptation for long-tongued
fly pollination. Ann Mo Bot Gard 82:517–534
Goto R, Okamoto T, Kiers ET, Kawakita A, Kato M (2010) Selective flower abortion maintains
moth cooperation in a newly discovered pollination mutualism. Ecol Lett 13:321–329
Hartmann S, Nason JD, Bhattacharya D (2002) Phylogenetic origins of Lophocereus (Cactaceae)
and the senita cactus-senita moth pollination mutualism. Am J Bot 89:1085–1092
Holland JN, Fleming TH (1999) Mutualistic interactions between Upiga virescens (Pyralidae), a
pollinating seed-consumer, and Lophocereus schottii (Cactaceae). Ecology 80:2074–2084
Holland JN, Buchanan AL, Loubeau R (2004) Oviposition choice and larval survival of an obli-
gately pollinating granivorous moth. Evol Ecol Res 6:607–618
Hossaert-McKey M, Soler C, Schatz B, Proffit M (2010) Floral scents: their roles in nursery pol-
lination mutualisms. Chemoecology 20:75–88
Inouye DW, Larson BMH, Ssymank A, Kevan PG (2015) Flies and flowers III: ecology of foraging
and pollination. J Pollin Ecol 16:115–133
Jaeger N, Till-Bottraud I, Despres L (2000) Evolutionary conflict between Trollius europaeus and
its seed-parasite pollinators Chiastocheta flies. Evol Ecol Res 2:885–896
Janzen DH (1979) How to be a fig. Annu Rev Ecol Syst 10:13–51
Johnson SD, Neal PR, Peter CI, Edwards TL (2004) Fruiting failure and limited recruitment in
remnant populations of the hawkmoth-pollinated tree Oxycanthus pyriformis subsp. pyriformis
(Rubiaceaae). Biol Conserv 120:31–39
Kato M, Takimura A, Kawakita A (2003) An obligate pollination mutualism and reciprocal diver-
sification in the tree genus Glochidion (Euphorbiaceae). Proc Natl Acad Sci 100:5264–5267
Katzenberger J, Zacharias D (2015) Mutualism of Stratiotes aloides L. (Hydrocharitaceae) and
Hydrellia tarsata Haliday (Diptera: Ephydridae): tritrophic interaction of macrophyte, leaf-
mining dipteran pollinator and parasitoid Braconidae. J Pollin Ecol 15:23–29
References 61
Kawakita A, Kato M (2004) Evolution of obligate pollination mutualism in New Caledonian

Phyllanthus (Euphorbiaceae). Am J Bot 91:410–415
Kawakita A, Kato M (2009) Repeated independent evolution of obligate pollination mutualism in
Phyllantheae-Epicephala associations. Proc R Soc B 276:417–442
Kephart S (2006) Pollination mutualisms in Caryophyllaceae. New Phytol 169:637–640
Kephart S, Reynolds R, Rutter M, Fenster C, Dudash M (2006) Pollination and seed predation by
moths on Silene and allied Caryophyllaceae: evaluating a model system to study the evolution
of mutualisms. New Phytol 169:667–680
Kerdelhue C, Rasplus J-Y (1996) The evolution of dioecy among Ficus (Moraceae): an alternative
hypothesis involving non-pollinating fig wasp pressure on the fig-pollinator mutualism. Oikos
77:163–166
Krishna A, Guimares PR Jr, Jordano P, Bascompte J (2008) A neutral-niche theory of nestedness
in mutualistic networks. Oikos 117:1609–1618
Larson BMH, Kevan PG, Inouye DW (2001) Flies and flowers: taxonomic diversity of anthophiles
and pollinators. Can Entomol 133:439–465
Levin DA, Berube DE (1972) Phlox and Colias: the efficiency of a pollination system. Evolution
26:242–250
Machado CA, Robbins N, Gilbert MTP, Herre EA (2005) Critical review of host specificity and its
coevolutionary implications in the fig/fig-wasp mutualism. Proc Natl Acad Sci
102:6558–6565
Martins AC, Aguiar AJC, Alves-dos-Santos I (2013) Interaction between oil-collecting bees and
seven species of Plantaginaceae. Flora 208:401–411
Minckley RL, Roulston TH (2006) Incidental mutualisms and pollen specialization among bees.
In: Waser NM, Ollerton J (eds) Plant-pollinator interactions from specialization to generaliza-
tion. University of Chicago Press, Chicago/London, pp 69–98
Mitchell RJ, Flanagan RJ, Brown BJ, Waser NM, Karron JD (2009) New frontiers in competition
for pollination. Ann Bot 103:1403–1413
Nefdt RJC, Compton SG (1996) Regulation of seed and pollinator production in the fig-fig wasp
mutualism. J Anim Ecol 65:170–182
Pauw A (2006) Floral syndromes accurately predict pollination by a specialized oil-collecting bee
(Rediviva peringueyi, Melittidae) in a guild of South African orchids (Coryciinae). Am J Bot
93:917–926
Pauw A (2007) Collapse of a pollination web in small conservation areas. Ecology 88:1759–1769
Pauw A, Bond WJ (2011) Mutualisms matter: pollination rate limits the distribution of oil-secreting
orchids. Oikos 120:1531–1538
Pellmyr O (1989) The cost of mutualism: interactions between Trollius europaeus and its pollinat-
ing parasites. Oecologia 78:53–59
Pellmyr O (1997) Pollinating seed eaters: why is active pollination so rare? Ecology
78:1655–1660
Pellmyr O (2003) Yuccas, yucca moths, and coevolution: a review. Ann Mo Bot Gard 90:35–55
Pellmyr O, Krenn HW (2002) Origin of a complex key innovation in an obligate insect-plant mutu-
alism. Proc Natl Acad Sci 99:5498–5502
Pellmyr O, Segrave KA (2003) Pollinator divergence within an obligate mutualism; two Yucca
moths (Lepidoptera: Prodoxidae: Tegeticula) on the Joshua tree (Yucca brevifolia: Agavaceae).
Ann Entomol Soc Am 96:716–722
Pellmyr O, Balcazar-Lara M, Althoff DM, Segraves KA, Leebens-Mack J (2006) Phylogeny and
life history evolution of Prodoxus yucca moths (Lepidoptera: Prodoxidae). Syst Entomol
31:1–20
Phillips RD, Faast R, Bower CC, Brown GR, Peakall R (2009) Implications of pollination by food
and sexual deception for pollinator specificity, fruit set, population genetics and conservation
of Caladenia (Orchidaceae). Aust J Bot 57:287–306
Powell JA (1992) Interrelationships of yuccas and yucca moths. Trends Ecol Evol 7:10–15
Sakai S (2002) A review of brood-site pollination mutualism: plants providing breeding sites for
their pollinators. J Plant Res 115:161–168
Shenoy M, Borges RM (2008) A novel mutualism between an ant-plant and its resident pollinator.
Naturwissenschaften 2008:61–65
Song B, Chen G, Stocklin J, Peng D-L, Niu Y, Li Z-M, Sun H (2014) A new pollinating seed-
consuming mutualism between Rheum nobile and a fly fungus gnat, Bradysia sp., involving
pollinator attraction by a specific floral compound. New Phytol 203:1109–1118
Ssymank A, Kearns CA, Pape T, Thompson FC (2008) Pollinating flies (Diptera): a major contri-
bution to plant diversity and agricultural production. Biodiversity 9:86–89
Steffan-Dewenter I, Klein A-M, Gaebele V, Alfert T, Tscharntke T (2006) Bee diversity and plant-
pollinator interactions in fragmented landscapes. In: Waser NM, Ollerton J (eds) Plant-
pollinator interactions from specialization to generalization. University of Chicago Press,
Chicago/London, pp 387–407
Svensson GP, Okamoto T, Kawakita A, Goto R, Kato M (2010) Chemical ecology of obligate pol-
lination mutualisms: testing the ‘private channel’ hypothesis in the Breynia-Epicephala asso-
ciation. New Phytol 186:995–1004
Tartaglia ES, Handel SN (2014) Nectar plant preferences of Hemaris (Sphingidae: Lepidoptera) on
co-occurring native Cirsium and non-native Centaurea (Asteraceae) inflorescences. J Pollin
Ecol 13:184–187
Terry LI, Walter GH, Donaldson JDS, Snow E, Forster PI, Machin PJ (2005) Pollination of
Australian Macrozamia cycads (Zamiacaeae): effectiveness and behavior of specialist vectors
in a dependent mutualism. Am J Bot 92:931–940
Triponez Y, Arrigo N, Espindola A, Alvarez N (2015) Decoupled post-glacial history in mutualistic
plant-insect interactions: insights from the yellow loosestrife (Lysimachia vulgaris) and its
associated oil-collecting bees (Macropis europaea and M. fulvipes). J Biogeogr 42:630–640
Vazquez DP, Aizen MA (2004) Asymmetric specialization: a pervasive feature of plant-pollinator
interactions. Ecology 85:1251–1257
Ware AB, Compton SG (1994a) Dispersal of adult female fig wasps. 1. Arrivals and departures.
Entomol Exp Appl 73:221–229
Ware AB, Compton SG (1994b) Dispersal of adult female fig wasps. 2. Movements between trees.
Entomol Exp Appl 73:231–238
Waser NM, Ollerton J (eds) (2006) Plant-pollinator interactions from specialization to generaliza-
tion. University of Chicago Press, Chicago/London
Waser NM, Chittka L, Price MV, Williams NM, Ollerton J (1996) Generalization in pollination
systems, and why it matters. Ecology 77:1043–1060
Weiblen GD (2002) How to be a fig wasp. Annu Rev Entomol 47:299–330
Westerbergh A (2004) An interaction between a specialized seed predator moth and its dioecious
host plant shifting from parasitism to mutualism. Oikos 105:564–574
Westerbergh A, Westerbergh J (2001) Interactions between seed predators/pollinators and their
host plants: a first step towards mutualism? Oikos 95:324–334
Wiklund C, Eriksson T, Lundberg H (1979) The wood white butterfly Leptidea sinapis and its
nectar plants: a case of mutualism or parasitism? Oikos 33:358–362
Woodcock TS, Larson BMH, Kevan PG, Inouye DW, Lunau K (2014) Flies and flowers II: floral
attractants and rewards. J Pollin Ecol 12:63–94
Chapter 4
Classic Themes: Ants, Plants and Fungi
Keywords Acacias • Ant gardens • Domatia • Epiphytes • Extrafloral nectaries •

Food bodies • Formicidae • Herbivory • Mutualism • Myrmecochory •
Myrmecophytes • Plant defences
4.1 Introduction: Ant-Plant Interactions
The interactions between ants and plants collectively display an enormous variety
of intricate and specific associations. As noted in the previous chapter, some ants
pollinate flowers, but their roles in that capacity are vastly overshadowed by the
predominance of other Hymenoptera. However, the variety of other interactions
between ants and plants has led to more being written on these themes than on any
other mutualism-related topics. Some of those interactions rank amongst the classic
mutualisms described. In his overview, Buckley (1982) listed six broad categories
of ant-plant interactions that he included as ‘mutualisms’. In addition to pollination,
these were associations with extrafloral nectaries, food bodies, domatia, ant-
epiphytes, ant gardens and seed dispersal. Each of these broad themes has received
considerable attention, and they collectively indicate the broad array of interdepen-
dences flowing from associations between different plant groups and different ant
taxa, with many interactions very specific. Key reviews include those by Janzen
(1966), Beattie (1985) and Rico-Gray and Oliveira (2007), the last being an ency-
clopaedic synthesis and reference source with numerous examples.
The varied mechanisms and roles present are a key element of considering ant-
plant mutualisms as model systems in ecology and evolutionary biology. Heil and
McKey (2003), for example, noted that these systems are tractable to experimental
study and manipulation as defences ‘worn on the outside’, in contrast to most chem-
ical defences developed by plants. Studies of ant-plant symbioses, including mutu-
alisms, emphasise the importance of extrafloral nectaries, domatia and the various
‘food bodies’ (p. 75) as key traits. Mayer et al. (2014) also emphasised the changing
conceptual framework that progressively involves evolutionary ecology and shifts
in how these relationships may be interpreted and noted the predominant themes in
which recent advances have been made as related to nutritional ecology, the influ-

DOI 10.1007/978-3-319-58292-4_4
64 4 Classic Themes: Ants, Plants and Fungi
Fig. 4.1 Ant-plant mutualisms, a scheme of the three main topics of evolutionary ecology in
which recent conceptual shifts have occurred and of the relationships between these (From Mayer
et al. 2014)
ences of global environmental changes and greater awareness of the full range of
costs and benefits, as well as how these may be modelled and interrelated (Fig. 4.1).
One caveat is necessary here: the names given for ant species over the next chap-
ters are those used in the specific accounts cited. In some cases, more recent taxo-
nomic studies may have led to nomenclatural changes or further delimitations
within species complexes. In some cases, this might entail different species being
referred to under the same name and the converse.
Writing generally on ant-plant interactions, Beattie (1985) opined that ‘it is little
wonder that ant assemblages are in large part fortuitous, interaction is diffuse and
facultative, and specialisation between particular ant and plant species an excep-
tional occurrence’. The more specialised mutualisms have evolved largely in
response to some form of ‘stress’, with ant protection of plants commonly consid-
ered a response to herbivores and seed predators (Beattie 1985): many studies (some
discussed below) have demonstrated that individual plants with patrolling ants and
resources for these to thrive suffer less herbivory, and can set more seed, than plants
deprived of their ant partners. Major benefits to plant fitness have been reported
repeatedly and attributed to ant mutualists. In extreme cases, the ants are vital to
plant survival, as Janzen (1966) demonstrated for Acacia cornigera occupied by
Pseudomyrmex ferruginea. Ten months after ants were removed from 69 of his 108
monitored trees, more than half of the ant-deprived trees had died through herbivore
4.1 Introduction: Ant-Plant Interactions 65
attack and overgrowth from neighbouring vines. Both these damaging trends are
countered by resident defending ants.
Heil (2010) categorised the ‘thousands of mutualisms’ between ants and plants
into the two main groups of ‘dispersal’ and ‘defence’. The first of these encom-
passed seed dispersal (myrmecochory, p. 87) and ant gardens (p. 85), and the
defence category incorporated extrafloral nectaries, food bodies and domatia – these
falling into groupings of ‘facultative defences’ (by providing food for foraging ants)
and ‘obligate defences’ (in which nest sites are also provided). Pollination was not
discussed by Heil, other than a brief allusion to its scarcity, but with the additional
relevant point that most ants are not hairy, so pollen might not easily adhere to them
as it does to strongly pilose bees and moths.
Ant protection mutualisms involve at least three species – the ant, the protected
species (plant or animal) and the threatening species. As Savage and Peterson (2007)
noted, this variety includes considerable potential for indirect effects to occur and to
confuse how those associations function. Facultative mutualisms between ants and
plants are very common defences against herbivores, and increased provision of
nectar is often associated with increased abundance of visiting ants. Such responses,
however, are very variable. For ants visiting Inga (Fabaceae) in Panama, Bixenmann
et al. (2011) suggested that it was adaptive for the plants to invest in nectar when
ants are present but that otherwise the costs of producing nectar would not be bal-
anced by the benefits of protection. Variable responses of nectar production to ants,
but not to herbivores, may be a favourable strategy for the plant. Disturbance can
affect the balance of mutualisms in many ways of influencing the supply of a reward
and the abundance of ants present. These may be from either ‘bottom-up’ effects
such as pruning or browsing of plants leading to increased attendance of both ants
and herbivores (Piovia-Scott 2011) or ‘top-down’ effects in which ant removals are
associated with increased herbivory but also allow alternative predators to increase
in density whilst unopposed by the ants.
The protective roles of ants have dominated studies of mutualisms between
insect defenders and plants, but a few intriguing parallels involving other insect
groups have also been reported. A small predatory mirid bug (Hemiptera) has been
implicated as a mutualist partner of the Neotropical shrub Piper urostachyum
(Piperaceae) (Hodson and Gastreich 2006), for example. Plants from which bugs
had been removed then suffered considerably more foliage loss than plants with
mirids. This undescribed bug is one of several predatory arthropods attracted to
Piper by presence of domatia and food bodies (both associated more usually with
ants), but is the most common of these, and has been observed feeding on the foliar
food bodies and a range of herbivorous insects.
Studies of specific mutualisms have indicated both the effectiveness of defence
against herbivory and the complexities of the mechanisms involved. A mutualism
between wild cotton (Gossypium thurberi, Malvaceae) and the predominant ant
visitor (Forelius pruinosus) in North America led to the ants reducing herbivore
damage, mostly caused by larvae of the cotton leaf-perforator moth (Bucculatrix
thurberiella, Lyonetiidae) (Rudgers 2002), and enhanced seed production. This

effect, however, arose in large part not from predation but from nonconsumptive
processes. The ants induced a variety of behavioural effects in the moth larvae –
when ants were absent, the larvae spent more time stationary (and potentially
feeding) and did not need to escape by dropping from the plants when disturbed.
Anti-predator behaviour of the larvae arose both from direct encounters with
Forelius and simply when ants walked onto the leaves on which they were feeding
(Rudgers et al. 2003).
Intriguingly, mutualist ants patrolling on their partner plants can be ‘regulated’
by some plants in relation to the likelihood of herbivory. Volatile compounds
released by damaged plants can attract the ants, with this form of communication
demonstrated in many such relationships (Blatrix and Mayer 2010). In Cameroon,
one of few attempts to identify the compounds involved (Vittecoq et al. 2011)
showed that two subspecies of Leonardoxa africana (Fabaceae, p. 93) occupied by
different mutualist ants (Petalomyrmex phylax, Aphomomyrmex afer) had different
signalling systems. P. phylax responded most strongly to methyl salicylate, but A.
afer only to hexanal, both chemicals commonly emitted by plants when foliage
damage occurred and, amongst other roles, attractive to some predators and parasit-
oids of the herbivores. Three taxa of Leonardoxa, the above and one non-
myrmecophytic taxon, expressed the same compounds, and existence of two
different signalling systems for these myrmecophyte-ants may represent local spe-
cialisation in these allopatric forms.
In part also using L. africana, Amsellem and McKey (2006) emphasised the
importance of considering plant phenology in assessing defences against herbi-
vores, with the pattern of leaf development seasonally affecting herbivore attack and
the behaviour of attending ants. P. phylax workers are much smaller than those of
the less specialised A. afer, and the restriction of predominant herbivores to young
leaves which can be protected by a smaller number of ants may also favour ‘defence
investment’ among many smaller ant individuals.
This system has also contributed to understanding evolutionary constraints of the
mutualisms. Pairwise coevolution between ants and plants is often postulated, but
rather few cases have been demonstrated convincingly. Possible relationships
between three myrmecophytic species of Leonardoxa and their specific ants are
related to the form of the plants’ prostomata, which function as specialised micro-
domatia harbouring the ants on the swollen twigs. Detailed measurements of the
prostomata, the specific associated ant taxa and the entrance holes excavated by the
ants inside the prostomata gave striking evidence of coevolution between size and
shape of prostomata and ants in one such association (Brouat et al. 2016). Dimensions
and shape of the prostomata effectively exclude ants that are larger than the mutual-
ist species but, clearly, will allow access to any species smaller or of similar size.
They do, however, provide an ‘exclusion filter’ that restricts the variety of ants that
might occur.
As one of the most ecologically pervasive and diverse insect groups, it is perhaps
surprising that ants do not have greater specialist roles as pollinators. Despite con-
4.1 Introduction: Ant-Plant Interactions 67
siderable attention to this, Buckley’s (1982) comment that ‘few plants are pollinated
by ants’ remains largely valid, with far greater importance of ants as nectar exploit-
ers. In particular, extrafloral nectaries play major roles in the ecology of some taxa.
Indeed, ant access to floral nectar that could sometimes increase their chances of
becoming passive pollinators has in some cases been thwarted by physical or
chemical deterrents. Nectar robbing by ants can often be severe and disrupt the
plant’s pollination system.
The root-parasite plant Cytinus hypocistis (Cytinaceae) grows in species of
Cistaceae in the Mediterranean area. The plants are visible above ground only dur-
ing the March–May flowering season, when clusters of inflorescences appear at
ground level on their host roots. Surveys of flower visitors revealed a predominance
of ants, with ten species together comprising 97.4% of a total 4638 floral visits
observed (de Vega et al. 2009). Ants are here the true pollinators and, whereas
Cytinus differs from several other case studies in that several ant species rather than
a single species are involved, de Vega et al. considered this to be a true mutualism.
Other cases of presumed ant-specific pollination occur – for example, pollination of
some Australian orchids by pseudocopulation with winged male ants. Males of a
‘bull-ant’, Myrmecia urens (which lacks metapleural glands, below), pseudocopu-
late with Leporella fimbriata and carry away pollen masses that can be deposited
later on the stigma of other flowers (Peakall et al. 1987), but most other ant pollina-
tion records are more generalised, and specialised pollination mutualisms involving
ants are indeed relatively scarce. A second confirmation of ant pollination of an
Australian orchid was of (in this case, flightless workers with limited dispersal
potential) Iridomyrmex gracilis pollinating Microtis parviflora (Peakall and Beattie
1991). Other ants (Rhytidoponera tasmaniensis, Meranops sp.) were also reported
pollinating this orchid, but I. gracilis was the most frequent pollen-carrying
species.
Buckley (1982) suggested that ants might have disadvantages as pollinators, for
four main reasons: (1) the fixed nests of ants, together with territoriality, might
mean that they do not carry pollen far enough for effective cross-pollination beyond
the parent plant; (2) most ants lack specificity and tend to visit any available nectar
source indiscriminately, rather than selecting particular species; (3) the constant
grooming and cleaning behaviour of many ants would remove pollen before it can
be transferred; and (4) the secretion by many ants of chemicals that inhibit pollen
development and germination – those chemicals secreted from metapleural glands
and evolved as antibacterial and antifungal measures – may extend their impacts to
pollen. They do not occur in all ants, however.
Many ants are thus nectar robbers but not pollinators so can be ‘aprovechados’
(Chap. 8) or parasites of more specialised mutualisms (Sachs 2015). Their presence
can also interfere with activities of legitimate pollinators, as noted by Cembrowski
et al. (2014) – whose ingenious experiments with artificial flowers showed that both
ants and their scents reduced levels of bumblebee pollination. Their study recog-
nised the twin impacts of ants as (1) consuming rewards sought by legitimate pol-
linators and (2) harassing those pollinators and disrupting their activities. Using
workers of Myrmica rubra (an invasive ant in North America) and the bumblebee
pollinator Bombus impatiens, the ants influenced bee behaviour by both direct
attack and by causing them to avoid flowers with ant scent. As indicated by transfer
of powdered food dyes (used as surrogate pollen) on the artificial flowers, bees
transferred significantly more ‘pollen’ to and from ant-free flowers, showing that
interference competition from ants could itself modify pollination behaviour.
4.2 Extrafloral Nectaries
Extrafloral nectaries that attract ants occur in numerous different plants (across
more than 90 families: Rudgers and Gardener 2004), and the wide basis for mutual-
isms is the ants gaining nectar as food whilst defending the nectary, or the entire
plant, against insect herbivores and other marauders, so reducing imposed damage.
Since Janzen (1966) first drew attention to the variety of such defences, from his
studies on the bull-horn acacia (Acacia cornigera) and the ant Pseudomyrmex fer-
ruginea in Mexico, numerous other cases have been found, with the ants often
showing what Way (1963) called ‘ownership behaviour’ by actively pursuing and
repelling intrusive herbivores. The bull-horn acacia-Pseudomyrmex interaction was
the first such to be determined experimentally as truly mutualistic – neither party
survived when the other was absent. For ants that reside on or in the plants, their
simple presence can be an effective deterrent, and all plant structures may be pro-
tected. That protective behaviour is also effective against potential nectar thieves, an
array of other insects that are attracted to the nectar sources.
The widespread occurrence of extrafloral nectaries, and that they have evolved
independently on numerous occasions in unrelated groups of plants, has suggested
that they have wide adaptive significance and, as well as providing rewards for
mutualistic protectors, other hypotheses have been raised in support of this. Those
alternatives, not mutually exclusive, include (1) distracting ants from flowers
(below); (2) distracting ants from Hemiptera (p. 108), luring ants away and so help-
ing to prevent them from inducing outbreaks of aphids or scale insects by allowing
attack by natural enemies; and (3) excretion of excess carbohydrate from the plants.
Reviewed by Koptur (2005), the ‘protection function’ still has far greater general
support as their primary role. However, the variety of extrafloral nectaries continues
to pose problems of interpreting their functions precisely – but with their impor-
tance in influencing the structure of plant communities through influencing activi-
ties of herbivores, predators and parasitoids recognised widely. Their roles in
mutualisms are an important stimulus for further investigations, but Marazzi et al.
(2013) emphasised that many aspects of their significance are poorly understood
and that ‘the extrafloral nectary puzzle remains an intriguing challenge for the
future’.
Two of the alternative hypotheses noted above relate directly to consumers tak-
ing floral nectar from the same species, as (1) rewards for legitimate pollinators and
4.2 Extrafloral Nectaries 69
(2) enticing illegitimate visitors, notably ants, away from floral nectaries, so dis-
tracting them from damaging flowers or interfering with legitimate pollinators – and
so associated with the harmful effects of ants as flower visitors. ‘Distraction’ has
two major components (Galen 2005) as (1) enticing ants from floral to extrafloral
nectaries and (2) ‘diluting’ the intensity and rates of ant visits through predator
satiation with surplus nectaries on the entire plant. Galen tested these experimen-
tally (using Formica rufibarbis gelida on Polemonium viscosum, Polemoniaceae),
with outcomes indicating that nectar in surplus flowers can satiate ants and reduce
their negative impacts on flower functions. The ‘distraction hypothesis’ in any form
implies that ants visited flowers before extrafloral nectaries evolved and their nega-
tive effects provided the selective pressures for those to be developed.
General ant activity fostered through provision of extrafloral nectar is widely
believed to primarily benefit the plants by consequent protection from herbivory but
might also have other roles. Thus, pollinators may be threatened or deterred by ants,
and decreased visitation rates by key pollinators might reduce the plant’s fitness by
reduced reproduction. Tests by Assuncao et al. (2014) using experimental manipu-
lations of ants feeding from nectaries of Heteropterys pteropetala (Malpighiaceae)
in Brazil encompassed ten ant species and five species of visiting bees and included
trials in which surrogate plastic ‘model’ ants (resembling naturally present
Ectatomminae) were one of the four treatments for comparison. Outcomes for fruit
production (Fig. 4.2) showed that bees could distinguish plastic ants from (control)
plastic circles and avoid those flowers significantly, but, overall, avoidance by pol-
linators caused by ants was not sufficient to negatively affect plant reproduction.
However, also in Brazil cerrado, the 15 species of ants visiting extrafloral nectaries
of Chamaecrista debilis (Caesalpiniaceae) were effective anti-herbivore agents,
their presence increasing fruit set and reducing foliage herbivory (Fig. 4.3) (do
Nascimento and Del-Claro 2010). Pollinators were not addressed specifically, but
17 visiting insect species were reported.
Extrafloral nectar is commonly the most important, occasionally the only, food
available for the ants on plants. However, the reciprocal protection postulated from
the ants is not universal, with some of the numerous studies stimulated through
Janzen’s pioneering work having failed to detect any deterrent effects of the ants on
herbivores. Nevertheless, protection has been validated for many plant taxa and
growth forms in different ecosystems and in many parts of the world. Most exam-
ples are from the tropics where, as reaffirmed by Rico-Gray and Oliveira (2007),
plants with extrafloral nectaries and associations with ants are reportedly more com-
mon than elsewhere. A wide range of generalised ‘facultative mutualisms’ between
ants and those plants has been postulated. Some experimental studies of the patterns
of ant visitations to co-occurring plants that differ in nectary features (such as struc-
ture, nectar composition and rate/amount of secretion) have revealed consistent dif-
ferences. These emphasise the need to appraise the associations in the local
community context, rather than as static interactions (Apple and Feener 2001), and
that plants may compensate for the protection offered by ants and other arthropods.
That study, on ants visiting three species of Passiflora in Costa Rica, built on an
Fig. 4.2 Interactions between pollinators, ants and extrafloral nectaries on Heteropterys pterop-
etala (Malpighiaceae), evaluated through the use of plastic ant models placed on flower petals.
Bees avoided visiting flowers with plastic ants, and those flowers produced far fewer fruits than
those without normal ant access or from which ants were excluded; (a) flower visitation pattern
(open, visited; black, not visited); (b) fruit production pattern (open, fruits produced; black, fruits
not produced) (Assuncao et al. 2014)
Fig. 4.3 Ants as anti-

herbivore agents on
Chamaecrista debilis in
Brazilian savanna: (a)
herbivory levels on plants
with (black) and without
(open) ants over 4 months;
(b) proportion of fruits
formed/buds produced
between plants with (black,
n = 10) and without (open,
n = 10) ants (do
Nascimento and Del-Claro
2010)
earlier survey of the same host species (Smiley 1978) in which the plants’ attractive-
ness to ants was characterised as, respectively, high, intermediate or low and inter-
preted by the principle that higher ant visitations could represent greater protection
from herbivory. Most ants found by Apple and Feener could be identified only to
genus level, but the visitors included ten genera, and, in general, the earlier correla-
tion of ant visits with extrafloral nectary structure was supported. In both studies,
ant visitation to Passiflora oerstedii was low, and the nectaries of that species were
very small. Such small nectaries might attract parasitoids rather than ants: Koptur
(1985) had earlier noted that parasitoids can be frequent visitors in places where
ants are relatively scarce.
Different between-species outcomes amongst acacia-ant-herbivore associations
also occur (Cronin 1998). In Costa Rica Acacia collinsii hosts three species of
Pseudomyrmex, but with any individual tree usually only harbouring one species.
Pseudomyrmex spinicola and P. nigrocinctus are rather similar in appearance, but
the latter removes more litter and plant material from around the base of its host, and
P. flavicornis completes the trio, whose activity patterns were compared to deter-
mine which might be the ‘best mutualist’ and how. The three ant species were very
similar in patrolling and post-disturbance activity. They respond to the presence of
herbivores at similar rates and inflict similar number of attacks. However, differ-
ences emerged when their daily activity patterns were assessed. The first two spe-
cies responded more rapidly and attacked more aggressively in the morning and P.
flavicornis in the afternoon (Fig. 4.4), and such differences might become relevant
in assessing defences against herbivores with different temporal activity patterns, by
both day and season.
Extrafloral nectaries occur on many parts of plants and range from simple glan-
dular areas to elaborate structures, for some of which the functional significance is
continually debated. Bentley (1977) noted that the two major supported groups of
suggested functions devolved on ‘protectionists’ (with the ants protecting the plant
from herbivore attack) and ‘exploitionists’ (viewing the plants as being used, and as
not using the ants for their own benefits). The first context is by far the more widely
supported. Most direct interactions between the nectaries and ants were inferred to
be ‘diffuse’ and unspecialised (Rico-Gray and Oliveira 2007), largely fortuitous and
facultative, so that in general most interactions lack any major level of specialisa-
tion, and obligate mutualisms are also relatively rare. Most studies have failed to
confirm that both the ants and the plants benefit significantly but, rather, have shown
that those plants attracting multiple ant species may benefit from those ants disrupt-
ing a correspondingly wider array of potential herbivores. Ants vary widely in their
defensive activities – some confine defences to the nectaries or their immediate
vicinity, whilst others patrol the plants, with wider deterrence of incursive insects.
In general, the plant’s energy investment in production of the nectaries is low rela-
tive to the benefits obtained through such defences.
The diversity of insects visiting extrafloral nectaries implies that they can play
important and wide-ranging roles in communities, extending well beyond the most
immediate mutualisms. Records of visitors to the nectaries summarised by Rudgers
and Gardener (2004) included herbivores, pollinators and parasitoids spanning at
Fig. 4.4 Activity patterns

of three species of
Pseudomyrmex ants on
Acacia collinsii in Costa
Rica in the morning and
afternoon. P spinicola/P.
nigrocinctus (squares), P.
flavicornis (triangles); (a)
pre-disturbance activity
(open symbols) and
post-disturbance activity
(solid symbols); (b) time
taken for an introduced
katydid to be first attacked:
(c) number of times/90 s
that a katydid was attacked
following attachment to an
ant-inhabited Acacia
(Cronin 1998)
least ten orders of arthropods, all of which presumably benefit in some way from the
nectar, notably carbohydrate, supply and potentially interact. One scenario dis-
cussed there is that habitats supplying extrafloral nectar might create ‘islands’ of
high arthropod diversity and abundance – in a principle resembling some imposed
aspects of habitat modification to foster natural enemies for conservation biological
control in agroecosystems (New 2005). There, resource enrichment is an important
tool in attracting natural enemies and concentrating numbers where their impacts
are needed. The spatial distribution of extrafloral nectar in communities could also
possibly prevent a single arthropod species from monopolising the resource
(Bluthgen et al. 2000), so promoting species coexistence and diversity.
Some of those other, and generally little-heeded, arthropods attracted to nectar
can also participate in plant defences against herbivores. Koptur (2005), for exam-
ple, cited spiders and wasps. Jumping spiders (Eris sp., Metaphidippus sp.,
Salticidae) chose individual plants of Chamaecrista nictitans with active extrafloral

nectaries over those with the nectaries artificially blocked (by covering them with
nail polish) and enhanced the level of seed set on those plants (Ruhren and Handel
1999). The spiders collected nectar as well as preying on flower visitors, so might
provide unexpected defences to those plants, at least by sporadic and facultative
association.
A somewhat different view of extrafloral nectaries was discussed by Becerra and
Venable (1989), in proposing that they are sometimes favoured even when ants do
not defend the plant from herbivores. Rather, they may help to defend the plants
against ant-Hemiptera mutualisms (Chap. 5). The basis for this viewpoint was that
the main selection pressure for the nectaries was not protection of the plant from
herbivores, as the most commonly accepted, but distraction of ants from tending
Hemiptera (Fig. 4.5), so thwarting ant defence of the Hemiptera and reducing the
damage they may cause to the plant, by their being attacked by natural enemies,
perhaps including the ants as predators rather than tenderers. This hypothesis flowed
from recognition that the ‘ant guard’ scenario had many exceptions – despite the
widespread defence roles in some obligate mutualisms, many others are facultative
and with very variable outcomes, in which lack of protection of Hemiptera by the
ants can lead to increased mortality and decreased fecundity. The ant-Hemiptera
mutualism is regarded by some authors as ‘a mixed blessing’ (after Carroll and
Janzen 1973).
Only a small proportion of the total possible component species may contribute
to the ‘active core’ of ant-plant networks in which the major resource is extrafloral
nectar (Dattilo et al. 2014, for Brazil). That core may comprise mainly the competi-
tively superior species capable of excluding others, which are thereby directed
toward less favourable resources. The major core genera all exhibited high levels of
recruitment so that they both dominated and outnumbered other ants. They also
often show adaptations to feeding on liquid resources (Dattilo et al. 2014).
Ant protection of plants with extrafloral nectaries has thus provided a suite of
systems, from obligate to facultative mutualisms, that have proved rewarding to
study (Lanan and Bronstein 2013). Most such plants interact with several omnivo-
rous ant species which differ in their responsiveness to the plant, the effectiveness
of defence against herbivores and their potential impacts on pollinators (Ness 2006).
The identity of the ‘best ant mutualist’ is likely to change in time and space, to
reflect the changing balance and importance of pollination and protection from her-
bivory. Varying qualities of extrafloral nectar can influence levels of aggressiveness
of mutualist ants, for example (Ness et al. 2009). Because ant colonies may forage
across many plants simultaneously, variations in the resources at one plant can
cause changes in foraging activity that affect other plants. Studies of colonies of
Crematogaster opuntiae tending the cactus Ferocactus wislizeni showed that work-
ers from large colonies visited 5–34 nectar-secreting cacti within their territory,
leading Lanan and Bronstein (2013) to highlight the importance of the ant colony as
relevant for study in appraising ant-plant interactions.
Fig. 4.5 The effects of changes that can occur and influence plant fitness amongst factors consid-
ered in (a) the ‘ant-guard hypothesis’ and (b) the ‘ant-distraction hypothesis’ of how extrafloral
nectaries may evolve for defence against ant-Hemiptera associations. ‘Fitness’ is indicated as WS,
a base fitness for plants without herbivores, ants or extrafloral nectaries; WW, fitness with herbi-
vores but without ants or nectaries; WN, fitness with ants; WA, fitness with nectaries and ants.
Arrows indicate changes in fitness due to indicated factors, comparing plants with or without nec-
taries and ants (after Becerra and Venable 1989)
4.3 Food Bodies and Domatia 75
4.3 Food Bodies and Domatia
The ‘space’ (such as cavities and hollows) available to resident ants for nesting on
plants can impose upper limits on the size of the colony that can be supported, in
conjunction with the food resources (both direct, as food bodies, below, and indi-
rect, as derived from other insects) needed by the colony. The model derived by
Fonseca (1993) illustrates this relativity with the assumption that the cost to the
plant may increase directly to support larger ant colonies (p. 97, Fig. 4.12).
Simultaneously, defence of the plant increases with the number of patrolling worker
ants, but, as the number increases, the importance of each individual ant in defence
against herbivores decreases – so that the benefit to the plant reaches a limit. At the
intersection point between ‘cost’ and ‘benefit’, the functional emphasis changes
from mutualism to the ant being parasitic. At some earlier point, a smaller colony
size than this maximum, benefit will be greatest in relation to the costs incurred.
Resources that enable ants to live on or in a plant, or forage there and promote
defence of the plant, have been provided in several ways. Plants may produce spe-
cialised ‘food bodies’ other than extrafloral nectaries and which are highly attractive
sources of nutrients for the ants, which may also be enabled (or encouraged) to nest
near this food supply. Defences of the plant may correspondingly be extended from
just against herbivores to also thwarting plant competitors that, by growing vigor-
ously nearby, may affect the plant and its long-term capability to sustain ant colo-
nies. These associations are wholly tropical or subtropical. Variations in effectiveness
of defence can be substantial when facultative or opportunistic associations occur,
and the ants are not resident. If the latter (as on the nectary-bearing vine Dioscorea
prehensilis, Dioscoreaceae: Di Giusto et al. 2001), ants must be recruited by the
plant, and defence against herbivores (in this example, the chrysomelid beetle
Lilioceris latipennis) differed considerably over 3 years of survey, as well as reflect-
ing the relative susceptibility of different beetle larval instars. Ant attacks on larger
larvae were usually unsuccessful. Presence of ants also reflected the likelihood of
immediate rewards, in this case of extrafloral nectar.
Several of the classic mutualisms based on these associations are recapitulated
here briefly. Each has been described and discussed fully and repeatedly elsewhere,
but they are fundamental in demonstrating parallels in development of mutualisms
across widely different plant groups, geographical regions and ant taxa.
1. Acacia-Pseudomyrmex associations in the neotropics have attracted wide atten-
tion since their initial characterisation by Janzen (1966, 1967). The specialised
swollen thorns of the acacias (Mimosaceae) harbour the ant nests, established
there as founding queens cut an entrance hole in new or empty thorns and, if
necessary, excavate the inner parenchyma to construct a nest chamber. Ants feed
on extrafloral nectar, and their brood is fed on specialised foliar structures,
Beltian bodies, whose presence may be a prerequisite for queen ants to produce
eggs. Ant colonies can grow until all available thorns – on one or a group of
adjacent trees – are occupied. They can thereby become large – Janzen reported
a single colony of Pseudomyrmex belti in Mexico to occupy 100 trees.
2. Cecropia-Azteca associations are also neotropical, but differ from the acacias
noted above, in that the plants (Cecropiaceae) lack any extrafloral nectaries. The
ants nest in the hollow internodes, entered through the stem walls, and feed on
specialised food bodies – here, glycogen-rich Mullerian bodies (Janzen 1969).
At least ten species of Azteca ants have obligate relationships in this way with
Cecropia and defend the plant in similar ways to Pseudomyrmex on Acacia. All
myrmecophytic (‘ant-supporting’) Cecropia species give similar benefits to the
ants.
3. Macaranga-Crematogaster associations in Southeast Asia also involve ants
nesting within hollowed stems, and the plants (Euphorbiaceae) provide food as
‘Beccarian bodies’. The nesting sites, specialised food bodies and extrafloral
nectaries are all provided by the plants whether or not ants are present.
In addition, another group of food bodies (‘pearl bodies’), which are often asso-
ciated with extrafloral nectaries, are produced by many different tropical and sub-
tropical plants and utilised by a number of different ant genera.
The various ‘food bodies’ noted above differ substantially in primary nutritive
value. From the above examples, which encompass the major categories that have
attracted attention in mutualisms, Beltian bodies contain large amounts of protein,
as well as lipids and carbohydrates; Mullerian bodies contain high levels of glyco-
gen and lipids; Beccarian bodies are especially rich in lipids; and pearl bodies are
also rich in lipids but with those on different plant taxa varying considerably (Rico-
Gray and Oliveira 2007). The bodies may be adapted to meeting special nutritional
requirements and attractiveness to the ants, as principles that have been investigated
most thoroughly for Macaranga (Heil et al. 1998), in which only some species form
regular associations with ants. In this genus, the food bodies of myrmecophytic spe-
cies provide large amounts of lipids, proteins and soluble carbohydrates, and those
of non-myrmecophytic species have rather different sugars in the soluble carbohy-
drate fractions. These are adapted to their specific roles in ant attraction and nutri-
tion. Whilst all Macaranga species with food bodies are presumed to maintain or
attract ants for defence, the differences between myrmecophytic and non-
myrmecophytic taxa indicate the additional subtleties across species with differing
forms of association. Further, the intensity of ant defence differs predictably across
sympatric species of myrmecophytes (Itioka et al. 2000). Macaranga species show
a gradient from obligate myrmecophytes, through facultative or transitional
myrmecophytes to non-myrmecophytes. Itioka et al. noted that (1) plants may ben-
efit more from ant defences if they provide more food for them, because (2) ant
colonies receiving more food may produce more workers and so (3) increase effec-
tiveness of defence through greater number of defenders. However, a higher invest-
ment in defence may divert resources from other aspects of plant wellbeing, such as
growth and reproduction, so that the various trade-offs between investment in
defence and other processes are likely to differ amongst species. The species of
Macaranga studied were considered likely to have different counter-herbivore strat-
egies. In comparing three co-occurring pioneer species in Malaysia, Itioka et al.
(2000) confirmed such differences, with the three species differing also in their
herbivore complements; two suffered most of their damage from generalist polyph-
agous insects such as grasshoppers, but the third incurred more damage from a
specialist gall-making fly.
Macaranga spp. have many and varied associations with ants, spanning obligate
mutualisms to much more sporadic and facultative interactions. They range from
myrmecophytic species with domatia to those that attract ants through extrafloral
nectar and food bodies, and the roles of ants in protection against herbivory vary
correspondingly. A number of Macaranga species are considered ‘transitional’ in
that when young they are visited by different ant species but are later colonised by
an obligate residential mutualist. Study of one such species, M. hosei, in Malaysia
(Fiala et al. 1994) showed that the plants benefitted from early facultative associa-
tions with different ant species (10 visiting species, compared with 16 on M. tanar-
ius, a non-myrmecophyte used for comparison, and with a pool of 21 ant species
across both plants) that collectively protected the juvenile plants against herbivores.
As stem diameter increased sufficiently to allow nest construction, extrafloral nec-
tary production stopped and led to a much more effective protection as the obligate
Crematogaster partner established.
A similar conclusion of greater protection against herbivores in obligate mutual-
isms amongst Macaranga was reached by Heil et al. (2001), who noted the need for
long-term studies rather than those confined to short periods, which do not always
give a realistic picture of the functional associations. Both the ‘obligate’ species
studied (M. hosei, M. triloba) were compared with and without the mutualist ants,
both on whole trees and on individual isolated branches. Within a year, the ant-free
plants of both species had lost high proportions of their foliage area (80% in M.
triloba, 70% in M. hosei), a clear indication of the need for ant mutualists. Short–
term studies gave wholly inadequate estimates of this longer-term outcome, in sug-
gesting only low levels of protection by the ants. Many characteristic functional
traits of obligate ants do not occur amongst facultative species on non-myrmecophytic
plants, which have no need to prune competing plants or extensively clean plant
surfaces.
Taxonomic subtleties abound in ant-myrmecophyte associations, with some
early suggestions of ant species being relative generalists in some cases giving way
to recognition that these each constitute a separate lineage. Two of the four
Crematogaster ants on Macaranga studied by Itino et al. (2001) each associate with
several Macaranga species, in contrast to the other two species occurring on a
single tree species. Phylogenetic studies using mtDNA to clarify lineages implied
that each of the ‘generalists’ comprised two or more distinct genetic lineages.
Further interpretation implied that the two partners have co-speciated and co-diver-
sified over the past 6–7 million years, a process facilitated by the long-term stability
of the Southeast Asian tropical forests. The evolution of Crematogaster-Macaranga
mutualisms, elucidated by phylogenetic analyses based on CO1 reconstructions
(Quek et al. 2004), indicated that host shifts by ants have been constrained by stem
features that affected their ability to nest within the plants and that each of the two
Macaranga clades involved have separate origins for their myrmecophily. The
mutualistic habit may have enabled Macaranga to exploit enemy-free space, whilst
Fig. 4.6 Outline map of Southeast Asia to indicate circumscribed area of distribution (dashed
line) of species of Neonauclea that support resident ants by development of hollow chambers
(‘myrmedomes’) in internodes (Ridsdale 1989)
ant diversification has been fostered by stem texture, host specialisation and local
geography.
In comparison with the greater diversification of ant-plant mutualisms in tropical
America, this is one of only two such major radiations to have occurred in Southeast
Asia, the other being with Neonauclea (Rubiaceae) (Ridsdale 1989). The region
supports only three species-rich genera of myrmecophytic ants, Crematogaster,
Camponotus and Cladomyrma. Persistence of the mutualisms during speciation
may be fostered by the lack of alternative taxa that could be exploited, as hypothe-
sised by Davidson and McKey (1993). Ants associate with many different Rubiaceae
(below), with Neonauclea one of the predominant genera of myrmecophytes in the
region. Ridsdale (1989) noted 17 of the 65 species as myrmecophytes, with the 45
non-myrmecophytic species collectively having a much wider geographical range
(Fig. 4.6). In Borneo, several independent colonisations by Cladomyrma form these
mutualisms, with the three major endemic ant species varying in specificity and host
range (Razafimandimbison et al. 2005). It has been claimed that Cladomyrma in
Malaysia may be entirely myrmecophytic (Maschwitz et al. 1991), with the found-
ing ant queens chewing holes in the plant internodes and hollowing out stems to
form nest chambers.
Ants colonise myrmecophytic Macaranga in the early, seedling, stages of devel-
opment, after which their protection reduces damage from folivores. In contrast,
ant-free trees suffer high damage levels as seedlings, but this declines as the trees
grow – a situation explored by Itino and Itioka (2001) to show that coexisting spe-
cies of Macaranga can differ markedly in the form of ant-herbivore defence and
whether this is primarily by ants or by secondary plant compounds. In addition,
Fig. 4.7 The phasmid

Orthomeria cuprinus and
Crematogaster ant
defences on
myrmecophytic
Macaranga in Borneo. The
number of ants found in
trees with and without
phasmids: data shown are
for number of ant workers
initially present (open) and
maximum numbers
reached (black) (Shimizu-
kaya and Itioka 2016)
some ecological properties of the plants – such as crown light levels and the position
of ant food bodies – are partially correlated with levels of herbivory. Individual
host-specific ant mutualists may deter herbivores that are adapted to either chemical
or structural defences, but, within the complex arrangements of the Macaranga-ant-
herbivore associations, mutual counter-adaptations have occurred both between
plants and herbivores and between herbivores and ants (Itino and Itioka 2001).
Particular herbivore species may, at least partially, evolve features that overcome the
defences imposed by myrmecophytic ants.
In Borneo, the winged phasmid Orthomeria cuprinus is a specific feeder on two
myrmecophytic Macaranga species having mutualisms with Crematogaster ants.
Phasmids forage mainly on young leaves, and their presence led to ants walking
faster; phasmids then walk away from approaching ants, and nymphs sometimes
‘swing’ legs if touched. A low intensity of ant defences (activity) may be needed for
phasmids to settle, and ants responded strongly to any introduction of a phasmid
nymph (Shimizu-kaya and Itioka 2016). Experimental trials implied that ant
defences on noninfested plants prevented the phasmids from staying on the plants.
Markedly greater number of ants occurred on noninfested trees than on trees with
Orthomeria (Fig. 4.7), on which the number of plant-ants on young leaves immedi-
ately before introduction of a phasmid and the later maximum numbers were
recorded over an hour of observation.
Intriguingly, food bodies are in some cases produced only when the plants are
occupied by ants. The classic example is of Piper cenocladum (Piperaceae), which
produces them only when occupied by Pheidole bicornis. If the ants are removed,
food body production falls greatly but resumes if the ants are reintroduced (Risch
and Rickson 1981), seemingly as a direct response to the ant being present. In Costa
Rica, the ants feed on the food bodies provided inside the petiole cavities in which
they nest, and the ants are effective defences against herbivores. In addition to stim-
ulating food body production on Piper, the ants also appeared to keep the food
source free from fungi – in plants without ants, existing food bodies were com-
monly covered with fungal hyphae.
Fig. 4.8 The trophic

interactions among Piper
cenocladum, its insect
herbivores (mostly
coleopterans or
lepidopterans), the resident
mutualist ants (Pheidole)
and the resident ant-
predator beetle
Tarsobaenus letourneauae.
Arrows indicate positive
contribution to biomass of
individuals; bulletheads
indicate interactions
resulting in mortality of
individuals; line thickness
indicates importance of the
interactions (based on
study in Costa Rica: Dyer
and Letourneau 1999)
This system was investigated in more detail by Dyer and Letourneau (1999), to
incorporate relationships amongst P. cenocladum and its associated arthropods.
More complex interactions across trophic levels within the community associated
with Piper shrubs inhabited by Pheidole are summarised in Fig. 4.8. Pheidole
bicornis harvests food bodies and provides substantial defence against herbivores
and intrusive vines. Larvae of the predatory clerid beetle Tarsobaenus letourneauae
live in the hollow petioles, where they feed both on the food bodies and the ant
brood, and also kill adult ants. Major outcomes from that manipulative study sup-
ported the structure being a ‘top-down’ trophic cascade whereby the herbivores are
limited by their predators but the plants are not limited by their herbivores. There
was no evidence of indirect effects of plant biomass on higher level predators, but
the ants indeed had indirect effects on plant biomass through decreasing herbivory.
In some Cecropia species, debris deposited by ants in their nest chambers can be
an important nutrition source for the tree. For example, ant-occupied C. peltata
receive 93% of their nitrogen in this way (Sagers et al. 2000), as a source that had
largely been overlooked previously. Stable isotope analyses showed that ants were
providing far more nutrients than they received from the plants and would substan-
tially reduce the ‘costs’ to the plant of sustaining the ants. Whilst the plants can
potentially exploit ants for their nitrogen, this – when derived from frass and ant
corpses – would not be any cost to the ants. In the absence of mutualistic ants, other
species of Cecropia compensate by increasing other, more classical, forms of
defence involving increased chemical defence in young foliage and physical tough-
ness of older leaves (Latteman et al. 2014). Production of Mullerian bodies to sup-
port ants in C. membranacea may constitute a trade-off between biological and
chemical defences, with increased ant presence and abundance lessening need for
chemical defences. Some species of Cecropia are presumed to have lost an earlier
mutualism with ants but still produce Mullerian bodies from the specialised glandu-
lar areas (trichilia) from which those structures develop.
The combination of plants providing both food for ants and a place for them to
live fosters close and specific relationships, and the reciprocal defences provided by
the ants can range from simple patrolling as deterrents to vigorous aggressive
counter-attacks. Thus, Janzen (1966) noted that Pseudomyrmex attack all other
insects on the swollen-thorn acacias and give the plants considerable protection
from herbivores (by comparison, acacias without ants lose foliage and growing tips)
and competitors. The ants mutilate vines and other plants that grow into contact
with their host acacia and also clear an area of ground around its base. Both mea-
sures benefit the acacia by reducing competition from neighbours, and it has also
been suggested that the cleared annulus might also serve as a firebreak (Bentley
1977). The ants may seek nectar on other plants from their host, and this otherwise
obligate association seems to have evolved at least twice, involving different species
groups of acacias and ants.
The ecological complexities that can arise are illustrated well through another
ant-Acacia mutualism, in East Africa. Acacia drepanolobium associates with sev-
eral different ant species, but each tree is occupied by only one of four mutually
exclusive mutualistic ant species. Those ants have rather different interactive roles.
Only Crematogaster mimosae is a strong mutualist, attacking browsing insects and
removing herbivores. C. nigriceps is also a strong defender of the trees against her-
bivores, but in addition it eats many of the tree’s shoots. The others, C. sjostedti and
Tetramorium penzigi, offer little if any mutualistic benefit. Photosynthetic rates of
trees attended by the first two ant species above were about 50% greater than those
with the nonmutualist ants (King and Caylor 2010), giving a tangible demonstration
of the benefits of protection from herbivory. Two distinct ways by which this protec-
tion might be conferred were advanced, as (1) the far more aggressive swarming of
the two mutualist ants to disturbed branches and its positive correlation with deter-
rence of browsers constitutes ant behaviour that should reduce rates of herbivore
damage directly, and (2) by enabling increased photosynthetic rates, presence of
those strongly mutualist ants can contribute to defence against herbivory also by
providing greater resources to the plant for its possible use in defence or replace-
ment of lost tissue. Recognition of these options could change perspective more
generally on the roles of biotic interactions in plant defence, as summarised in
Fig. 4.9.
The two predominant mutualists thereby confer considerable protection of A.
drepanolobium, but other consequences include (1) Crematogaster nigriceps can
temporarily sterilise the tree because it does not defend the reproductive structures
(Palmer and Brody 2007) and (2) C. sjostedti can facilitate invasion, by boring bee-
tles, and weaken the tree (Palmer et al. 2008). The differences in ant recruitment to
defend new foliage or fruits after disturbance were substantial for all four species,
but especially pronounced for C. mimosae (Fig. 4.10), indicating a rarely proven
subtlety of differentiation of mutualist defence across plant structures with different
primary roles. Effects on other potentially invading taxa also occur: the level of
infestation by a bud-galling midge (a new species of Acacidiplosis, Cecidomyiidae)
was associated closely with the species of ant present and was much higher on trees
without C. mimosae (Schumer et al. 2013). However, the level of herbivore defence
Fig. 4.9 Evaluating biotic interactions and plant defence: (a) a prevailing paradigm in which
effects of herbivores and indirect defenders (mutualists) are recognised as affecting plant alloca-
tion of resources to defence, but not the primary metabolic acquisition of those resources; (b) a
revised paradigm that includes critical effects of herbivores and mutualists also on rates of carbon
assimilation, subsequently changing the pools of resources available for allocation to defence traits
(King and Caylor 2010)
by each ant species predicted only poorly the net impact on the trees. Experimental
exclusions of each of the four ant species from A. drepanolobium trees (Stanton and
Palmer 2011) showed that (1) removal of each Crematogaster species, despite their
impacts on herbivore presence, led to considerable gains in tree ‘fitness’ as growth
Fig. 4.10 Effects of

simulated disturbance on
activity of symbiont ants
on Acacia drepanolobium
in Kenya: (a) the number
of worker ants swarming
aggressively in response to
simulated browsing of
branch tips of host trees
occupied by each species;
(b) the number of worker
ants recruiting to new
leaves (black) versus fruits
(open) following
disturbance. Ant species
from left to right are (Cs)
Crematogaster sjostedti,
(Cm) C. mimosae, (Cn) C.
nigriceps and (Tp)
Tetraponera penzigi
(Palmer and Brody 2007)
and/or reproduction and (2) the low cost of Tetramorium penzigi maintenance elimi-
nated any negative impacts on tree growth and increased new growth. Acacia drepa-
nolobium was thus highly tolerant of insect herbivory, and costs of maintaining
mutualistic ants exceeded any benefit that ensued. Those costs include production
of extrafloral nectar and the tending of pollen-feeding hemipterans by the ants. At a
rather different scale, occasional (perhaps, catastrophic) browsing by elephants
might be influenced strongly by aggressive mutualist ants, with other (non-
myrmecophytic) acacias more susceptible than A. drepanolobium to elephant attack.
Indeed, Stanton and Palmer (2011) suggested that elephants could act as selective
agents to promote the high plant investment that myrmecophytes need to maintain
aggressive ant mutualists. Exclusion of large herbivores (all those >15 Kg) from
plots in a Kenya savanna over a decade (Palmer et al. 2008) led to trees decreasing
their investment in supporting mutualistic ants. Both active nectaries (food) and
swollen thorns (domatia) were reduced, and absence of large herbivores may lead to
a breakdown in the mutualisms.
Ant domatia, usually derived from structures formed as part of the plant’s normal
growth, can occasionally be induced by the ants, in addition to their more common
modifications to existing structures. This was claimed, for example, for the neo-
tropical rainforest tree Vochysia vismiaefolia (Vochysiaceae), in which
Pseudomyrmex workers were observed initiating domatia by biting holes in twig
internodes and removing inner pith to create cavities (Bluthgen and Wesenberg
2001). In this plant, those holes could subsequently house a variety of different ant
species. Availability of domatia, however, does not necessarily engender
mutualisms. Some species of Piper (p. 65) that produce domatia only on leaf sur-
faces (rather than in hollow petioles) as well as food bodies are not known to have
mutualisms (Gastreich and Gentry 2004). Extensive investigations of the four spe-
cies of Piper with hollow petioles and interacting with Pheidole ants (notably, P.
bicornis) showed the expected reciprocal benefits of food/protection. Contrasting
the two categories of Piper (P. obliquum, mutualism; P. urystachyum, no obligate
mutualism) showed that the latter formed mutualisms with omnivorous mirid bugs
that fed on both the plant and on insects and for which increasing density was asso-
ciated with decreased loss of leaf area (Gastreich and Gentry 2004). However, the
experimental information needed to confirm whether this is indeed a causative asso-
ciation is still not available. Influences from co-occurring spiders on both Piper
species emphasised the varying roles of the embedding community in any such
mutualisms – in this case, as predators or disruptors of ant behaviour. Availability of
shelter from the ants also affected impacts.
4.4 Ant-Epiphytes
A number of plant genera are characterised as ‘ant-epiphytes’, as growing on trees

and providing nesting cavities in which the ants appear to be predominantly nutrient
providers rather than defenders. The ants are typically scavengers rather than
aggressive defenders. They accumulate food and waste materials in the cavities in
the host plants, where those materials are progressively broken down and used by
the plant. The symbioses are based on the reciprocal benefits of dry enclosed nesting
sites for ants in hollow foliage or stem structures and nutrient provision for the
plants. At least some of the plants have been confirmed to absorb nutrients from this
store, and, as for other ant-plant associations, these relationships have attracted con-
siderable attention.
From Huxley’s (1980, 1982) listing of plant species involved, the Southeast
Asia-Australia region is a major centre for development of this association, and the
habit was reported first for Southeast Asian Rubiaceae (Janzen 1974), in which
separate plant chambers can be used for nesting and/or for prey storage by the ants.
More than 160 species of Rubiaceae develop domatia, irrespective of the presence
of ants. The major genera of plants involved include Hydnophytum, Myrmecodia
(Rubiaceae) and Dischidia (Asclepiadaceae), the last also with many non-
myrmecophytic species. Fewer species of ferns and Bromeliaceae also harbour ants,
but this habit is much more sporadic and generally less prominent in these – for
example, Huxley cited Tillandsia (Bromeliaceae, tropical America) as containing 4
ant species and 400 non-ant species.
More recently, markedly increased knowledge of ant-plant nutritional mutual-
isms has extended this to other taxa, and modified the original scenario, as inter-
preted by Janzen (1974). Some shrubs and trees, in addition to epiphytes associated
with low-quality soils, also obtain nutrients from ant-accumulated debris, to demon-
strate that ‘ant-fed plants are a varied and complex group in terms of their evolution’
(Rico-Gray and Oliveira 2007).
4.5 Ant Gardens 85
4.5 Ant Gardens
Associations between ants and epiphytes can lead to the most complex extensions
of that form of mutualism, in which the detritus accumulation within the host plant
chambers is also supplied with epiphyte seeds that subsequently grow there – with
early suggestions that the roots could strengthen the ant nest chamber walls and
with implications of deliberate purpose by the ants later refuted in favour of simple
opportunism by them. This habit, creation of ‘ant gardens’, is restricted to ants that
form nests of ‘carton’ in tropical forests, and both the ants and plants involved are
diverse. Collectively, the ants are also ecologically diverse, with many levels of
mutualism involved, and, as Rico-Gray and Oliveira (2007) noted, the term ‘ant
gardens’ applies to some of the most specialised associations between ants and epi-
phytes, although sometimes with more than one ant species nesting in the same
plant and partitioning the space available. Projected advantages of this system may
also include (1) ability to build larger nests than could be made without the epi-
phyte, (2) these being protected from heavy rains and (3) building in more exposed
environments where more food may be available (Yu 1994). Experimental manipu-
lations by Yu in Peru showed that epiphytes indeed contributed to the structural
strength of the ant gardens, suggesting that an obligate association between epi-
phytes and the garden ants could prevent the carton nests from destruction by heavy
rains.
Ant gardens link strongly to myrmecochory (p. 87), and many of them are formed
by direct dispersal of epiphyte seeds to the ant nests, in which the ant larvae feed on
seed attachments. In Peru, Davidson (1988) noted that ants can also obtain seed
directly from the faeces of vertebrate fruit dispersers. They also ‘enriched’ nest
gardens with those faeces. In that survey, transcending a number of different forest
habitats in the region, ten epiphytes (representing seven families, including both
monocotyledons and dicotyledons) were established almost solely on ant gardens;
two others also probably fell into this category but were found only in very small
numbers, too few to confirm this allocation. Species sometimes grew together in the
same garden. That diversity also encompassed a substantial variety of ants –
Davidson and McKey (1993) noted that 5 of the 12 ant subfamilies were repre-
sented, with the habit absent only amongst some subfamilies of predatory or
legionary ant groups and from the two tiny monotypic subfamilies (Aneuretinae and
Nothomyrmeciinae) (Table 4.1). All five subfamilies they reported included species
with obligate associations with epiphytes. Considerable flexibility exists in other
interactions, with some ants associated with several different epiphytes, and these
hosts persisting in association with any of several ant species. Some associations are
more frequent than others, and development of these may reflect the outcome of
competition amongst the contending ant taxa. Rapid and efficient host location by
queens may confer considerable advantage by ‘priority’.
Mutualistic ants that protect the plant from herbivores in ant gardens may also
destroy flowers and reduce reproductive output. This extreme conflict (Yu and
Pierce 1998, sometimes termed ‘plant castration’) may be favoured by natural
Table 4.1 The worldwide taxonomic distribution of ants with specialised and regular symbiotic
associations with plants
Subfamily/tribe Major genera involved
Ponerinae Pachycondyla
Pseudomyrmecinae Myrcidris, Pseudomyrmex, Tetraponera
Myrmicinae
Cephalotini Zacryptocerus
Solenopsidini Allomerus, Solenopsis, Monomorium
Crematogastrini Crematogaster
Pheidolini Pheidole
Leptothoracini Leptothorax
Dacetini Strumigenys
(Unspecified) Atopomyrmex, Cataulacus, Podomyrma, Tetramorium,
Wasmannia
Dolichoderinae
Tapinomini Anonychomyrma, Axinidris, Azteca, Tapinoma,
Technomyrmex
Formicinae
Myrmelachistini Aphomomyrmex, Cladomyrma, Myrmelachista,
Petalomyrmex
Plagiolepidini Plagiolepis
Camponotini Camponotus s.l., Polyrhachis
Davidson and McKey (1993)
s election but mediated through mechanisms such as plant chemicals acting against
ant attack of flowers (Ona and Lachmann 2011). Conflict reduction can become
complex and involve aspects of ant aggression and dispersal or, even, of the attack-
ing ants comprising both discriminating and nondiscriminating cohorts – the former
avoiding attacking pollinators and the latter attacking both pollinators and
herbivores.
The evocative term ‘devil’s gardens’ is sometimes used for a specialised form of
ant mutualism in the Amazonian region, involving the association between
Myrmelachista schumanni and Duroia hirsuta (Rubiaceae) (p. 24). The ants kill
neighbouring plants, using formic acid, with the consequence of largely monospe-
cific stands of Duroia, groups of which occur (as ‘devil’s gardens’ or ‘supay
chakras’, the latter term being the name given by Andean peoples in Peru for the
clearings created: Edwards et al. 2009) in areas largely devoid of other plants
(Frederickson et al. 2005, Frederickson and Gordon 2007). Gardens can reach up to
600 or more trees, and a single garden can occupy >1000 m2. Each garden is occu-
pied by a single polygynous ant colony that can contain up to about three million
workers and 15,000 queens. Although Myrmelachista reduces herbivory on the
plants, the nearly pure stands of Duroia created as devil’s gardens suffer higher
levels of herbivory, reflecting the high plant density and freedom from encroach-
ment by other plants. The behaviour of Myrmelachista in creating ‘gaps’ around
Duroia improved light conditions for the trees and helped to secure the supply of
4.6 Myrmecochory 87
new domatia needed for the ant colony to expand (Oleson et al. 2002). Colony size
is limited by nest space, and M. schumanni actively increases nest sites by causing
swellings (galls: Edwards et al. 2009) on multiple plant species, and that become
domatia. Direct measurements of herbivory on plants in and outside the gardens,
and with or without ants, gave estimates of annual leaf loss to herbivores of 16% (in
gardens, with ants), 43% (in gardens, without ants) and only 5.5% (outside gardens,
where surrounded by a diverse rainforest understorey). Frederickson and Gordon
(2007) noted also that herbivore pressure increased as the garden grows, with
increasing lack of interspecific competition with Duroia from other plants, leading
to clear correspondence with the ‘resource concentration’ and ‘enemies’ hypotheses
of Root (1973).
4.6 Myrmecochory
The phenomenon of myrmecochory, which involves dispersal of seed by ants, also

depends on a nutritional reward system proffered by the plant. The mutualisms are
thereby based on nutritional benefit for ant vectors transporting seed away from the
vicinity of the parent plant – so with clear analogies to pollination, in which ants are
relatively minor components (Chap. 3). Myrmecochory is associated fundamentally
with ‘elaiosomes’, as attractive to ants and nutritionally rewarding structures
attached to seeds and used as ‘handles’ by which ants can carry the bulky seeds
away without damaging them. Within the ant nest, the elaiosomes are detached and
used as larval food and the seeds discarded either in or near the nest, where they can
germinate.
Myrmecochory is sometimes thought of as a ‘diffuse mutualism’ involving
guilds of plants and ants in which the basic reciprocal benefits of food reward and
propagule dispersal vary greatly, but the associations considered widely to advan-
tage plants by (1) distance dispersal of seeds, (2) directed dispersal toward higher
nutrient areas such as ant nests and (3) protection or escape from seed predators by
ant activity. These three themes (or working hypotheses for deciphering myr-
mecophily) are not mutually exclusive and may overlap considerably. Long-distance
seed dispersal, for example, may enhance escape from natural enemies clustered
around the parent plants. A literature review of the relative support for each hypoth-
esis across several major biomes (Giladi 2006) revealed considerable support for
each – but a few contexts (notably predation in tropical forests) were insufficiently
represented to appraise them properly (Table 4.2). Likewise, distribution across
plant growth form showed considerable bias toward studies on herbs (42 studies),
followed by shrubs (17) and with very few studies (7) focussed on trees. Giladi
acknowledged the possible bias through inconclusive studies, or those that failed to
support the tested hypothesis, not being published, and also that intricacies of the
various hypotheses are difficult to confirm.
In common with many other ant-plant associations, myrmecochory is often not
specific, but the variety of possible selective advantages (Table 4.3) demonstrates a
Table 4.2 Overview of studies that have tested the three major hypotheses on the evolution of
myrmecochory and their inferences
Habitat
Hypothesis Conclusion S TeO TeF ToF Total
Directed dispersal
Supported 4 3 2 3 12
Not supported 3 3 5 1 12
Non-conclusive 0 1 1 0 2
Total (% support) 7(57) 7(43) 8(25) 4(75) 26(46)
Distance dispersal
Total (% support) 4(75) 3(67) 8(75) 2(100) 17(76)
Predator avoidance
Total (% support) 7(86) 9(89) 10(80) 1(0) 27(81)
Total 18 19 26 7 70
Figures are for number of published studies categorised by major ecosystem (Giladi 2006)
Collectively, 62 published studies tested one or more of these hypotheses
Habitats are S sclerophyll, TeO temperate open, TeF temperate forest and ToF tropical forest
Table 4.3 Some selective advantages to plants that occur from seed dispersal by ants
1. Avoidance of predators – rapid transport of seeds to enclosed environments such as within
nests, where protected from consumption by many predators
2. Avoidance of competitors – reduced interspecific competition for germination and sites
for seedling growth through increased dispersal
3. Precision placement – seeds in ant nests protected from forest fire but still within regimes
that enable sufficient heat for germination
4. Decrease in seed mortality with increased dispersal distance – seedling mortality may be
inversely related to dispersal distance, so wider transportation may be a selective
advantage
5. Dispersal distance per se – dispersal distance itself may be a benefit from myrmecochory
6. Deposition in nutrient-rich environment – differences in soil features between ant nests
and surrounding areas may be substantial and promote heterogeneity of plant community
and fitness of seedlings from ant nests
7. Myrmecochory associated with low-nutrient soils – for sclerophyll vegetation (Australia),
high levels of ant dispersal contrast with vegetation on richer soils and costs for ant
dispersal systems are less than for fleshy fruit production for vertebrate dispersal more
characteristic of the latter
As listed by Rico-Gray and Oliveira (2007) (see text)
4.6 Myrmecochory 89
wide range of ecological contexts that indicate substantial benefits to both parties
(Rico-Gray and Oliveira 2007) within the unifying theme of seed dispersal.
The various functional roles ascribed to myrmecochory have each been explored
in focussed studies, with examples cited below. Thus, the ‘predator-avoidance
hypothesis’ claims that myrmecochorous seeds are protected from predation by
being hidden in ant nests or buried. This principle has sometimes been claimed to
have influenced the evolution of myrmecochory but was only partially supported by
Manzaneda et al. (2005), despite widespread acceptance of its benefits in many dif-
ferent ecosystems. They studied Helleborus foetidus (Ranunculaceae), attended by
several genera of seed vector ants, and with field mice (Apodemus sylvaticus) the
major predator of dispersed seeds. Surveys in four separate areas in Spain showed
that ant nests varied considerably in the number of seeds remaining there, reflecting
differences in geographic structure and the locally determined interaction between
the particular ant and its habitat. However, probability of seed survival was not cor-
related with mouse abundance, but the mice tended to avoid open areas – so, for
example, they did not forage near nests of the ant Camponotus cruentatus, which
thereby constituted a relatively ‘safe microhabitat’.
Collectively, the major services by seed-carrying ants to the host plants comprise
(1) moving seeds away from the parent plants, (2) depositing viable seeds in suit-
able places for germination and (3), less regularly, decreasing seed predation by
removing the attractive elaiosomes. The three themes have only rarely been distin-
guished in any embracing single study. Using Asarum canadense (p. 162) seeds
with the white-footed mouse (Peromyscus leucopus) as a predator in Ohio, Kwit
et al. (2012) showed that advantages of myrmecochory occurred mostly through
burial and elaiosome removal.
The benefits to ants from myrmecochory have, in general, been documented in
less detail than those for plants and are founded in the nutritional values of the
elaiosomes. These typically have higher concentrations of proteins, lipids and car-
bohydrates than most other plant tissues but may also contain specialised essential
nutrients that the ants cannot synthesise, so providing an alternative rationale to
benefit from nutritional value alone. Gammans et al. (2005) noted that solutions to
this dilemma must flow from well-controlled experiments in which the range of
alternative foods available in the field is quantified or removed. Their trials involved
Myrmica ruginodis, established in laboratory colonies and fed standard diets, with
elaiosomes from seeds of Ulex, a myrmecophyte used regularly by this ant, supplied
for 10 weeks over the period corresponding with their usual season of field avail-
ability. These regimes were applied to colonies with and without queens, in which
elaiosomes yielded different outcomes. Although elaiosome-fed colonies produced
heavier overwintering larvae in both treatments, with eggs producing larvae 48%
heavier than in control (non-elaiosome) colonies, ‘queenless’ workers produced far
more larvae (102% more) than in control colonies. This contrasted with colonies
with queens, in which there was a tendency for control colonies to produce more
overwintering larvae. Tentative explanation for this difference (Gammans et al.
2005) involved the elaiosome-fed larvae entering the development cycle that would
produce more workers for summer foraging, whilst the lower quality of food to
control colonies might have slowed larval development so that more individuals
overwintered. The effects were attributed to increased diet quality, possibly through
provision of essential nutrients, these probably including fatty acids and sterols.
In addition to effective seed transport to ant nests, many seeds may be dropped
en route and either be left where they fall or retrieved by other ants. For any such
movements, the size and shape of the elaiosome may limit the size and capability of
potential vectors, and the ‘carrying position’ may be very characteristic (Gorb and
Gorb 1999). The process of seed dispersal comprises several stages – distinguished
by Gorb and Gorb as ‘removal, transport, secondary relocation’, each of which is a
filter with the level of accomplishment depending on the ant species, the size(s) and
abundance of workers and the sizes of seeds and elaiosomes. Transport of seed
influences the conditions that can determine subsequent development, through both
reducing seed predation (by burying seeds) and protection from hot surface fires
(Beaumont et al. 2013), as well as increased nutrients for germinating seedlings
from the enriched soil of ant middens. Wider dispersion of transported seeds reduces
competition both with the parent plant and within seedling groups. In Australian
sclerophyll forests, long recognised as supporting numerous myrmecochorous
plants (Berg 1975), Beaumont et al. evaluated ‘multiphase myrmecochory’, a term
meaning the subsequent further transport of seeds that have already been trans-
ported by one ant species to its nest environs, by additional movement by other ant
species or, indeed, by the initial vector. Rhytidoponera metallica, for example,
moved seeds (without detaching the elaiosomes) of the myrmecochorous shrub
Pultenaea daphnoides (Fabaceae), relocating them from the nest entrance by up to
almost 80 cm in burned areas and about 60 cm in less open unburned habitat, with
little difference between experimental tests of seeds with or without elaiosomes.
More than one dispersal phase can thereby occur in myrmecochory.
The importance of Rhytidoponera as seed dispersers in Australia was endorsed
by a study of R. violacea in Western Australian heathlands (Lubertazzi et al. 2010),
where the ant participates in myrmecochory over numerous plant species. This pat-
tern is more general (Warren and Giladi 2014), with implication that possibly only
‘a few dozen keystone species worldwide’ of myrmecochorous ants link predomi-
nantly with some 11,000 plant species that benefit from their activities. Those asso-
ciations may be highly asymmetrical – whilst the plants rely on ants for numerous
benefits and suffer when ants are absent, seed-dispersing ants usually have no
dependence on the plants. Despite the benefits of elaiosomes as larval or other sup-
plementary food, this seems to not reflect any specific benefit from the elaiosomes
but simply an additional food material. Even some myrmecochores that are highly
attractive to ants seem not to provide them with any special benefit (Turner and
Frederickson 2013). Features indicating keystone status of participating ants were
appraised for Aphaenogaster rudis in North America, where a high proportion of
myrmecochorous seeds from the rich flora are dispersed by this ant (Ness et al.
2009). Aphaenogaster were far more likely to collect those seeds when discovered
than were any of the other ten genera of ants occurring with them in forests, and
their densities are correlated positively with those of the myrmecochorous plants.
‘Keystone status’ was supported by a combination of high encounter rates, high
4.6 Myrmecochory 91
Table 4.4 Seed dispersal by myrmecochorous ants: distances moved by five species in Mexico
% of interactions Dispersal distance
Ant species (n = 134) cm (SD) Maximum
Pachycondyla harpax 54.5 76 (43) 247
P. apicalis 5.9 925 (635) 2050
Pheidole spp. 17.2 0
Solenopsis geminata 19.4 7.9 (8.5) 37
Wasmannia auropunctata 3.0 4.0 (2.9) 7
Horvitz and Schemske (1986)
subsequent seed collection rates and the benefits to the seeds/plants involved. Those
outcomes differed substantially from interactions with several ants that were more
abundant than Aphaenogaster – such as species of Lasius, Leptothorax and
Stenamma – and physical features (size, jaw ‘gape’) may also favour A. rudis.
Further study of a similar myrmecochore-Aphaenogaster association suggested that
seed sizes may also be relevant, with myrmecochores setting small seeds earlier in
order to avoid competition for dispersal mutualisms with larger seeds (Warren et al.
2014). Competition might thus influence the structure of these mutualisms.
Different ant species may characteristically transport seeds over different dis-
tances and at different rates. Observations on Calathea ovandensis (Marantaceae) in
Mexico showed that, although five ant species were attracted to seeds, the seeds
were transported beyond the parent canopy by only two of these, both species of
Pachycondyla that differed considerably in activity. The more abundant P. harpax
moved seeds about 75 cm, but the rarer P. apicalis averaged about twelve times this
distance (Table 4.4) (Horvitz and Schemske 1986).
A broader survey of myrmecochory confirmed the involvement of members of
five ant subfamilies (Dolichoderinae, Ectatomminae, Formicinae, Myrmicinae,
Ponerinae) and a wide range of seed dispersal distances up to slightly more than
10 m – but with a mean of 1.49 m based on 7889 observations (Gomez and Espadaler
2013). The distance is influenced by the density and distribution of the ant nests, as
well as the structure of the local vegetation These authors also reported examples of
far greater distances, the longest being 180 m, for seeds of Acacia ligulata by
Iridomyrmex viridiaeneus in Australia. They noted also claims that myrmecochory
has evolved independently at least 100 times and that as many as 11,000 plant spe-
cies could participate and benefit from this process.
Two distinct guilds of seed-collecting ants are sometimes recognised, as (1)
‘poor-quality dispersers’, granivorous ants that forage as groups that recruit numer-
ous workers to seeds and secrete the seeds within their nests, and (2) ‘high-quality
dispersers’ (sometimes termed ‘removalists’), species that forage individually, carry
diaspores toward their nests, feed on elaiosomes and discard the seeds. Most seed
removal results from species of the second guild and specificity may be aided by the
elaiosomes having chemical stimuli similar to those of the ants’ more usual prey
(Hughes et al. 1994). Thus, whilst the ‘diffuse’ nature of myrmecochory is fre-
quently cited, this may sometimes be illusory – many ant species may remove seeds,
but only a few of them provide efficient seed dispersal as more regular mutualists.
The evolutionary dilemma for the plants involved is thus of attracting optimal myr-
mecochores as the most beneficial partners and avoiding the neutral or antagonistic
ant taxa. Attraction of particular ant species may be an important role of elaiosomes,
with responses of ants to some fatty acids produced by elaiosomes suggesting an
‘elaiosomes are insect analogues’ hypothesis. Hughes et al. (1994) suggested that
one adaptive value of elaiosomes is to attract carnivorous/omnivorous ants rather
than granivorous species, so shifting the balance of attendance to ants that will not
eat the seeds themselves but simply disperse them. Following from Giladi (2006),
an ideal myrmecochorous ant should simultaneously (1) disperse seeds to sites that
are suitable for establishment and are preferably nutrient enriched, (2) disperse
seeds sufficiently rapidly to reduce pre-dispersal predation and (3) disperse seeds
far enough from the parent plants to reduce any related ‘costs’. Each may, in some
way, influence disperser selection and behaviour.
Many of the ant participants involved in myrmecochory are relatively large spe-
cies, but Fokuhl et al. (2012) drew attention to the important roles also of smaller
ants, in their study of the European Temnothorax crassispinus, with worker body
length only 2.3–3.4 mm, as an omnivore that occasionally uses elaiosomes. They
carried seeds of Chelidonium majus (Papaveraceae) for up to 162 cm and to their
nests and provide a counter to the sometimes-cited ‘bigger means better’ maxim of
myrmecochory. The roles of tiny ants, most commonly ignored in such studies, may
bear more critical examination. The values of feeding on elaiosomes for T. crassis-
pinus were demonstrated in comparative feeding trials, in which female dry weight
was significantly higher with this supplementary larval feeding and taken as an
example of the so-called resource-dependent theory that highlights the nutritional
benefits of elaiosomes for female larvae and promotion of queen development.
Another example of involvement by small ants, the role of Tetramorium tsushi-
mae (body length 2–3 mm) in Japan, where it disperses seeds of Lamium amplexi-
caule (Lamiaceae), supported the ‘predator-avoidance hypothesis’. Seed predation
by the sucking bug Adomerus rotundus (Cydnidae) was significantly lower (both in
number and damage intensity) when the ant was present (Tanaka et al. 2015), and
interpreted as the bug being deterred by the ants, this occurring when the seeds were
exposed rather than buried.
Condition of the host plant can strongly influence behaviour of attending ants.
Attraction to particular plant parts – not simply to extrafloral nectaries or domatia
but to such features as sites of herbivore damage and new shoots – may help to con-
centrate protective ants where and when they are most urgently needed. Ant attrac-
tion to particularly vulnerable plant parts could represent highly adapted situations,
but Edwards et al. (2007) discussed a second possible explanation that the attraction
is stimulated by a ‘sensory trap’. This represents false signals from the plant, which
mimic valid signals to stimulate ant behaviour in its normal context. Several possi-
ble cases of this have been postulated amongst ant-plant mutualisms. They include
the mutualism between Acacia and Pseudomyrmex, in which the plant produces ant
repellents that, supposedly, are mimics of ant alarm pheromones valuable in the
context of the balance of benefit from preying on pollinators and the cost of reduced
4.7 Ants, Plants and Fungi 93
pollination success of the plant. Experimental investigation of another context, of

the ant-plant Cordia nodosa (Boraginaceae) and the ant Allomerus octoarticulatus
variety demerarae, led Edwards et al. to reject this hypothesis in favour of worker
attraction to new shoots being by active selection. Allomerus can destroy flowers of
C. nodosa, reducing fruit production completely in many host plants (Yu and Pierce
1998). Those ‘castrated’ plants occupied by Allomerus benefit the ant by increased
number of domatia over plants occupied by truly mutualistic Azteca ants. That cas-
tration behaviour thereby benefits Allomerus to the cost of Cordia, by preventing
association with the beneficial Azteca – a scenario characterised by Yu and Pierce as
‘a striking example of cheating in a mutualism’.
In this case, as in some others, sterilisation/castration is not complete so that
some of the plants occupied by the parasite can still reproduce. Examined both
experimentally and theoretically by Tarnita et al. (2014), this outcome can arise
from ecologically induced variations in parasite loads on the trees. Acacia drepa-
nolobium showed increased reproduction when small colonies of the sterilising ant
Crematogaster nigriceps were present. It seemed that those colonies arose from
interference competition between different colonies and facilitated host plant repro-
duction because the number of parasites was simply too low to achieve complete
sterilisation of the acacia.
4.7 Ants, Plants and Fungi
The relatively recent recognition of fungi growing in the domatia formed by many
ant-plants has added further complexity to understanding the apparently simpler
bipartite ‘ant-plant’ interactions, with increasing evidence that ant-fungus interac-
tions can also be mutualistic (Blatrix et al. 2009). The ant-plant they studied
(Leonardoxa africana africana, Fabaceae) is endemic to lowland rainforest in
southern Cameroon, where each plant is occupied by a colony of the mutualist ant
Petalomyrmex phylax – which, however, can be replaced by the parasitic Cataulacus
mckeyi. Domatia with the mutualist ant usually contain fungus, which is absent
from unopened domatia and parasite-occupied plants. The fungus (order
Chaetothyriales) is thus associated very strongly with the mutualist ant, by which it
is actively propagated, and appears to contribute to the tripartite and apparently
specific association amongst ‘plant-ant-fungus’. It seems clear that the fungus is
inoculated by the ant (Defossez et al. 2009), whose video recordings of ants within
domatia showed them moving pieces of the fungus mat around. Ants living within
domatia may tend and feed on fungi, and the latter may aid in nutrient dynamics as
intermediaries within ant-to-plant nitrogen transfer. This tripartite association was
demonstrated for sister species of Chaetothyriales in domatia of two sympatric ant-
plants, Barteria fistulosa and B. dewevrei (Passifloraceae) in Gabon, where each
plant contains a single ant colony of one of three ant species (Tetraponera aethiops,
T. latifrons, Crematogaster sp.) (Kokolo et al. 2016).
The role (or roles) of the fungi remains somewhat unclear. Blatrix et al. advanced
two hypotheses, not mutually exclusive, as (1) the fungus patch being a nutrient
recycler, transferring refuse from the ant middens into nutrients available to the
plant, and (2) the fungus being a direct food source for the ants. Wider surveys indi-
cate that fungi are widespread in association with ant-plants and occur on all
continents.
Extending to yet another, but perhaps related, dimension of mutualisms involv-
ing ants, the ants may offer some protection against plant pathogens on foliage. The
presence of mutualist ants can reduce the number of fungal spores on the host’s
leaves, in some cases as an additional benefit to the plant from ants attracted to
extrafloral nectaries (de la Fuente and Marquis 1999). The presence of mutualistic
Crematogaster ants prevents infection of Macaranga by fungi (Heil et al. 2002).
This scenario was extended to cover other epiphyllic organisms on Acacia hindsii
(Mimosaceae) (Gonzalez-Teuber et al. 2014), where indirect defences by
Pseudomyrmex ferrugineus actively patrolling foliage also protected the plant from
bacterial pathogens.
The 200 or so species of fungus-growing ants (Attini, most commonly desig-
nated ‘leaf-cutter’ or ‘leaf-cutting’ ants) are a notable neotropical phenomenon, in
which the workers cut pieces of foliage and transport these to their nests. They
maintain a specialised symbiosis through which the ants use foliage as manure for
a fungus, creating fungus gardens as supplies of food, mainly for their brood. The
ants are ‘true farmers’ (Mueller 2002), with successive gardens established by
transfer of fungus from older to newly prepared gardens. The ants may subsequently
both ‘fertilise’ (by application of faecal matter) and ‘weed’ those cultivated areas
(through grooming the fungal mycelium to remove contaminant spores and excising
fragments of the garden that include fungal parasites). A considerable variety of
fungal cultivars occur, and the ant-fungus cultivar associations exhibit mutualisms
through dependencies on nutrition, physiology and antibiotic traits. Ant larvae feed
exclusively on the cultivated fungi, whilst worker ant diets can be more variable and
also include plant materials. The mutualisms are, therefore, based on (1) fungus
providing critical nutrients for ants, especially brood, and gaining a clean and tended
substrate and growth environment; (2) complex behavioural-physiological integra-
tion, whereby the ants actively distribute fungal digestive enzymes within the gar-
den, the fungus thus exploiting the ants in distributing the enzymes; and (3) the ants
providing antibiotics from their metapleural glands and preventing germination of
alien fungal spores in their gardens. All these themes are complex (Mueller 2002),
but Caldera et al. (2009) considered the associations to be obligate mutualisms in
which an array of specialised microfungal symbionts and parasites is far more
diverse than supposed earlier.
The true ‘leaf-cutting ants’, by far the best-known Attini, comprise only a portion
of the tribe, with many genera collecting fallen plant materials or other ‘wastes’
such as insect corpses and faeces to nourish their fungal partners. These taxa, in
general, have been studied far less than the more advanced leaf-cutters (Leal et al.
2011) but also form a variety of fungus gardens. Leal et al. distinguished (1) pieces
of substrate enveloped by yeasts, (2) laminar fungus suspended within the nest
chambers and (3) amorphous fungus on the chamber floor. The true leaf-cutting taxa
are viewed as a highly derived group from these likely more ancestral agricultural
systems involving ants and fungi.
The intricacies of these associations within wider communities vary greatly, but
the interactions affect both the grazing and decompositional food webs that involve
numerous other species and interactions. However, the obligate mutualism between
Attini and fungi (mostly Basidiomycetes) is based on the antagonistic interaction
whereby ants exploit plants by active herbivory, grazing and removing foliage. In
general these ants are highly successful polyphagous herbivores, with their gather-
ings providing the substrates for polyphagous fungi. Rico-Gray and Oliveira (2007)
referred to studies demonstrating high richness and taxonomic variety of plants
exploited: a year of observations of a single colony of Atta colombica in Panama,
for example, showed that ants harvested foliage from 126 plant species over that
period (Wirth et al. 1997, 2003 – the latter a comprehensive natural history account
of lasting interest).
Recognition that the fungi involved, rather than being passive recipients of ant
attention, may play more active roles by influencing ant behaviour introduces a
range of wider considerations for these interactions (Mueller 2002) through roles of
different fungal cultivars. Mueller suggested that attine fungus cultivars may ‘have
evolved chemical and physiological schemes that alter ant behaviour to serve the
fungus’ reproductive interests’, possibly also affecting the ants’ reproductive per-
formance. The dynamics of ‘conflict versus cooperation’ in these complex
interactions, with cultivars transported frequently between sites, need much more
investigation to clarify their evolution (Green et al. 2002). Cultivation of fungi for
food by ‘pre-attine ants’ apparently evolved 50–60 million years ago, and transfer
of cultivars occurs both ‘vertically’ (through generations, by transfer by ant off-
spring as they found new nests) and, more rarely, ‘horizontally’ (between related ant
species). The latter was reported for two species of Cyphomyrmex in Panama,
through an extensive phylogenetic analysis (Green et al. 2002) of the 72 fungal
cultivars associated with two sympatric ant species.
The extent of attine nest constructions can cause massive disturbances to local
environments. Wirth et al. (2003) noted that construction of a single Atta nest can
involve moving 23 m3 (or 40 tons) of soil to the nest surface. Similarly, Meyer et al.
(2013) claimed that a single species, Atta cephalotes, could cause complete turnover
of soil of a Costa Rican rainforest within 200–300 years, with accompanying eco-
logical impacts that included alterations to forest structure through creating light
gaps above nests and associated changes in microclimate. They noted leaf-cutting
ants as ‘ecosystem engineers’ which, in human-modified environments where they
are especially abundant (such as near forest edges), can lead to increased homogeni-
sation of environmental conditions – in part through decreased soil nutrient levels
and decreased litter cover around nest sites. Individual nests of A. cephalotes can
exceed 250 m2 in surface area, and impacts include (1) improving soil penetrability
in relation to non-nest soils, (2) enhancing soil nutrient availability through accu-
mulation of leaf material used to nourish fungi in the underground chambers, (3)
promoting bare ground around the nests as leaf litter is removed and (4) creating
Fig. 4.11 The mean

colony densities of attine
ants ((a) Atta cephalotes,
(b) A. sexdens) in different
distance zones from the
forest border toward the
interior of Atlantic forest in
Brazil (Wirth et al. 2007)
substantial nest clearings within the forests. Such wider changes augment herbivory
and broaden the range of effects, and, not surprisingly in view of their abundance
and predominance, the various impacts of Attinae have received considerable atten-
tion. The far-reaching impacts are clearly very variable within any individual con-
text – thus, for Atta nests in their Coimbra (Brazil) study, Correa et al. (2010)
showed that disturbance from ant activity may lead to impoverished plant assem-
blages along edges of Atlantic forest by both directly and indirectly reducing recruit-
ment of trees and shrubs. Ant nests were associated with environments that were not
suitable for (1) shade-tolerant, light-sensitive plant species; (2) plants with seed-
lings targeted by ants; and (3) plants with seeds for which germination needs undis-
turbed conditions. Those impacts occur for several tens of metres around an A.
cephalotes nest.
Herbivory by leaf-cutting ants can reach levels up to 13% (Wirth et al. 2003) or
15% (Leal et al. 2011) of standing leaf crop per year, extending over more than half
the plant species present and reducing vegetation cover substantially (Leal et al.
2011). Impacts occur on individual plants, communities and ecosystems. Ant popu-
lations can increase markedly in areas subject to human impacts such as deforesta-
tion and agricultural conversion. Wirth et al. (2007) specifically noted that colony
densities increase with pasture conversion, in plantations, in early successional for-
est and in isolated forest remnants. Populations of Atta cephalotes and A. sexdens in
Brazil both appear to benefit from presence of forest edges (Fig. 4.11), with colony
numbers declining within deeper undisturbed forest.
Their increased presence in more open disturbed areas has been attributed to
presence of fast-growing and poorly defended pioneer plants that can be exploited
Fig. 4.12 The costs and

benefits to myrmecophytic
plants of maintaining
protective ants of different
colony sizes. The greatest
net benefit is from colony
size ‘N’, and where the
costs of maintaining the
ant colony exceed the
benefits provided by the
ants, the mutualism
changes to become
parasitic (Bronstein 1998,
from Fonseca 1993)
easily. Indeed, it has been claimed that some Atta are becoming ‘hyper-abundant’ in
parts of Brazil due to increasing levels of forest fragmentation (Correa et al. 2010).
In short, edge creation can promote high densities of leaf-cutting ants, with dra-
matic environmental impacts that may augment other concerns from deleterious
edge effects. Atta sexdens removed about three times more leaf area from drought-
stressed plants than from normal ‘control’ plants in Brazil (Neto et al. 2012). This
trait was suggested to be a more general aspect of food plant choice by leaf-cutting
ants – and may also contribute to their greater relative abundance on forest edges
where drought stress may be most evident. Although leaf-cutting ants are both
polyphagous and highly selective in the food plants they choose, attractiveness of
drought-stressed foliage is widespread (Meyer et al. 2006). It is, however, unclear
whether that selection is primarily to fulfil immediate food/energy needs of the for-
aging ants or provide for optimal fungus growth – or a combination of these and, so,
beneficial to both partners.
Ant-fungus mutualisms pose rather different evolutionary opportunities than the
better-studied ant-trophobiont mutualistic associations with other insects sum-
marised in the next two chapters, and for which the general balance between cost
and benefit outlined earlier (p. 74, Fig. 4.12) can apply. The three a priori hypothe-
ses on the structures of trophobioses between ants and other insects noted by Oliver
et al. (2008) help to clarify this discrepancy. Those hypotheses are (1) ecologically
dominant, polygynous ants are more likely to monopolise resources and form close
relationships with sugar-producing insects; (2) mutualisms involving ants-Hemip-
tera, ants-Lycaenidae and ants-extrafloral nectaries are likely to be positively asso-
ciated, and ants that associate with Hemiptera or Lycaenidae are largely a subset of
those that visit extrafloral nectaries, so that capability for liquid feeding by ants may
provide opportunity for all three mutualisms; and (3) fungus-gardening is nega-
tively associated with trophobiosis, with the trade-off preventing specialisation in
the two major but different kinds of mutualism – in that context, fungal cultivation
Fig. 4.13 Scheme illustrating that multiple mutualisms involving ants cannot occur if there is a
trade-off in requirements and redundancy of awards. Ants can simultaneously participate in spe-
cialised mutualisms with hemipterans (A) and domatia-bearing plants (B). In contrast (right of
diagram), there is a negative correlation between ant-hemipteran (A) and ant-fungus (C) mutualism
(Oliver et al. 2008)
necessitates establishment and maintenance of growth substrates, whilst maintain-

ing trophobionts necessitates a patrolling defence as a very different strategy that
involves trade-offs between different potential mutualistic partners for the ants.
From their extensive literature survey, Oliver et al. (2008) suggested that specialised
obligate mutualisms involving ants will not develop under such trade-off conditions
(Fig. 4.13).
References
Amsellem L, McKey DB (2006) Integrating phenological, chemical and biotic defences in ant-
plant protection mutualisms; a case study of two myrmecophyte lineages. Chemoecology
16:223–234
Apple JL, Feener DH Jr (2001) Ant visitation of extrafloral nectaries of Passiflora: the effects of
nectary attributes and ant behavior on patterns in facultative ant-plant mutualisms. Oecologia
127:409–416
Assuncao MA, Torezan-Silingardi M, Del-Claro K (2014) Do ant visitors to extrafloral nectaries
repel pollinators and cause an indirect cost of mutualism? Flora 209:244–249
Beattie AJ (1985) The evolutionary ecology of ant-plant interactions. Cambridge University Press,
Cambridge
References 99
Beaumont KP, Mackay DA, Whalen MA (2013) Multiphase myrmecochory: the roles of different
ant species and effects of fire. Oecologia 172:791–803
Becerra JX, Venable DL (1989) Extrafloral nectaries: a defense against ant-Homoptera mutual-
isms? Oikos 55:276–280
Bentley BJ (1977) Extrafloral nectaries and protection by pugnacious bodyguards. Annu Rev Ecol
Syst 8:407–427
Berg RY (1975) Myrmecochorous plants in Australia and their dispersal by ants. Aust J Bot
213:475–508
Bixenmann RJ, Coley PD, Kursaw TA (2011) Is extrafloral nectar production induced by herbi-
vores or ants in a tropical facultative ant-plant mutualism? Oecologia 165:417–425
Blatrix R, Mayer V (2010) Communication in ant-plant symbioses. In: Baluska F, Nincovic V
(eds) Plant communication from an ecological perspective. Springer, Berlin, pp 127–158
Blatrix R, Bouamer S, Morand S, Selosse M-A (2009) Ant-plant mutualisms should be viewed as
symbiotic communities. Plant Signal Behav 4:554–556
Bluthgen N, Verhaagh M, Goitia W, Jaffe K, Morewetz W, Barthloot W (2000) How plants shape
the ant community in the Amazonian rainforest canopy: the key role of extrafloral nectaries and
hemipteran honeydew. Oecologia 125: 229-240
Bluthgen N, Wesenberg J (2001) Ants induce domatia in a rain forest tree (Vochysia vismiaefolia).
Brouat C, Garcia N, Andary C, McKey D (2016) Plant lock and ant key: pairwise coevolution of
an exclusion filter in an ant-plant mutualism. Proc R Soc Lond B 268:2131–2141
Buckley RC (1982) Ant-plant interactions: a world review. In: Buckley RC (ed) Ant-plant interac-
tions in Australia. W. Junk, The Hague, pp 111–141
Caldera EJ, Poulsen M, Suen G, Currie CR (2009) Insect symbioses: a case study of past, present,
and future fungus-growing ant research. Environ Entomol 38:78–92
Carroll CR, Janzen DH (1973) Ecology of foraging by ants. Annu Rev Ecol Syst 4:231–257
Cembrowski AR, Tan MG, Thomson JD, Frederickson ME (2014) Ants and ant scent reduce bum-
blebee pollination of artificial flowers. Am Nat 183:133–139
Correa NN, Silva PSD, Wirth R, Tabarelli M, Leal IR (2010) How leaf-cutting ants impact forests:
drastic nest effects on light environment and plant assemblages. Oecologia 162:103–115
Cronin G (1998) Between species and temporal variation in Acacia-ant-herbivore interactions.
Dattilo W, Marquitti FMD, Guimares PR Jr, Izzo TJ (2014) The structure of ant-plant ecological
networks: is abundance enough? Ecology 95:475–485
Davidson DW (1988) Ecological studies of Neotropical ant gardens. Ecology 69:1138–1152
Davidson DW, McKey D (1993) The evolutionary ecology of symbiotic ant-plant relationships.
J Hymenopt Res 2:13–83
de la Fuente MAS, Marquis RJ (1999) The role of ant-tended extrafloral nectaries in the protection
and benefit of a neotropical rainforest tree. Oecologia 118:192–202
de Vega C, Arista M, Ortiz PL, Herrera CM, Talavera S (2009) The ant-pollination system of
Cytinus hypocistis (Cytinaceae), a Mediterranean root holoparasite. Ann Bot 102:1065–1075
Defossez E, Selosse MA, Dubois MP, Mondolot L, Faccio A (and 3 other authors) (2009) Ant-
plants and fungi: new threeway symbiosis. New Phytol 182:942–949
Di Giusto B, Anstett M-C, Donnias E, McKey DB (2001) Variation in the effectiveness of biotic
defence: the case of an opportunistic ant-plant protection mutualism. Oecologia 129:367–375
do Nascimento EA, Del-Claro K (2010) Ant visitation to extrafloral nectaries decreases herbivory
and increases fruit set in Chamaecrista debilis (Fabaceae) in a Neotropical savanna. Flora
205:754–756
Dyer LA, Letourneau DK (1999) Relative strengths of top-down and bottom-up forces in a tropical
forest community. Oecologia 119:265–274
Edwards DP, Arauco R, Hassall M, Sutherland WJ, Chamberlains K, Wadhams LJ, Yu DW (2007)
Protection in an ant-plant mutualism: an adaptation or a sensory trap? Anim Behav
74:377–385
Edwards DP, Frederickson ME, Shepard GH, Yu DW (2009) A plant needs ants like a dog needs
fleas: Myrmelachista schumanni ants gall many tree species to create housing. Am Nat
174:734–740
Fiala B, Grunsky H, Maschwitz U, Linsenmaier KE (1994) Diversity of ant-plant interactions:
protective efficacy in Macaranga species with different degrees of ant association. Oecologia
97:186–192
Fokuhl G, Heinze J, Poschlod P (2012) Myrmecochory by small ants – beneficial effects though
elaiosome nutrition and seed dispersal. Acta Oecol 38:71–76
Fonseca CR (1993) Nesting space limits colony size of the plant-ant Pseudomyrmex concolor.
Oikos 67:473–482
Frederickson ME, Gordon DM (2007) The devil to pay: a cost of mutualism with Myrmelachista
schumanni ants in ‘devil’s gardens’ is increased herbivory on Duroia hirsuta trees. Proc R Soc
Lond B Biol Soc 274:1117–1123
Frederickson ME, Greene MJ, Gordon DM (2005) ‘Devil’s gardens’ bedeviled by ants. Nature
437:495–496
Galen C (2005) Catching ants with honey: an experimental test of distraction and satiation as
alternative modes of escape from flower-damaging ants. Oecologia 144:80–87
Gammans N, Bullock JM, Schonrogge K (2005) Ant benefits in a seed dispersal mutualism.
Oecologia 146:43–49
Gastreich KR, Gentry GL (2004) Faunal studies in model Piper spp. systems, with a focus on
spider-induced indirect interactions and novel insect-Piper mutualisms. In: Dyer LA, Palmer
ADN (eds) Piper: a model genus for studies of phytochemistry, ecology, and evolution. Kluwer
Academic/Plenum Publishers, New York, pp 97–116
Giladi I (2006) Choosing benefits or partners: a review of the evidence for the evolution of myrme-
cochory. Oikos 112:481–492
Gomez C, Espadaler X (2013) An update of the world survey of myrmecochorous dispersal dis-
tances. Ecography 36:1193–1201
Gonzalez-Teuber M, Kaltenpoth M, Boland W (2014) Mutualistic ants as an indirect defence
against leaf pathogens. New Phytol 202:640–650
Gorb SN, Gorb EV (1999) Dropping rates of elaiosome-bearing seed during transport by ants
(Formica polyctena Foerst.): implications for distance dispersal. Acta Oecol 20:509–518
Green AM, Mueller UG, Adams RMM (2002) Extensive exchange of fungal cultivars between
sympatric species of fungus-growing ants. Mol Ecol 11:191–195
Heil M (2010) Ant-plant mutualisms. In: Encyclopedia of life sciences. Wiley, Chichester.
doi:10.1002/978047001902.a0022558
Heil M, McKey D (2003) Protective ant-plant interactions as model systems in ecological and
evolutionary research. Annu Rev Ecol Evol Syst 34:425–453
Heil M, Fiala B, Kaiser W, Linsenmaier KE (1998) Chemical contents of Macaranga food bodies:
adaptations to their role in ant attraction and nutrition. Funct Ecol 12:117–122
Heil M, Fiala B, Maschwitz U, Linsenmaier KE (2001) On benefits of indirect defence: short- and
long-term studies of antiherbivore protection via mutualistic ants. Oecologia 126:395–403
Heil M, Hilpert A, Fiala B, Bin Hashim R, Strohm E, Zotz G, Linsenmaier KE (2002) Nutrient
allocation of Macaranga triloba ant plants to growth, photosynthesis, and indirect defence.
Funct Ecol 16:475–483
Hodson AK, Gastreich KR (2006) Evidence for a novel mutualism in the tropical understorey
shrub Piper urostachyum. Biotropica 38:127–131
Horvitz CC, Schemske DW (1986) Seed dispersal of a neotropical myrmecochore: variation in
removal rates and dispersal distance. Biotropica 18:319–323
Hughes L, Westoby M, Jurado E (1994) Convergence of elaiosomes and insect prey: evidence from
ant foraging behaviour and fatty acid composition. Funct Ecol 8:358–365
Huxley CR (1980) Symbiosis between ants and epiphytes. Biol Rev 55:321–340
Huxley CR (1982) Ant-epiphytes of Australia. In: Buckley RC (ed) Ant-plant interactions in
Australia. W. Junk, The Hague, pp 63–73
References 101
Itino T, Itioka T (2001) Interspecific variation and ontogenetic change in antiherbivore defense in
myrmecophytic Macaranga species. Ecol Res 16:765–774
Itino T, Davies SJ, Tada H, Hieda Y, Inoguchi M, Itioka T, Yamane S, Inoue T (2001) Cospeciation
of ants and plants. Ecol Res 16:787–793
Itioka T, Nomura M, Inui Y, Itino T, Inoue T (2000) Difference in intensity of ant defense among
three species of Macaranga myrmecophytes in a south-east Asian dipterocarp forest. Biotropica
32:318–326
Janzen DH (1966) Coevolution of mutualism between ants and acacias in Central America.
Evolution 20:249–275
Janzen DH (1967) Interaction of the bull’s-horn acacia (Acacia cornigera L.) with an ant inhabit-
ant (Pseudomyrmex ferrugineus F. Smith) in eastern Mexico. Univ Kansas Sci Bull
47:315–558
Janzen DH (1969) Allelopathy by myrmecophytes: the ant Azteca as an allelopathic agent of
Cecropia. Ecology 50:147–153
Janzen DH (1974) Swollen thorn acacias of Central America. Smithson Contrib Bot 13:1–131
King EG, Caylor KK (2010) Herbivores and mutualistic ants interact to modify tree photosynthe-
sis. New Phytol 187:17–21
Kokolo B, Atteke C, Ibrahim B, Blatrix R (2016) Pattern of specificity in the tripartite symbiosis
between Barteria plants, ants and Chaetothyriales Fungi. Symbiosis 69:169–174
Koptur S (1985) Alternative defenses against herbivores in Inga (Fabaceae: Mimosoideae) over an
elevational gradient. Ecology 66:1639–1650
Koptur S (2005) Nectar as fuel for plant protectors. In: Wackers FL, van Rijn PCJ, Bruin J (eds)
Plant-provided food and herbivore-carnivore interactions. Cambridge University Press,
Cambridge, pp 76–108
Kwit C, Marcello GJ, Gonzalez JL, Shapiro AC, Bracken RD (2012) Advantages of seed dispersal
for a myrmecochorous temperate forest herb. Am Midl Nat 168:9–17
Lanan MC, Bronstein JL (2013) An ant’s–eye view of an ant-plant mutualism. Oecologia
172:779–790
Latteman TA, Mead JE, DuValt MA, Bunting CC, Bevington JM (2014) Differences in anti-
herbivore defences in non-myrmecophyte and myrmecophyte Cecropia trees. Biotropica
46:652–656
Leal IR, Silva PSD, Oliveira PS (2011) Natural history and ecological correlates of fungus-growing
ants (Formicidae: Attini) in the neotropical cerrado savanna. Ann Entomol Soc Am
104:901–908
Lubertazzi D, Lubertazzi MAA, McCoy N, Gove AD, Majer JD, Dunn RR (2010) The ecology of
a keystone seed disperser, the ant Rhytidoponera violacea. J Insect Sci 18:1–15
Manzaneda AJ, Fedriani JM, Rey PJ (2005) Adaptive advantages of myrmecochory: the predator-
avoidance hypothesis tested over a wide geographic range. Ecography 28:583–592
Marazzi B, Bronstein JL, Koptur S (2013) The diversity, ecology and evolution of extrafloral nec-
taries: current perspectives and future challenges. Ann Bot 111:1243–1250
Maschwitz U, Fiala B, Moog J, Saw LG (1991) Two new myrmecophytic associations from the
Malay Peninsula: ants of the genus Cladomyrma (Formicidae: Camponotinae) as partners of
Saraca thaipingensis (Caesalpiniaceae) and Crypteronia griffithii (Crypteroniaceae). Insect
Soc 38:27–35
Mayer VE, Frederickson ME, McKey D, Blatrix R (2014) Current issues in the evolutionary ecol-
ogy of ant-plant symbioses. New Phytol 202:749–764
Meyer ST, Roces F, Wirth R (2006) Selecting the drought-stressed: effects of plant stress on intra-
specific and within-plant herbivory patterns of the leaf-cutting ant Atta colombica. Funct Ecol
20:973–981
Meyer ST, Neubauer M, Sayer EL, Leal IR, Tabarelli M, Wirth R (2013) Leaf-cutting ants as eco-
system engineers: topsoil and litter perturbations around Atta cephalotes nests reduce nutrient
availability. Ecol Entomol 38:497–504
Mueller UG (2002) Ant versus fungus versus mutualism: ant-cultivar conflict and the deconstruc-
tion of the attine ant-fungus symbiosis. Am Nat 160:S67–S98
Ness JH (2006) A mutualism’s indirect costs: the most aggressive plant bodyguards also deter pol-
linators. Oikos 113:506–514
Ness JH, Morin DF, Giladi I (2009) Uncommon specialization in a mutualism between a temperate
herbaceous plant guild and an ant: are Aphaenogaster ants keystone mutualists? Oikos
118:1793–1804
Neto JDR, Pinho BX, Meyer ST, Wirth R, Leal IR (2012) Drought stress drives intraspecific choice
of food plants by Atta leaf-cutting ants. Entomol Exp Appl 144:209–215
New TR (2005) Invertebrate conservation and agricultural ecosystems. Cambridge University
Press, Cambridge
Oleson JM, Lindberg AB, Eskildsen LI, Svenning JC, Lindberg R (2002) Plants in the devil’s
garden: intruders in an ant-plant mutualism. Ecotropica 8:81–86
involving ants. J Evol Biol 21:1597–1608
Ona L, Lachmann M (2011) Ant aggression and evolutionary stability in plant-ant and plant-
pollinator mutualistic interactions. J Evol Biol 24:617–629
Palmer TM, Brody AK (2007) Mutualism as reciprocal exploitation: African plant-ants defend
foliar but not reproductive structures. Ecology 88:3004–3011
Palmer TM, Stanton ML, Young TP, Goheen JR, Pringle EM, Karban R (2008) Breakdown of an
ant-plant mutualism follows the loss of large herbivores from an African savanna. Science
319:192–195
Peakall R, Beattie AJ (1991) The genetic consequences of worker ant pollination in a self-
compatible, clonal orchid. Evolution 45:1837–1848
Peakall R, Beattie AJ, James SH (1987) Pseudocopulation of an orchid by male ants: a test of two
hypotheses accounting for the rarity of ant pollination. Oecologia 73:522–524
Piovia-Scott J (2011) The effect of disturbance on an ant-plant mutualism. Oecologia
166:411–420
Quek S-P, Davies SJ, Itino T, Pierce NE (2004) Codiversification in an ant-plant mutualism: stem
texture and the evolution of host use in Crematogaster (Formicidae: Myrmicinae) inhabitants
of Macaranga (Euphorbiaceae). Evolution 58:554–570
Razafimandimbison SG, Moog J, Lantz H, Maschwitz U, Bremer B (2005) Re-assessment of
monophyly, evolution of myrmecophytism, and rapid radiation in Neonauclea s.s. (Rubiaceae).
Mol Phylogenet Evol 34:334–354
Rico-Gray V, Oliveira PS (2007) The ecology and evolution of ant-plant interactions. University of
Chicago Press, Chicago/London
Ridsdale CE (1989) A revision of Neonauclea (Rubiaceae). Blumea 34:177–275
Risch SJ, Rickson FR (1981) Mutualism in which ants must be present before plants produce food
bodies. Nature 291:149–150
Root RB (1973) Organization of plant-arthropod association in simple and diverse habitats: the
fauna of collards (Brassica oleracea). Ecol Monogr 43:95–124
Rudgers JA (2002) Evolutionary ecology of ant-wild cotton associations. Ph D thesis, University
of California, Davis. (not seen, cited by Rudgers et al. 2003, below)
Rudgers JA, Gardener MC (2004) Extrafloral nectar as a resource mediating multispecies interac-
tions. Ecology 85:1495–1502
Rudgers JA, Hodgen JG, White JW (2003) Behavioral mechanisms underlie an ant-plant mutual-
ism. Oecologia 135:51–59
Ruhren S, Handel SN (1999) Jumping spiders (Salticidae) enhance the seed production of a plant
with extrafloral nectaries. Oecologia 119:227–230
Sachs JL (2015) Exploitation of mutualism. In: Bronstein JL (ed) Mutualism. Oxford University
Sagers CL, Ginger SM, Evans RD (2000) Carbon and nitrogen isotopes trace nutrient exchange in
an ant-plant mutualism. Oecologia 123:582–586
References 103
Savage AM, Peterson MA (2007) Mutualism in a community context: the positive feedback
between an ant-aphid mutualism and a gall-making midge. Oecologia 151:280–291
Schumer M, Birger R, Tantipathananandh C, Aurisano J, Maggioni M, Mwangi P (2013) Infestation
by a common parasite is correlated with ant symbiont identity in a plant-ant mutualism.
Shimizu-kaya U, Itioka T (2016) Reduced ant defenses in Macaranga myrmecophytes
(Euphorbiaceae) infested with a winged phasmid. Ecol Res 31:665–672
Smiley J (1978) Plant chemistry and the evolution of host specificity: new evidence from Heliconius
and Passiflora. Science 201:745–747
Stanton ML, Palmer TM (2011) The high cost of mutualism: effects of four species of East African
ant symbionts on their myrmecophyte host tree. Ecology 92:1073–1082
Tanaka K, Ogata K, Mukai H, Yamawo A, Tokuda M (2015) Adaptive advantage of myrmecochory
in the ant-dispersed herb Lamium amplexicaule (Lamiaceae): predation avoidance through the
deterrence of post-dispersal seed predators. PLoS One 10(7):e0133677. doi:10.1371/lournal.
pone.0133677
Tarnita CE, Palmer TM, Pringle RM (2014) Colonisation and competition dynamics can explain
incomplete sterilisation parasitism in ant-plant symbioses. Ecol Lett 17:1290–1298
Turner KM, Frederickson ME (2013) Signals can trump rewards in attracting seed-dispersing ants.
PLoS One 8:e71871. doi:10.1371/journal.pone.0071871
Vittecoq M, Djieto-Lordon C, Buatois B, Dormont L, McKey D, Blatrix R (2011) The evolution of
communication in two ant-plant mutualisms. Evol Biol 38:360–369
Warren RJ II, Giladi I (2014) Ant-mediated seed dispersal: a few ant species (Hymenoptera:
Formicidae) benefit many plants. Myrmecol News 20:129–140
Warren RJ II, Giladi I, Bradford MA (2014) Competition as a mechanism structuring mutualisms.
J Ecol 102:486–495
Way MJ (1963) Mutualism between ants and honeydew-producing Homoptera. Annu Rev Entomol
8:307–344
Wirth R, Beyschlag R, Tyel J, Holldobler B (1997) Annual foraging of the leaf-cutting ant Atta
colombica in a semideciduous rain forest in Panama. J Trop Ecol 13:741–757
Wirth R, Herz H, Tyel TJ, Beyschlag W, Holldobler B (2003) Herbivory of leaf-cutting ants: a case
study of Atta colombica in the tropical rainforest of Panama, Ecological studies, vol 164.
Springer, Berlin
Wirth R, Meyer SR, Almeida WR, Araujo MV Jr, Barbosa VS, Leal IR (2007) Increasing densities
of leaf-cutting ants (Atta spp.) with proximity to the edge in a Brazilian Atlantic forest. J Trop
Ecol 23:501–505
Yu DW (1994) The structural role of epiphytes in ant gardens. Biotropica 26:222–226
Yu DW, Pierce NE (1998) A castration parasite of an ant-plant mutualism. Proc R Soc Lond B
265:375–382
Chapter 5
Classic Themes: Ants and Other
Insects – Hemiptera
Keywords Aphids • Auchenorrhyncha • Honeydew • Membracidae • Mutualism •

Scale insects • Sternorrhyncha • Trophobiosis
5.1 Introduction: Trophobiotic Associations
This chapter and the next deal predominantly with a variety of specialised mutual-
isms and myrmecophily that primarily involve interactions of ants with two large
insect orders: Hemiptera (this chapter) and Lepidoptera (Chap. 6). The wider rami-
fications of myrmecophily involve diverse assemblages of species that reflect the
wide variety of ants and the stable, well-protected environments they offer to associ-
ates. Ivens et al. (2016) noted that the types of ant-myrmecophyte (Chap. 4) interac-
tions are ‘as diverse as the myrmecophiles themselves’. Within that array and its
parallels treated here, mutualisms are often implied but the nuances among these
relationships commonly render them difficult to validate or to characterise. Larvae
of aphytophagous miletine lycaenid butterflies (p. 120), for example, are predators
on Hemiptera but may also be associated strongly with single ant species, so that the
hemipterans may be mutualists with the ant(s) involved and, in turn, be restricted to
particular host plants. Lohman and Samarita (2009) noted the extreme specialisa-
tion shown in Malaysia by the lycaenid Allotinus apries, which associates with an
uncommon ant (Myrmicaria arachnoides). First instar Allotinus larvae fed on ant-
attended coccids but later stages are passed in the ant nest, presumed to be predators
on ant brood or ‘cuckoos’ (p. 125). The scarcity of the butterfly may reflect this
specialised development on a rare ant and on patchily distributed or ephemeral
Hemiptera prey.
Many of the associations outlined below fall into the broad scope of ‘trophobio-
sis’, in which the attending ants gain food from their ‘trophobionts’ and protect
them from natural enemies, as the basis of a mutualism. Many such associations
may be rather casual, opportunistic and non-specific, and the two major categories
of (1) ants and Hemiptera (Sternorrhyncha, Auchenorrhyncha) and (2) ants and
Lepidoptera demonstrate many evolutionary and ecological subtleties from long-
term associations. Many of them are also of considerable antiquity.

DOI 10.1007/978-3-319-58292-4_5
106 5 Classic Themes: Ants and Other Insects – Hemiptera
The above two categories have dominated considerations of trophobiotic mutual-

isms involving insects, but several more unusual associations have also been
reported, and their functional status may be complex to determine. The Japanese
moth Nudina artaxidia (Arctiidae [or Erebidae: Arctiinae]) is one of a diverse and
widespread group whose larvae feed predominantly on lichens. Nudina larvae, how-
ever, feed on honeydew produced by the Giant scale insect Drosicha corpulenta
(Margarodidae) and actively solicit honeydew by tapping the rear of the scale
(Komatsu and Itino 2014). Larvae associate with the mutualistic ants (several spe-
cies of Lasius subgenus Dendrolasius) attending the scale insects, and follow the
ant trails to encounter the scales. The consistency of this behaviour implied that N.
artaxidia is a myrmecophilous moth that depends on ants and ant-tended Hemiptera,
and that has developed intricate behaviour to avoid attack by the ants and exploit the
honeydew. The moth larvae have dense long body hairs that may prevent contact
with the ants, and did not appear to depend on any nutrition from the ants. Nudina
is thus likely to be an exploiter (Chap. 8) rather than a mutualist.
Trophobiotic interactions also occur in other insects. In Malaysia and Africa,
plataspid bugs (Heteroptera) are tended by ants within the carton nest ‘pavilions’
constructed by the ants at the base of tree trunks. One such bug, Caternaultiella
rugosa, is tended in Cameroon by two ant species (Gibernau and Dejean 2001) that
protect the bug’s egg masses from attacks by an encyrtid wasp parasitoid. Inside the
pavilions, the usual parental care from the bugs may be abandoned to the ants, and
the ants gain honeydew from the bugs. Differences were found between the two ant
species, with Camponotus brutus providing more effective protection than
Myrmicaria opaciventris. Ant activity is also linked with the bug’s own intensity of
brood protection, with adults and later instars grouping around the first instar
plataspids.
5.2 Ant-Hemiptera Mutualisms
Ant-Hemiptera interactions have been considered ‘keystone interactions’ because of

their wide impacts on arthropod communities and their host plants (Styrsky and
Eubanks 2007). Not all those interactions are mutualisms but, as in many other asso-
ciations between species, the ecological consequences may be both diverse and
poorly understood. Styrsky and Eubanks emphasised that relatively few studies have
actually quantified the real effects of ant-Hemiptera interactions, and that many pub-
lished studies were weakened by design flaws. As Fig. 5.1 displays, the interaction
may (1) provide a net benefit to the host plant if the positive effects of the ants sup-
pressing other herbivores outweigh the direct impact of the Hemiptera on the plant,
or (2) harm plants if the impact of the hemipterans on the plant outweighs the posi-
tive aspects of increased ant suppression of other herbivores. Influences of herbivory
by other herbivores may affect this balance between ants and honeydew-producing
Hemiptera (Fig. 5.2). More generally, from the extensive literature overview by
Styrsky and Eubanks (2007), the consequences and importance of many
5.2 Ant-Hemiptera Mutualisms 107
Fig. 5.1 Some potential

consequences of
interactions between ants
and honeydew-producing
hemipterans on plants.
Those interactions may
provide (a) net benefit to
the plants if the indirect
positive effect of increased
ant suppression of other
(not honeydew-producing)
herbivores outweigh the
direct effects of
hemipterans on the plants,
or (b) may harm plants if
the direct negative effects
of hemipterans on the
plants outweigh the
indirect positive effects of
increased ant suppression
of other herbivores. Note
that, not shown here,
honeydew-producing and
other herbivores may also
act independently through
effects on host plant
quality (See also Fig. 5.2,
Styrsky and Eubanks 2007)
ant-Hemiptera interactions have not received adequate attention ‘whether because

they are assumed, ignored, dismissed or simply overlooked’.
Many accounts refer to the bugs involved as ‘Homoptera’. That nomenclature
preserves the traditional division of the order Hemiptera into two suborders
(Homoptera and Heteroptera). More recent arrangements, reflecting the paraphy-
letic nature of the group, have led to wide recognition of three major suborders, with
‘Homoptera’ an artificial grouping of two of these (Sternorrhyncha, the scale
insects, aphids and whitefly; Auchenorrhyncha, the leafhoppers, planthoppers and
cicadas). The suborder Heteroptera (true sucking bugs) completes the trio.
Many ant-Hemiptera mutualisms fall into a category termed ‘host-visitor mutu-
alisms’, in which the sap-feeding bugs are tended and protected by visiting ants.
Those Hemiptera, however, may respond differently to the ants and to predators
such as ladybirds (Coccinellidae). The membracid treehopper Publilia concava pro-
duces vibrational alarm signals in the presence of ladybirds (Harmonia axyridis)
but no such signals ensued when only ants (Formica fusca) were present (Morales
et al. 2008). Those signals increased ant activity, in turn increasing protection of the
treehopper and deterring the predators. The signals appeared to be a specific
Fig. 5.2 The consequences of a mutualistic interaction between ants and honeydew-producing
hemipterans on a host plant, based on intensity of herbivory by other herbivores, including
Hemiptera that do not feed from phloem or produce honeydew. Cost to the plant, as loss caused by
honeydew-producing Hemiptera damage to the plants (dotted line), increases with increasing den-
sity of the hemipterans; benefit of the interaction to the plant, as increased fitness caused by ant
suppression of plant damage by other herbivores (solid lines), is shown for two contexts. At a high
level of herbivory by non-tended herbivores, benefits of the interaction to the plant are predicted to
outweigh costs except at extremely high hemipteran densities, whilst at low level of herbivory by
non-tended herbivores, mutualism either produces only slight benefit in fitness, at low hemipteran
densities, or a net loss of fitness at high hemipteran densities (Styrsky and Eubanks 2007)
response to predator attack. The mechanism may thus differ from that of some ant-
Lepidoptera mutualisms (p. 130) in which acoustic signalling by larvae can be a
response to both ant presence and the simulated threat of ant predation. One possi-
ble explanation for this difference (Morales et al. 2008) could be that the lower
frequency of signalling by Publilia was ‘honest communication’ (alerting ants to
predator pressure) but in Lepidoptera included an element of ‘dishonest communi-
cation’ that regulates their investment in ant-tending. Morales et al. cited the context
of some lycaenid larvae reducing the costs of producing a reward by ‘training” ants
to respond to signals, by using a strategy of sporadic reinforcement.
Based on their studies of Dalbulus leafhoppers (p. 215), Larsen et al. (1992)
hypothesised that the evolution of ant-Hemiptera mutualisms might have involved
the stages of (1) ants exploiting sugar sources, including honeydew deposited on
plant surfaces occupied by the bugs, and (2) the ants developing from gatherers to
harvesters, a step that necessitates the ants reducing the normal ‘escape’ behaviours
and enabling physical contact with the ants. In turn, the Hemiptera adapt to such
changes by modifying the amount and mode of honeydew production to facilitate
the harvesting.
Most of the mutualisms involving ants and noted in the previous chapter are
mediated through the reward of nectar provided by a host plant with which ants in
some way associate, and with the reciprocal that the ants are presumed to protect the
plant from herbivory. Hemiptera include many of the most serious pests of trees,
ornamental plants and field crops, and their associations with ants have attracted
wide attention in relation to their impacts and ‘control’ (Way 1963) and as a central
need of planning agricultural pest management, in particular. These associations are
also mediated through nutrition: in this case, the carbohydrate-rich ‘honeydew’

excreted by the bugs, which, in turn, are purported to be protected from predators
and parasitoids by ants seeking this resource. This protection may often increase the
detrimental effects of insects such as aphids and scale insects on desirable plants
(such as crops and ornamentals), so increasing ‘pest impacts’ through enabling
larger populations to develop. However, if the protection of the plant by the ants
counters or outweighs the herbivory from the Hemiptera, the consequent indirect
mutualism involves benefits to all three parties, the plant, ant and hemipteran.
Following Janzen (1985) and classic examples studied earlier by him (Janzen 1972,
1973), it is clear that some ants, as ‘farmers’, maintain a population of scale insects
or other bugs as a major food source. Janzen regarded the hemipterans as ‘zoologi-
cal devices used by the plants to maintain an ant colony’. High populations of
Hemiptera, correlated with high densities of attending ants, may lead to better
defence by the ants against herbivores, as shown for the ant Azteca pittieri on Cordia
alliodora in Costa Rica (Pringle et al. 2011). The ants host several species of
Coccidae and Pseudococcidae and feed on their honeydew, but without any trace of
direct predation on the hemipterans. The ants were more aggressive defenders when
provided with increased honeydew levels, and three indirect benefits to the trees
were determined: (1) larger ant colonies were associated with more hemipterans, so
that coccoids were associated with greater defence; (2) the carbohydrate supply
from the coccoids to the ants stimulated ant aggression; and (3) the abundance of
coccoids was greatest in young domatia near young foliage, so that defence of
growing foliage was especially effective.
Honeydew production by Hemiptera is very widespread and has been deemed an
ecological equivalent to nectar production from extrafloral nectaries of plants, and
the trophobiont ant associations with both these key resources are largely linked
with defence of the host plants from insect herbivores. Removal of the excess hon-
eydew produced by most Sternorrhyncha and Auchenorrhyncha is vital to the bugs
in order to avoid the producers drowning in it or succumbing to fungal contamina-
tion of this sugar-rich resource that represents egested excess from the ingestion of
plant sap in which the crtical nutritional need, nitrogen, is at much lower concentra-
tions. Hemiptera have developed various ways to redress this – wax secretions may
isolate the body of some taxa to some extent, and wax filaments may form a tube
ducting honeydew away from the body, whilst other species may kick or shake the
body to ‘throw’ the droplets away. Ants removing the honeydew, as mutualists, may
also become trophobionts by feeding on the Hemiptera themselves. For aphids,
Stadler and Dixon (2005, 2008) suggested that trophobiosis arose as a predator-prey
interaction in which the production of honeydew reduced the ants’ predatory ten-
dency, and has now led to a wide range of facultative to obligate mutualisms
(Delabie 2001) as true relationships between the participants. However, Delabie
emphasised that most knowledge of these interactions has come from direct studies
of only the two mutualistic species, and the far wider community ramifications and
interactions have rarely been examined. Even the key third party, the plant host of
the bugs, has been only a rare primary focus in studies of the mutualisms. Figure 5.3
indicates possible interactions in that wider context in which the various adjunct
Fig. 5.3 Schematic organisation of the ant-hemipteran interactions, indicating the factors provid-
ing positive (1–8) or negative (9–12) influences on the arrowed participants; negative influences
are shown by dotted lines. Numbers are as follows: 1 predation, parasitisation; 2 avoidance; 3
transport, dissemination; 4 honeydew; 5 sap; 6 nectar; 7 proteins/lipids; 8 nesting/habitat site; 9
herbivory control; 10 honeydew spray, contamination by fungi; 11 phytotoxicity; 12 plant disease
(Delabie 2001)
species will clearly differ greatly across cases and contexts. The primary mutualism
is thus a ‘hub’ for the wide network and fates of other interacting species.
Studies on all the major groups of Hemiptera involved in mutualisms have indi-
cated the wide variety of contexts present and range of ecological significance.
Treehoppers (Membracidae) may benefit in ways besides ant protection from preda-
tors. Publilia concava tended by Formica obscuriventris in the United States had
higher survivorship levels than unattended bugs (Morales 2000), even when preda-
tors were excluded. The benefits from ant attendance were greatest at low hopper
densities and increased as densities rose, irrespective of predator pressure. Maternal
care by these membracids may effectively attract mutualistic ants, with the simple
presence of female treehoppers increasing nymphal survival even without obvious
direct care. Thus ant-tending (mainly by Formica obscuripes) of Publilia modesta
in Colorado increased the numbers of surviving nymphs by about 50% (Fig. 5.4).
Maternal presence increased survival when ants were present, but not when they
were absent (Billick et al. 2001).
Fig. 5.4 Role of maternal care in nymphal survival of the membracid Publilia modesta, and influ-
ence of tending ants. When ants (predominantly Formica obscuripes) are present, maternal care
increased the number of new Publilia adults, but had no effect or a small negative impact when ants
were absent (Billick et al. 2001). Black, mothers present; open, mothers removed; numbers of
nymphs initially present shown
Relationships between Publilia modesta and Formica obscuripes are dependent

on the host plant species involved, as the other key species in the tripartite association.
Such influences may occur from both ‘top-down’ processes imposed by the ants as
predators and ‘bottom-up’ processes in which the host plant provides the operating
environment in which the ant-hemipteran mutualism can occur and variously thrive
or strive to persist. Dependence of the ants on the membracid were compared by
manipulating ant densities on the two predominant host plants of the membracid
(Chrysothamnus viscidiflorus, Wyethia spp., Asteraceae) in Colorado (Reithel and
Billick 2006). The hosts displayed different senescence patterns, which strongly
influenced the mutualism. Thus, membracid nymphs on Chrysothamnus reached
adulthood well before host senescence, but the far earlier senescence of Wyethia led
to large-scale loss (through death and movement) of nymphs and essentially pre-
cluded formation of a mutualism. Relatively few nymphs succeeded in reaching a
more suitable host plant, and the earlier loss of Wyethia results in part from the large
aggregations of nymphs that can occur. Any positive effects of ants on nymphs on
Wyethia species was eliminated by large mortality associated with senescence.
These host plant species grow intermingled and close together, and membracid pop-
ulations can mix freely amongst the plants.
The associations are fundamentally trophobiotic interactions based on conver-
gent evolution across different groups and with the most abundant forms facultative
and, so, very diverse in their roles and functions (Delabie 2001). They comprise one
of the major ‘farming systems’ of considerable value in studying mutualisms (Ivens
2015). The associations are usually asymmetric, with one partner the ‘farmer’ (host
ant) and the other mutualistic ‘symbionts’ (the favoured Hemiptera). The ant is
commonly the longer-lived partner and comprises a single reproductive entity (as a
long-lived colony), and the symbiont (terminology of Ivens 2015) is a group of
individuals with a more limited collective lifespan – perhaps for only a single sea-
son. Using Ivens’ phrase, in those interactions the interest of the partners will always
differ and ‘conflict lurks around the corner’.
Many Hemiptera show considerable host plant specificity, but their feeding
modes (which involve insertion of narrow piercing stylets to imbibe plant sap) ren-
der many of them vulnerable to attack, as they cannot withdraw or flee rapidly.
Many are also gregarious or sedentary and are thus conspicuous and an easily
accessible source of nutrients to abundant opportunist or more regularly associating
ants. The more extreme changes occur amongst more specialised associations in
which the bug (or bug group) is consistently attended by the same ant species to give
a high level of specificity. With aphids, and perhaps marking a more widespread
evolutionary path, Stadler and Dixon (1999, 2008) believed it likely that such asso-
ciations began (as noted above) through modification of predator-prey interactions
in which droplets of honeydew from the potential prey inhibited predatory behav-
iour and progressively fostered the use of this alternative food. The numbers of
Hemiptera taxa involved (especially amongst stationary gregarious taxa such as
many Coccoidea) and the quantity and high quality of honeydew (as from many
aphids) have generated numerous facultative interactions with ants and, because
honeydew is essentially a waste product through which Hemiptera eliminate excess
carbohydrates from their liquid diets, the interaction poses very little cost for them.
Many ants, often fortuitously, protect the bugs from parasitoids or predators, in
some cases in ways that reflect activity patterns. Attending ants can, for example,
protect ‘their’ Hemiptera from nocturnal generalist spider attacks (Buckley 1990).
The major effect of the attending Iridomyrmex was to reduce nocturnal mortality in
the two species of Eurymelidae studied in northern Australia.
Rico-Gray and Oliveira (2007) recapitulated the three main criteria that must be
met in order for plant protection to arise from the ants in such interactions, as (1) the
hemipteran tended by the ant should not be the main herbivore present, and others
should also be present on the plant; (2) ant activities should not permit high densi-
ties or feeding rates in the Hemiptera population and (3) the ants must be effective
removers of other, non-hemipteran, herbivores and seed predators. Essentially, the
plant’s benefit in such associations commonly depends on the intensity and dynam-
ics of the ant-hemipteran interaction. Likewise, the wider community ramifications
of those mutualisms are both widespread and ‘may alter the structure of entire
arthropod communities’ (Eubanks and Styrsky 2006). These authors considered the
interactions sufficiently important to be ‘keystone interactions’, based on the dra-
matic increase of ant abundance facilitated by the honeydew resource, and parallel
increases in suppression of both herbivores and higher trophic level arthropods. In
some cases the effectiveness of ants as biological control agents is enhanced,
because the ants remove herbivores that are far more important pests than their
hemipteran associates. Presence of the ants may be vital to allow some Hemiptera
to proliferate, by removing their natural enemies and in this context perhaps reduc-
ing local biological control by native or alien natural enemies. However, the mutual-
ism can often be beneficial in increasing predation by ants on more damaging plant
pests, and in some agricultural contexts enhances biological control. Eubanks and
Styrsky (2006) suggested that such mutualisms may be manipulable to enhance
suitable pest management – because many Hemiptera are amongst the most serious
crop pests, this suite of insect-plant mutualisms and accompanying interactions
between ants and other predators or parasitoids is one of the most intensively stud-
ied of all such contexts. In some cases, the hemipteran cannot survive unless ants
are present, because of the impacts of predators or parasitoids. The aphid

Chaitophorus populicola, for example, occurs only in association with ants, and
numbers declined substantially (by 88%) on host trees at 6 m or more from a mound
nest of Formica propinqua (Wimp and Whitham 2001), due to impacts of the ant on
the aphid’s natural enemies.
Defensive benefits of ants for Hemiptera, reviewed by Buckley (1987a, b), are
clear, but the effectiveness of different ant species can differ greatly. Observations
on 11 interacting pairs of ‘ant-hemipteran’ associations in Papua New Guinea
(Buckley and Gullan 1991) confirmed that the coccoids attended by less aggressive
ant species were more heavily parasitised than those attended by more aggressive
ants. Experimental exclusions of ants led to increased parasitisation.
Part of the importance of these associations is simply that, in involving more than
two species in mutualisms, they partially bridge studies of two-species interactions
and those of life histories of single species (Buckley 1987a), whilst their variety
provides a range of possible models for studying the interactions and the factors that
influence them. Thus, ants may prey on bugs, cause them to escape (such as by drop-
ping off the plant, as is common in many insect herbivores when they are disturbed
or threatened), remove and consume honeydew, inflict any combination of these,
and may or may not defend the bugs from predators and/or parasitoids at any stage
of their life cycle, perhaps increasing survival rates, population sizes or fecundity.
These, in turn, may affect the plant. Buckley (1987a) suggested that the tripartite
interactions may be costly to the plant and provide it with few – if any – benefits.
The details of ant predation on Hemiptera, not unexpectedly, will differ with both
species and context. It is not always clear why one hemipteran species participates
in a mutualism with ants whilst its close relatives do not. Approaches to clarifying
this have focused on mutualisms that are labile in having several evolutionarily
independent origins within a taxonomic group. Shingleton et al. (2005) advocated
the values of mutualisms between ants and aphids as useful models. In some aphid
groups, a mutualism may be lost or gained multiple times within a single genus
(such as Chaitophorus: Shingleton and Stern 2003). A study of 13 species of
Chaitophorus with varying biologies revealed a range from obligate mutualisms
with ants (6 species), through facultative mutualisms (6 species) to complete
absence of attending ants (1 species). The last species sometimes occurred on the
same host as mutualistic congenors, so that absence of mutualism did not link with
absence of ants. Relationships were demonstrated between aphid feeding position,
length of mouthpart stylets, ability to escape and risk of predation. Simply, aphid
species feeding on deeper phloem tissues (on petioles rather than leaf blades) have
longer mouthparts, and the greater depth of feeding thwarts rapid mouthpart with-
drawal and, hence, rapid escape from predators. Those traits thereby equate to a
greater need for protection by mutualistic ants.
Amongst these Chaitophorus species, differences of mouthpart lengths and feed-
ing sites were a common trait between species pairs sharing the same host plant
species – so that tending ants could also represent a means of escaping interspecific
competition by allowing the tended aphid to feed at sites not accessible to the
untended species. Shingleton et al. (2005) pondered whether the origin of ant-aphid
mutualisms, as suggested by this study, may depend on traits that evolved under
unrelated processes but together influence the benefits of mutualism. Ant-tending
has developed or been lost at least five times during the evolution of Chaitophorus,
and the genus has also made several ‘switches’ in feeding positions from leaf-
feeding to stem-feeding or the converse, as well as at least two host switches from
Populus to Salix (Shingleton and Stern 2003). These phylogenetic changes, coupled
with observations that, for example, ants will tend non-myrmecophytic species if no
other carbohydrate source is available (and this despite a preference for myrmeco-
phytic hosts if a choice is available), suggest that gaining or losing regular associa-
tion with ants is not infrequent or difficult. Shingleton and Stern suggested that the
pattern may be regarded as a hierarchy of ant preference for different aphid species,
mediated in part through honeydew quality, amount and accessibility. Differences in
species may thus be very subtle, and ant-tending is indeed very labile within this
group.
With these provisos, and the many cases purportedly showing benefits of ants to
aphids, an anomaly raised by Bristow (1991), namely, that only about a quarter (117
species) of the 479 aphid species found in the Rocky Mountain area of the western
United States are tended by ants and most species appear to forgo the possible ben-
efits of ant-guarding, has been addressed repeatedly. The major subsequent studies,
discussed by Yao (2014), have explored aspects of the interactions to show that they
are commonly unstable and that the chemical composition of honeydew had earlier
been largely overlooked as an influential factor. The trisaccharide melesitose, found
in the honeydew of ant-tended aphids, may play an important role. Yao noted that
non ant-attended aphids need only to satisfy their own nutritional needs from the
phloem sap diet, whilst ant-tended species must also provide enough honeydew to
attract and support the attendant ants, a requirement in part satisfied by sugar-rich
honeydew.
Likewise, the identity of the hemipteran can strongly influence the relationship
between the ant colony and the host plant. Data accumulated by Gaume et al. (1998)
showed that costs/benefits to Leonardoxa africana (of the segregate known as taxon
T3) depended strongly on whether the bug present was a coccid (Houardia abdita)
or a pseudococcid (Paraputa anomala), neither being host plant specific, and with
the ant colony biomass also affected. Various relationships between the three par-
ticipants were investigated, to indicate that (1) the relationship between large ant
colony size and greater protection from herbivory was more pronounced for trees
with pseudococcids than for those with coccids and (2) for trees of comparable size,
hemipteran biomass was larger on trees with coccids than those with pseudococ-
cids, so that the cost to the tree of maintaining ants may be higher with coccids, and
net benefits greater with the pseudococcid partner. Possible mechanisms included
relationships with numbers of extrafloral nectaries and nest site volume available
for ants, but the study revealed considerable subtleties over interpreting detail of
how such mutualisms function and are maintained.
Different ant species, each a mutualist, may have very different effects on their
attended Hemiptera and, as noted above, the ‘fitness’ of the trophobiont bugs will be
affected by which ant species becomes its predominant associate. In a Japanese
citrus orchard, the Citrus mealybug (Pseudococus citriculus) forms mutualisms

mainly with one or other of the two abundant ants Lasius niger and Pristomyrmex
pungens. Both are common gleaners of honeydew from pseudococcids (Itioka and
Inoue 1999). Growth rates of Pseudococcus populations were considerably higher
when either ant was present than when ants were absent, but populations became
significantly greater on trees with L. niger than those with P. pungens. In the field,
which ant attends Pseudococcus primarily reflects which ant occupies the individ-
ual tree so that, whilst the ant markedly affects the mealybug fitness, the mealybug
can only passively adopt its mutualistic partner. Interspeciifc competition between
ant candidates within the wider ant assemblage may thus affect trophobiont fitness.
When the favourable L. niger was replaced by P. pungens, the mealybug population
declined substantially in relation to one which continued with Lasius. Reciprocally,
if P. pungens was replaced by L. niger, the mealybug population increased. The dif-
ference reflected contrasting attack strategies by the two ant species on natural ene-
mies, with L. niger by far the more aggressive protector (Fig. 5.5), as investigated
by responses to ‘moving models’ using freeze-dried bodies of the ladybird
Chilocoris kuwanae, a specialist scale-eating coccinellid.
The opportunistic nature of many trophobiotic associations involving ants and
Hemiptera increases the difficulties of functional interpretation, not least because
Fig. 5.5 The responses of two mutualistic ants tending the citrus mealybug Pseudococcus citricu-
lus to artificial models of two key natural enemies of the mealybug. The outer circle in each dia-
gram represents responses to the ladybird Chilocoris kuwanae; the inner circle represents responses
to a parasitoid wasp model. Responses for (a) Lasius niger and (b) Pristomymex pungens are
shown categorised as severe attack (black), moderate attack (dotted) or no attack (open); N = 200
trials for each circle (Itioka and Inoue 1999)
considerable numbers of taxa can become involved in various ways. Thus, 28

species of ants tend 10 species of Hemiptera on the Borneo climbing bamboo
(Dinochloa trichogona, Poaceae) and form some 70 different trophobioses (Mezger
and Bluthgen 2007). The most intricate associations involved Dolichoderus ants
which, in addition to close associations with mealybugs, also attend other Hemiptera.
D. trichogona might be a keystone species for ants in the regional forest understo-
rey, by hosting hemipterans that collectively furnish an important honeydew
resource.
Some of the Dolichoderus species have adopted a unique habit leading to their
common name of ‘herdsman ants’, which consolidate the mutualisms by carrying
mealybugs to new host plants (Dill et al. 2002). At least 12 species of Dolichoderus
in south-east Asian rainforests collectively associate with more than 30 species of
Allomyrmococcini, with associated colonisation of new feeding sites and plants
aided by the ants solely forming bivouac nests rather than any more permanent or
solid structures. Study of Dolichoderus cuspidatus attending Malaicoccus formica-
rii (Maschwitz and Hanel 1985) confirmed the mutualistic nature of the interaction,
as (1) ant colonies deprived of the mealybug died out because of competition from
other ants and dependence on the bugs for food, and (2) ant-less mealybugs were
also not viable. Portage of the mealybugs is very frequent, with an average of more
than 10% of ants on trails at all times carrying Malaicoccus individuals. Any distur-
bance to the bugs may stimulate ants to move them, and the ants also react aggres-
sively to other arthropods approaching the feeding sites. Movements of nests may
occur as frequently as once or twice a week, with nest site shifts associated strongly
with changing feeding sites. This case has drawn considerable attention due to its
novelty.
References
Billick I, Weidmann M, Reithel J (2001) The relationship between ant-tending and maternal care
in the treehopper Publilia modesta. Behav Ecol Sociobiol 51:41–46
Bristow CM (1991) Are ant-aphid associations a tritrophic interaction? Oleander aphids and
Argentine ants. Oecologia 87:514–521
Buckley RC (1987a) Interactions involving plants, Homoptera, and ants. Annu Rev Ecol Syst
18:111–135
Buckley RC (1987b) Ant-plant-homopteran associations. Adv Ecol Res 16:53–85
Buckley RC (1990) Ants protect tropical Homoptera against nocturnal spider predation. Biotropica
22:207–209
Buckley RC, Gullan PJ (1991) More aggressive ant species (Hymenoptera: Formicidae) provide
better protection for soft scales and mealybugs (Homoptera: Coccidae, Pseudococcidae).
Delabie JHC (2001) Trophobiosis between Formicidae and Hemiptera (Sternorrhyncha and
Auchenorrhyncha): an overview. Neotrop Entomol 30:501–516
References 117
Dill M, Williams DJ, Maschwitz U (2002) Herdsman ants and their mealybug partners.
E. Schweizerbart’sche Verlagsbuchhandlung, Stuttgart
Eubanks MD, Styrsky JD (2006) Ant-hemipteran mutualisms: keystone interactions that alter food
web dynamics and influence plant fitness. Consequences of ant-hemipteran mutualisms for
biocontrol. In: Brodeau J, Boivin G (eds) Trophic and guild interactions in biological control.
Springer, Dordrecht, pp 171–189
Gaume L, McKey D, Terrin S (1998) Ant-plant-homopteran mutualism: how a third partner affects
the interaction between a plant-specialist ant and its myrmecophyte host. Proc R Soc Lond B
265:569–575
Gibernau M, Dejean A (2001) Ant protection of a heteropteran trophobiont against a parasitoid
wasp. Oecologia 126:53–57
Itioka T, Inoue T (1999) The alternation of mutualistic ant species affects the population growth of
their trophobiont mealybug. Ecography 22:169–177
Ivens ABF (2015) Cooperation and conflict in ant (Hymenoptera: Formicidae) farming mutual-
isms – a review. Myrmecol News 21:19–36
Ivens ABF, von Beeren C, Bluthgen N, Kronauer DJC (2016) Studying the complex communities
of ants and their symbionts using ecological network analysis. Annu Rev Entomol
61:353–371
Janzen DH (1972) Protection of Barteria (Passifloraceae) by Pachysima ants (Pseudomyrmeciinae)
in a Nigerian rain forest. Ecology 53:885–892
Janzen DH (1973) Dissolution of a mutualism between Cecropia and its Azteca ants. Biotropica
5:15–28
Janzen DH (1985) The natural history of mutualisms. In: Boucher DH (ed) The biology of mutual-
ism. Croom Helm, London/Sydney, pp 40–99
Komatsu T, Itino T (2014) Moth caterpillar solicits for homopteran honeydew. Sci Reports 4:3922.
doi:10.1038/srep03922
Larsen KJ, Heady SE, Nault LR (1992) Influence of ants (Hymenoptera: Formicidae) on honey-
dew excretion and escape behaviors in a myrmecophile, Dalbulus quinquenotatus (Homoptera:
Cicadellidae), and its congeners. J Insect Behav 5:109–122
Lohman DJ, Samarita VU (2009) The biology of carnivorous butterfly larvae (Lepidoptera:
Lycaenidae: Miletinae: Miletini) and their ant-tended hemipteran prey in Thailand and the
Philippines. J Nat Hist 43:569–581
Maschwitz U, Hanel H (1985) The migrating herdsman Dolichoderus (Diabolus) cuspidatus: an
ant with a novel mode of life. Behav Ecol Sociobiol 17:171–184
Mezger D, Bluthgen N (2007) Trophobioses on Borneo climbing bamboo – diversity and ecology
of ant-hemipteran associations on Dinochloa trichogona (Poaceae). Asian Myrmecol 1:59–68
Morales MA (2000) Mechanisms and density dependence of benefit in an ant-membracid mutual-
ism. Ecology 81:482–489
Morales MA, Barone JL, Henry CS (2008) Acoustic alarm signaling facilitates predator protection
of treehoppers by mutualist ant bodyguards. Proc R Soc B 275:1935–1941
Pringle EG, Dirzo R, Gordon DM (2011) Indirect benefits of symbiotic coccoids for an ant-
defended myrmecophytic tree. Ecology 92:37–46
Reithel JS, Billick I (2006) Bottom-up mediation of an ant-membracid mutualism: effects from
different host plants. Evol Ecol 20:27–38
Rico-Gray V, Oliveira PS (2007) The ecology and evolution of ant-plant interactions. University of
Chicago Press, Chicago/London
Shingleton AW, Stern DL (2003) Molecular phylogenetic evidence for multiple gains or losses of
ant mutualism within the aphid genus Chaitophorus. Molec Phylog Evol 26:26–35
Shingleton AW, Stern DL, Foster WA (2005) The origin of a mutualism: a morphological trait
promoting the evolution of ant-aphid mutualisms. Evolution 59:921–926
Stadler B, Dixon AFG (1999) Ant attendance in aphids: why different degrees of myrmecophily?
Ecol Entomol 24:363–369
Stadler B, Dixon AFG (2005) Ecology and evolution of aphid-ant interactions. Annu Rev Ecol
Evol Syst 36:345–372
Stadler B, Dixon AFG (2008) Mutualism. Ants and their insect partners. Cambridge University
Press, Cambridge
Styrsky JD, Eubanks MD (2007) Ecological consequences of interactions between ants and
honeydew-producing insects. Proc R Soc Lond B 274:151–164
Way MJ (1963) Mutualism between ants and honeydew-producing Homoptera. Annu Rev Entomol
8:307–344
Wimp GM, Whitham TG (2001) Biodiversity consequences of predation and host plant hybridiza-
tion on an aphid-ant mutualism. Ecology 82:440–452
Yao I (2014) Costs and constraints in aphid-ant mutualism. Ecol Res 29:383–391
Chapter 6
Classic Themes: Ants
and Other Insects – Lepidoptera
Keywords Acrodipsas • Exocrine glands • Host specificity • Lycaenidae •

Maculinea • Mutualism • Myrmica • Myrmecophily • Parasitism • Stridulation •
Tripartite associations
6.1 Introduction: Ants and Lycaenid Butterflies
Perhaps the most intensively studied and informative suite of interactions involving
insect mutualisms is the enormous variety of associations between ants and the
larvae of lycaenid butterflies. Lycaenidae are by far the predominant group of myr-
mecophilous Lepidoptera, but mutualistic associations have developed also in sev-
eral groups of moths (Hinton 1951). The parallels between the tripartite interactions
of ‘ant-hemipteran-plant’ and ‘ant-lepidopteran-plant’ both reflect ecological com-
plexities ranging from near independence to obligate mutualisms, with a continuum
of relationships categorised broadly as ‘myrmecophily’ bridging these extremes.
Associations between ants and lycaenid butterflies have been represented as ‘an
ideal system for research on mutualisms’ (Pierce et al. 1991), because of the exten-
sive collective range of associations and levels of intimacy and intricacies they
display.
Amongst species with larvae that associate with ants, relationships may be mutu-
alistic, parasitic, commensal or combinations of these (De Vries 1991).
The Lycaenidae (now treated by many workers as including the formerly
accepted family Riodinidae, as the subfamily Riodininae) thus exhibit many levels
of obligate or facultative associations between their early stages and ants (Hinton
1951), with advances in understanding over the subsequent half century discussed
by Pierce et al. (2002). Larvae of some species have developed to assume more
antagonistic habits as predators or parasites after they become obligate feeders on
ant brood once they are sheltered within the ant nest. Lycaenidae (the ‘blues’, ‘cop-
pers’ and ‘hairstreaks’) is the largest family of butterflies and contains more than
6000 species - about a third of all butterfly species. Numerous taxa are, in one way
or another, myrmecophiles that are variously mutualistic, parasitic or simply com-

DOI 10.1007/978-3-319-58292-4_6
120 6 Classic Themes: Ants and Other Insects – Lepidoptera
mensal with ants, but collectively they exhibit an enormous range of nuances and
specialisations that continue to be studied and elucidated.
Studies of associations between ants and Lycaenidae, however, have become one
of the dominant contributions to understanding the functions and evolution of
mutualisms, not least because of the popularity of these butterflies amongst hobby-
ists and naturalists, who have elucidated many of the critical details of butterfly life
histories during their observations – with myrmecophily a theme of great variety,
novelty, fascination and curiosity - and their values as tools for serious study by
ecologists. Those associations are also critical considerations in conservation plan-
ning: myrmecophilous lycaenids comprise a high proportion of signalled threatened
butterfly species. Collectively, these interactions have ‘developed into a paradig-
matic example of the evolutionary ecology and dynamics of interspecific associa-
tions’ (Fiedler 2012), with the mutualistic benefits sometimes highly asymmetrical.
Some species of ant-associated lycaenids are essentially ‘parasites’ of the ants and
especially susceptible to risks of coextinction. Indeed, many of the most threatened
species of Lycaenidae (such as the Palaearctic Large blues, Maculinea spp., some-
times treated as Phengaris spp.) fall into this category. ‘Parasitism’ includes two
main contexts of interaction, as (1) the lycaenid larvae feeding on ant brood within
the nests, as predators, and (2) the larvae feeding directly by trophallaxis from the
host ants, and commonly termed ‘cuckoos’. Illustrating the variety of possible inter-
actions, Fiedler (2012) noted the ambiguities of interpreting some other contexts
that could be considered parasitic. Amongst these are (1) myrmecophilous lycae-
nids feeding wholly and obligately on myrmecophytic ant plants, such as Arhopala
on Macaranga (p. 128) and regarded as competitors of the ants rather than partners,
and (2) species with larvae that feed on ant-attended Hemiptera and which often
also feed on honeydew. In both these cases, competition occurs with the ants for the
same resources (as trophobionts), but the larvae do not prey on ant brood or benefit
from trophallaxis. Collectively, only 11 genera of ants, of the 54 genera known by
Fiedler to attend Lycaenidae in some guise or other, are hosts of the parasitic but-
terflies. Those parasites, however, are a small component of myrmecophilous
Lycaenidae but highly significant in conservation practice and advocacy.
6.2 Myrmecophily
The extensive account and overview of lycaenid myrmecophily by Fiedler (1991b)

is a vital source of information on phylogenetic and geographical distribution of
myrmecophily in its various forms. As had some previous authors, Fiedler empha-
sised that adaptations of lycaenid larvae for myrmecophily fall into two categories
of (1) passive protection from ant attack by possession of a thick larval cuticle and
the characteristically flat ventral surface that can be appressed to the substrate and
(2) active exocrine glands (‘myrmecophilous organs’) with secretions that variously
attract and/or appease ants through chemical signals or provision of nutrients.
6.2 Myrmecophily 121
Three main categories of myrmecophilous organs are recognised widely. All are
diverse and distributed irregularly amongst different lycaenid taxa. They are:
1. Pore cupola organs (PCOs) with varied and debated roles including functioning
in avoidance of ant attacks and appeasement – but differing considerably across
taxa – and suggestions of species specificity (and, even, instar-specific effects,
Polyommatus) in roles, as reflected in variable attractiveness to ants
2. Dorsal nectary organs (DNOs), especially important in maintaining stable ant-
larva associations but, again, variable across taxa
3. Tentacle organs (TOs), eversible structures that can attract ants through releasing
volatile signals
Brakefield et al. (1992) commented that ‘they release a cocktail of volatile phero-
mones which alarm and activate the ants’. Discussed also by De Vries (1991), other
specialised structures have also been designated, including vibratory papillae found
on some myrmecophilous riodinines which produce sounds that appear to attract
ants, and dish organs and bladder setae with less well-defined roles. Dish organs
produce a liquid imbibed by ants. More generally, nutritional rewards produced by
lycaenid larvae are an important source of nitrogen for attending ants and may link
also with larval food plant choices. Nitrogen–rich food plants have long been known
to be especially desirable to many insect herbivores and are actively sought and
selected. Both myrmecophilous (broadly, with some form of mutualistic or other-
wise symbiotic association with ants) and myrmecoxenous (broadly, those showing
avoidance of ant aggressiveness) lycaenids actively preferred plants with higher
nitrogen content (Pellissier et al. 2012), and feeding on these may thus confer some
fitness advantage to both lycaenid categories, as well as for some other butterflies
that benefit through increased growth or survival rates. However, for those lycaenids
that feed directly on ant brood, the evolutionary ‘pressure’ to use nitrogen-rich
plants may become relaxed, allowing those taxa to colonise alternative, less nutri-
tious, plants.
Approximately 60% of ant-associated species are facultative myrmecophiles that
are attended by several different ant species (Trager and Daniels 2009). Some asso-
ciations are very widespread amongst the species involved. Larvae of the Isola blue
butterfly (Hemiargus isola) in southern North America feed on a variety of legumi-
nous host plants, and second-fourth instars are tended facultatively by several ants,
predominantly single species each of Formica and Forelius. Over 2 years of larval
sampling (Weeks 2003), ants were associated with almost all (88–99%) of larvae
collected, and, whilst a single species predominated (tending >80% of larvae), the
ant species involved differed in the 2 years. Experimental removal of either ant dem-
onstrated their substantial role in reducing parasitisation by the braconid wasp
Cotesia cyaniridis, probably by deterring wasp oviposition in early larval instars.
Facultative myrmecophily thus enhanced larval survival to a similar level whichever
predominant ant occurred.
Aspects of behavioural and chemical communication between lycaenid larvae
(and pupae) and ants, and the diverse evolutionary ramifications involved, render
this suite a major source of ideas and understanding on mutualistic interactions.
More than 200 years of constructive observations contribute to this awareness

(Pierce et al. 2002). Reflecting the significance of butterflies in conservation, and in
contrast to many other mutualisms discussed in this book, understanding of myr-
mecophily has played direct roles in conservation management in which the
idiosyncrasies of individual lycaenid species must be heeded, with needs of a spe-
cific host ant sometimes a key adjunct to those of the butterfly itself (Chap. 9). The
interactions also involve adoption of aphytophagy (Cottrell 1984), itself an impor-
tant conservation theme that can add further complexity to optimal resource man-
agement and supply, and provide challenging dimensions to understanding sound
practical conservation (New 1993). As a family, Lycaenidae (s.l.) participate in a
wider range of ecological interactions than any other group of Lepidoptera, and
increasingly informative and elegant searches for transitions and evolutionary path-
ways continue.
Overall, perhaps more than 99% of all Lepidoptera are wholly phytophagous
(Lohman and Samarita 2009). However, specific predation by Lepidoptera on
Hemiptera or ants, with consequent complexities in their life cycles, is associated
commonly with those Lepidoptera being rare – and some Lycaenidae are of high
conservation concern as ‘endangered’ (Pierce 1995). Protection by ants may afford
lycaenid larvae enemy-free space. More broadly ants are ‘important research sub-
jects in the efforts to restore endangered lycaenid butterfly species’ (Pascale and
Thiet 2016). Despite many associations being facultative rather than obligatory,
access to suitable ants may still be a key factor in the survival of threatened butterfly
populations in danger of succumbing to habitat loss or disturbance.
The endemic Australian imperial hairstreak, Jalmenus evagoras, is one of the
most intensively investigated myrmecophilous lycaenids and has been described as
a model system for elucidating some aspects of its mutualism. Much earlier work on
this species was summarised by Pierce and Nash (1999), whose review reflected
studies that included 7 doctoral theses and 11 honours/senior theses on this butter-
fly. Butterfly colonies can persist on individual trees of their food plants, any of
several species of Acacia, for more than 20 years. The gregarious larvae are pro-
tected by ants and provide the ants with a food reward. Ant exclusion experiments
(Pierce et al. 1987) showed that individual larvae without attendant ants were
unlikely to survive. Several different attending ants are found, the most frequent
being the species of Iridomyrmex, with names such as I. anceps and I. rufoniger
cited in various publications, but in some cases reflecting more diverse species
groups. Both larvae and pupae of Jalmenus have specialised adaptations that encour-
age ant attendance, with both stages having stridulatory mechanisms. Travassos and
Pierce (2000) found that the larvae produce three different kinds of calls, emitted in
different contexts – either spontaneously when disturbed, when other larvae are
present, or in the presence of attendant ants. The calling pattern therefore varied in
relation to ant attendance and, in conjunction with tentacular organ eversions, may
jointly modify ant behaviour – stridulation may thereby be a component of defence.
Pupae of J. evagoras have two call elements, also produced in different contexts and
important in maintaining high ant attendance and recruiting greater number of
workers. Sophisticated communication may thereby occur in this (and, indeed,

other) myrmecophilous relationships.
J. evagoras is a common species distributed widely in eastern Australia, so pro-
viding considerable opportunities for manipulative investigations – a marked con-
trast to many of the myrmecophilous lycaenids of primary conservation interest.
The extensive experimental trials on J. evagoras demonstrated, amongst other
things, that:
1. The survival of larvae was far greater on trees with ants, because of the ‘enemy-
free space’ provided, with two of the three main parasitoids having far lowered
levels of attack. The exception, a braconid parasitoid of larvae, used the ants as
oviposition cues and so overcame the putative defences. Larvae tended by ants
develop about 5 days faster than untended larvae, so decreasing the time they are
exposed to natural enemies.
2. The butterfly weights were smaller when reared with ants than without them – a
complex ‘cost’ to the butterfly – but compensated by realised fecundity of
females being greater when ants were present. Ant-tended female larvae pupate
and eclose earlier, perhaps giving them some advantage in selecting oviposition
sites.
3. The ants gain food from larval secretions, where it was shown that a considerable
proportion of the food of an attendant colony (of I. anceps) was provided.
4. The major cost to the ants is the metabolic expense of finding, exploiting and
defending the larval food source.
A further benefit may flow from the gregarious feeding behaviour of J. evagoras
(and some other myrmecophilous lycaenids), in that the intensive feeding may
induce some form of ‘nutrient sink’ increasing the supply of soluble protein (Pierce
et al. 1987).
Ramifications of the protection of mutualist lycaenid larvae by ants are indeed
extensive. They may, for example, facilitate host plant range extensions to allow the
butterfly to exploit additional or less suitable hosts. Survival of larvae of Lycaeides
melissa on alfalfa (Medicago sativa, a novel widespread host in California) was
increased in the presence of ants, not all identified specifically but including at least
five species of Formica (Forister et al. 2011). Amelioration of predation by these
attending ants was believed to facilitate the use of this novel host – which is, in any
case, a poor host for L. melissa – so that populations could probably not be sustained
solely on that host, and continuing buffering against predation is vital for it to persist
there. Survival levels across treatments varied considerably (Fig. 6.1), but the
absence of ants on M. sativa branches was commonly associated with lower larval
survival.
Many ecologists have acknowledged that specificity of interactions between
Lycaenidae and ants may play important roles in butterfly evolution, such as the
mutualisms with ants accompanying, in many cases, specialisations for larval food
plants. Those specialisations may extend beyond the most widely understood level,
of species, to populations. Some butterflies are necessarily associated with different
ant species in different parts of their range, and butterfly dispersal between these is
Fig. 6.1 The proportional

survival of larvae of the
lycaenid butterfly
Lycaeides melissa placed
on the novel host
Medicago sativa with ants
or with ants excluded. Ten
replicate trials are shown,
each with (black) and
without (open bars) ants
on a single plant and eight
experimental branches.
Note that replicates 5 and 7
had no surviving larvae
when ants were absent
(Forister et al. 2011)
difficult (or impossible), fostering local differentiation in which the populations

may also come to differ in the larval host plants used. One informative study
involved the silver-studded blue, Plebejus argus, and its associated Lasius ants
(Jordano and Thomas 1992). Two populations in North Wales were associated natu-
rally with different ants (respectively, Lasius niger and L. alienus), and experimen-
tal trials compared ant responses to larvae in the alternative associations. Larvae
were slightly more attracted to their natural host ant so that, whilst any Lasius within
this species group is attracted to any P. argus larva, finer-scale differentiation may
hone the behavioural interaction involved.
Host ant specificity is perhaps most pronounced amongst ‘parasitic’ and preda-
tory lycaenids, with investigations of these associations providing critical informa-
tion for conservation of taxa such as Maculinea in Europe (Chap. 10) and Acrodipsas
in Australia. Whilst it is tempting to regard such apparently one-sided relationships
as not being ‘proper mutualisms’, it seems that the ants do indeed benefit - at least
for some period – from the secretions provided by larvae, but these interactions are
undoubtedly amongst the ‘most asymmetrical’ of such mutualisms. As Fiedler et al.
(1996) put it, in the spectrum of mutualisms between ants and lycaenids, ‘benefits
to ants range from minimal to substantial food rewards’. Some may be rather tran-
sient and restricted to a short phase of the larval life. For the British Large blue,
Maculinea arion, Porter et al. (1992) showed that ants discovering a larva at the
stage at which it needs to be transported into the nest (as early in the fourth instar)
lick the pore cupola and dorsal nectary organs and recruit other ants. Tending the
larva continues for several hours, after which it is carried into the nest. That pro-
longed contact may be a means by which the larva gains the ‘scent’ of the ant colony
as a possible future defence.
The apparently relatively small benefits to ants from these associations have been
discussed elsewhere. Bronstein (1998) pondered whether ‘the benefit [is] some-
times so negligible that it would be inappropriate to consider the interaction to be a
mutualism’. An attempt to address this (Cushman et al. 1994) involved the endemic
Fig. 6.2 Survival of

Paralucia aurifera larvae
reared with (dashed line)
or without (solid line)
workers of the ant
Iridomyrmex nitidiceps;
duration of each instar is
shown (Cushman et al.
1994)
Australian bright copper, Paralucia aurifera, in which the survival rate of host ants
increased by 40% when secretions from the butterfly larvae were available. In the
field, larvae persisted only on plants (of the sole host, Bursaria spinosa,
Pittosporaceae) colonised by the host ant, Anonychomyrma sp. nitidiceps group
(formerly known as Iridomyrmex nitidiceps). In turn, the ants nest only at the bases
of plants colonised previously by P. aurifera, itself likely to be a cue to future ben-
efit (Cushman et al. 1994). In laboratory conditions, some pronounced differences
were found between butterfly larvae reared with ants and without them, so that the
major balance of the association may favour the lycaenid. Larvae with ants devel-
oped into heavier pupae (by 20% from larvae that developed 31–76% heavier) and
developed over a shorter period (by 69%) from larvae that developed 37% faster, in
part through undergoing one or two fewer instars than untended larvae.
Developmental differences were evident throughout larval life (Fig. 6.2). As in
many other cases, the mutualism appears asymmetrical – P. aurifera is host plant
and ant specific, and Anonychomyrma obtains food also from other sources.
Nevertheless, the ants clearly gain from this association.
Larvae of Acrodipsas (the genus known as the ‘Ant blues’ in Australia) and
Maculinea, amongst others, pass much of their developmental period in under-
ground nests of their hosts. Maculinea larvae are initially phytophagous, with the
first three instars free-living but growing little during that phase, thereafter being
carried by host ants into the nest, where they feed on ant brood and pupae, staying
there for 9–10 months or more (Barbero et al. 2013, noted 11–23 months within the
ant nests, as larva and pupa, reflecting different growth rates). Two species are fed
trophallactically by the host ants, as ‘cuckoos’, and a proportion of these (M. rebeli,
M. alcon) take 2 years to develop, as does the predatory M. arion (and also the
predatory hoverfly Microdon mutabilis: Schonrogge et al. 2000), so that develop-
mental polymorphism in growth rates may be adaptive in some forms of myr-
mecophily (Thomas et al. 1998). Larvae of some species of Acrodipsas are wholly
myrmecophagous, with the first instar larvae of A. myrmecophila and some others
being carried into the ant nest soon after hatching. Each may have very specific host
requirements.
Fig. 6.3 Surveys of host

ants adopted by European
species of Large blue
butterflies, Maculinea spp.
For each of the five
butterfly species,
percentage adoption by
different Myrmica ants
species is shown, together
with sample size
(N = number of emerging
adults, pupae and fully
grown larvae found)
(Thomas et al. 1989)
Each Maculinea species largely depends on a single species of Myrmica ant

(Thomas et al. 1989), a finding that invalidated the view most commonly held until
then that larvae could parasitise any species of that genus. The perspective furnished
by Thomas et al. included all five European Maculinea species, encompassing con-
tents of 994 Myrmica nests and several hundred other ant nests on sites occupied by
the butterflies. Their major results, summarised in Fig. 6.3, have been referred to
repeatedly as a template for specificity patterns since they were published. Maculinea
larvae mimic host ants (below), and Thomas et al. (1989) suggested that adaptations
such as different pheromones, larval behaviours and the position adopted in the ant
nest may be associated with this specificity. Nevertheless, recognition of this speci-
ficity revolutionised thinking over conservation needs, with the major transforma-
tion from needs from (1) flowering food plants to coincide with abundant Myrmica
of any species to (2) the much finer restriction of the flowering food plant to co-
occur with a particular species of Myrmica, with additional considerations of the
suitability of the local microclimates (influenced, for example, by sward height and
plant successions) for that focal host.
However, some host patterns for Maculinea are far more complex, and the pri-
mary host ant of most species varies cross the butterfly’s range. Barbero et al. (2013)
summarised accounts demonstrating that the five European species may be locally
adapted to species of Myrmica earlier believed not to host them. Maculinea
nausithous is the only species that indeed has only a single host ant (Myrmica rubra)
over most of its range, but it still associates with M. scabrinodis in Transylvania
(Tartally et al. 2008). Both M. alcon and M. rebeli can exploit more than one host
ant on the same site (Als et al. 2002). Up to eight species of Myrmica occur on some
European Maculinea sites, and all may carry larvae into their nests but with marked
differences in outcomes because many larvae do not develop further and never reach
adulthood. In Denmark, Als et al. (2002) found only three Myrmica species within
2 m of the Large blue larval food plants. However, five populations of Maculinea
alcon differed in their host range. One population exclusively used Myrmica
ruginodis, one predominantly used M. rubra, and the other three used both hosts.
The third ant present (Myrmica scabrinodis) was not utilised. Earlier appraisal of M.
alcon using different ants along a south-north gradient in Europe (Elmes et al. 1994)
has led to suggestions that the butterfly might comprise a group of cryptic species or
host races. No such simple gradient (or cline) was found in Denmark, but the varia-
tions were associated with demographic differences across populations. Patterns of
host ant specificity in Maculinea appear to have evolved largely at the local level
rather than at any larger regional or global scales – and Tartally et al. (2008) regarded
this as unsurprising because of the differing availability and abundance of possible
host species varying so widely across the greater region.
Although Maculinea spp. are amongst the most intensively studied butterfly
groups, with much of that study motivated directly by conservation needs and the
complexities of their myrmecophilous associations, many gaps in detailed under-
standing persist. The emphasis on the five major European taxa has not yet been
paralleled by comparable work on the eastern Palaearctic members of the genus
(Als et al. 2004). Even for Europe, however, recent review of biology of M. arion
(Hayes 2015) emphasises this complexity. As examples, geographical variations in
mimicry of different Myrmica species, the origins of larval chemical profiles, intri-
cate details of stridulatory signals and fuller investigation of oviposition cues for the
female butterflies were all noted as needing further investigation.
Maculinea females have been thought to detect host ant odours before laying, to
assure eggs are being laid near a suitable host. They may also detect that overcrowd-
ing of eggs on plants near host nests may be avoided by assessing egg loads.
Discussed by Thomas and Elmes (2001), the two hypotheses on oviposition site
selection proposed for M. alcon by van Dyck et al. (2000) were taken as mutually
exclusive. These hypotheses were as follows: (1) patterns in egg distribution are
chance effects related to microhabitat variations that influence flowering date of the
host plants and Myrmica distributions, leading to patterns in which most ‘early
eggs’ are placed near one ant species, and most ‘late eggs’ laid near a second species
and (2) ovipositing females detect ant odours and initially select plants near ant
nests but later reject these as egg load increases, subsequently selecting sites near
nests of a second Myrmica species. Tests suggested that the first hypothesis was
more likely to ‘explain’ oviposition patterns, and this was recommended for conser-
vation models of the butterflies.
Fig. 6.4 Responses of the ant Tapinoma caespitum to simulated attack on larvae of its tended
lycaenid butterfly Polyommatus icarus. A number of ants are shown immediately before and 5 min
after simulated attack, together with sample size; for controls, no such attack occurred and the
larvae were left undisturbed (Leimar and Axen 1993)
In general, female myrmecophilous Lycaenidae might be expected to use ant

presence and availability as cues in selecting plants for oviposition and benefit from
ant protection of their offspring maximising chances of their survival. Ant protec-
tion may also extend, in some cases, to the eggs – although this is far rarer and some
trials (Bachtold et al. 2014) revealed that, although eggs of both Rekoa marius and
Allosmaitia strophius were more abundant on host plants with ants, parasitisation of
the eggs was also considerably higher. Mechanisms of butterfly attraction for ovipo-
sition can be complex. Maculinea arion females ready to lay are attracted to flow-
erheads of Origanum vulgare in Italy by the volatile monoterpenoid carvacrol, with
markedly increased concentrations of this associated with disturbances of the plant
roots by Myrmica ants (Patricelli et al. 2015). Unusually, and the first documented
example of the same volatile chemical signal influencing both trophic and parasitic
stages of the same butterfly species, the female butterfly can thus select its host plant
for both larval food and later benefit as the larvae become parasitic on the ants.
Exposure of lycaenid larvae to a simulated attack, such as by pinching them
lightly with forceps, can lead to increased larval secretions and increased ant atten-
dance, with the number of reward droplets secreted making the interaction more
favourable to the ants (Leimar and Axen 1993, from observations on Polyommatus
icarus and Tetramorium caespitum). In control trials, the larvae were left undis-
turbed, but ant attack led to increased ant visitations (Fig. 6.4).
Avoidance of ant attack by active defenders of myrmecophytes sought by lycae-
nid larvae can be very subtle, with intricate chemical means developed by parasitic
lycaenids in order to feed on those plants. Chemical mimicry of the obligate mutual-
ist ants (as for host ants by myrmecophilous predatory or cuckoo species) has been
suggested, but the variations amongst species can be substantial, as demonstrated
for Arhopala on Macaranga (Inui et al. 2015). Three co-occurring butterfly species,
Fig. 6.5 The responses of

ants to larvae of three
species of Arhopala
butterflies (a, A. dajagaka;
b, A. amphimuta; c, A
zylda) on trees of three
species of Macaranga
myrmecophytes in
Malaysia (1, M. rufescens;
2, M. trachyphylla; 3, M.
beccariana). Ant responses
are shown as attack
(black), attending (dotted)
or disregarding (open), and
a number of trees/ant
colonies tested are given in
parentheses (Inui et al.
2015)
each associated with a different Macaranga food plant in Sarawak, were also asso-
ciated with different Crematogaster ants. Responses of the ants to larvae of each
lycaenid species introduced to ‘their’ tree differed clearly (Fig. 6.5), but attacks by
ants associated with their normal host plant were infrequent. Chemical cues from
larval cuticular hydrocarbons differed, and their recognition (or otherwise) by the
ants as resembling their own cuticular chemicals influenced the interaction. Larvae
of A. dajagaka were accepted by ants on M. rufescens by recognising the composi-
tion of the hydrocarbons. Those same chemicals increased ant attacks on other tree
species, implying that A. dajagaka was there recognised as an intruder – but the
abundant nectar from the larvae apparently induced ‘tolerance’ rather than attack. In
contrast, Arhopala amphimuta did not have matching hydrocarbons and was often
attacked by ants other than those on its normal host plant. Its larvae were regarded
as intruders on non-hosts. The third lycaenid, Arhopala zylda, was mostly ignored,
and larvae were attacked less frequently on all three Macaranga spp., suggesting
that this butterfly is myrmecoxenous. It had far fewer larval hydrocarbons than
either of the other species. However, such chemical mimicry, well known as a nest-
mate recognition process amongst ants, may be widespread also amongst myrmeco-
phytic taxa. Larvae of the Japanese oak-blue, Narathura japonica, feed on Quercus
and are attended by several ant species, including the parthenogenetic Pristomyrmex
punctatus. Workers of this species learn to recognise the specific cuticular hydrocar-
bons of the butterfly larvae and associate these with food rewards (Hojo et al. 2014).
This may be an important component of selecting optimal partners and is probably
more widespread amongst ant protection mutualisms.
Arhopala larvae feeding on Macaranga almost always encounter defending
Crematogaster ants, so are ‘obligately myrmecophilous’ (Shimizu-kaya et al. 2013).
Of the species noted above, A. zylda larvae lack a dorsal nectary organ, so fitting the
definition of myrmecoxenous (Kitching and Luke 1985). They do not attract plant-
ants or form any distinctive associations with them.
Such patterns in cuticular chemicals are paralleled in profiles of Maculinea lar-
vae and ant larvae and are important in conservation. For Maculinea alcon, Nash
et al. (2008) indicated that the geographical patterns of this chemistry were part of
a continuing ‘arms race’ between the butterfly and its hosts, with greater matching
of the chemistry facilitating exploitation of the ant colonies, and the ‘race’ mediated
by (1) the butterflies evolving better mimicry and (2) hosts improving their recogni-
tion of the parasites.
Chemical signals are amongst the most important means of communication by
lycaenid larvae to ‘appease’ potentially attacking ants or in seeking ‘adoption’ by
rendering the larvae attractive to the potential host ants. Together with nectar secre-
tion as a nutritional reward, chemical signals are widespread and ecologically
diverse in their roles and impacts.
Larvae and pupae of some lycaenids produce stridulations, substrate–borne
vibrations that may attract guarding ants, so that the three complementary approaches
to avoiding ant attack in mutualisms reflect (1) a variety of chemical signals, (2)
mutual rewards and (3) mechanical stimuli that can attract and moderate the number
of ants in attendance. The larval need is clearly to gain acceptance by the host ants,
so that the ants must recognise, tend and nurture the larvae – whilst the larvae in
general do not have to seek out the host ants but simply be present in a suitable
environment that may, in some way, have been selected for oviposition by the parent
butterfly. Those associations are, at the least, potentially mutualistic. However, for
many species, once that initial encounter is accomplished and stabilised, larval stim-
uli such as secretion rates and amounts may decrease, a process studied by De Vries
(1988) in Thisbe irenea and termed by him ‘enticement and binding’. Some other
associations are more difficult to categorise. Secretions of the dorsal nectary organ
of larvae of Narathura japonica reduce activities of the attendant ant, Pristomyrmex
punctatus, implying that the interaction may be parasitic rather than mutualistic
(Hojo et al. 2015).
‘Communication’ with the ants and manipulation of their behaviour are comple-
mentary strategies, with a need also for a balance between (1) investment in com-
munication and food rewards and (2) the range of target ants, so that the needs for
signal specificity differ across taxa and environments (Fiedler et al. 1996).
Communication systems tend to be relatively unspecific in facultative mutualisms
with any of a range of potential ant partners, but are necessarily very specific to
maintain obligate host-specific associations. As one example, Barbero et al. (2013)
noted that partner Maculinea larvae mimic the acoustic signals of the queen of their
specific Myrmica host, a tactic that allows them not only to become integrated into
the ant society but also to gain ‘high rank’ status. Thus, Barbero et al. showed that
larvae of Maculinea rebeli are fed in preference to host ant (Myrmica schenki) lar-
vae if food is scarce and, if the colony is disturbed, are rescued ahead of the ant
brood. It seemed that acoustic cues mediated this preference status. Acoustic com-
munications play large and varied roles in ant social behaviour and, in general, can
help to convey complex information. Historically, most lycaenid larval stridulations
have been considered to be defensive, and, also generally, the stridulations of many
mutualistic lycaenids do not obviously mimic those of the host ants. Queens of
Myrmica schencki produce unusual stridulations that attract workers, seemingly the
mimetic base for the larvae of M. rebeli to exploit. The variety and specificity of
acoustical signals by Maculinea larvae have been documented quite extensively (De
Vries et al. 1993). Although produced by different mechanisms, the acoustic calls of
Maculinea larvae and host ants can show considerable similarities – and De Vries
et al. suggested the likelihood of general mimicry – but found little or no detailed
convergence between larvae and their predominant host species.
Maculinea rebeli larvae are fed by trophallaxis, requiring persistent and effective
association with the ants, and the mimicry hypothesis was supported by experimen-
tal trials on the responses of M. schencki workers to the butterfly larvae (Akino et al.
1999): almost all larvae adopted by other ants are killed. Similar mimicry was inves-
tigated for Niphanda fusca, also a trophallactic feeder, on Camponotus japonicus in
Japan (Hojo et al. 2009). First instar Niphanda feed initially on honeydew from
aphids tended by the ant, but third instars are taken into the ant nests, where they are
nurtured to adulthood over about 10 months. Cuticular chemicals of the larvae
changed from before to after ant adoption and increasingly resembled those of the
male ants.
The gross structure of the local environment may influence the structure and
incidence of ant-lycaenid mutualisms on the same host plants. In Brazil, incidence
of lycaenid larvae on Banistereopsis malifolia (Malpighiaceae) was much higher
(abundance, fivefold; richness, sixfold greater) in a disturbed open area than in a
protected savanna (Bachtold et al. 2016). More attendant ants occurred in the dis-
turbed area, where host plants were generally taller and more intensively flowering.
The lycaenids were florivorous species with apparently facultative myrmecophilic
associations, for which floral resources were clearly favoured by disturbance.
Ants can benefit from the presence of attending lycaenid larvae in facultatively
myrmecophilous associations. Even when the larvae produce only small amounts of
secretion, this may be sufficient to promote ant survival substantially. In laboratory
trials with workers of three ant species (Myrmica rubra, Lasius niger, L. flavus), ant
mortality was 3–5 times higher when confined with larvae of lycaenid species lack-
ing a ‘nectar organ’ (Lycaena phlaeas, L. tityrus) than those with this structure
(Polyommatus icarus, P. bellargus, Aricia agestis) (Fiedler and Saam 1995). These
outcomes, although from groups of foraging workers rather than the ant colony
itself, suggested that outcomes would translate into wider effects on colony fitness.
Nutritional content of lycaenid larval secretions can be significant, as displayed for
larvae of Polyommatus coridon (Fiedler and Maschwitz 1989). Using various
assumptions to aid estimates, each larva could produce 22–44 μl of dorsal nectary
organ secretion, containing 3.5–7 mg of carbohydrates that represent a physical
energy equivalent of 55–10 J. This could contribute significantly to the nutrition of
ant colonies, especially when lycaenid density is high. In Central Europe, where
that study took place, P. coridon is one amongst at least five other myrmecophilous
lycaenids which collectively comprise a substantial potential carbohydrate resource
for attending ants.
The variety of responses by different ants to lycaenid larvae was illustrated
through experimental trials with Thisbe irenea (Riodininae) in Panama, in which De
Vries (1991) allowed 26 species of locally common ants (across six subfamilies) to
encounter larvae on potted plants. Responses of some ants were equivocal, but oth-
erwise three general modes of encounter were seen, as (1) 14 of the 26 species
ignored larvae completely, so were neither mutualists nor predators; some were
specialist predators elsewhere or general scavengers; (2) four species attacked the
larvae and were clearly predators; and (3) six species investigated the larvae and
accepted secretion rewards from them, so were regarded as potential mutualists.
As the number of species for which life histories are understood in detail
increases, appreciation of the extent and variety of myrmecophily has also increased.
Thus, Downey (1962) noted that the life histories of more than 800 species had then
been characterised to varying extents and 245 of those had myrmecophilous larvae.
More recent synthesis (Pierce et al. 2002) showed that about 75% of the 665 species
by then studied in detail are myrmecophiles, with 30% of those obligate and attended
by single ant species. These authors also reaffirmed earlier opinion that obligate
ant-lycaenid associations are more common in the southern hemisphere, where sev-
eral ant-attended lineages have proliferated. The taxonomic distribution of myr-
mecophily amongst lycaenid tribes, based on analysis of all information available
on European and North American fauna, led Fiedler (1991a, b) to distinguish sev-
eral categories of interactions (Table 6.1), with >75% of the species myrmecophi-
Table 6.1 The broad categories of larval myrmecophily of Lycaenidae recognised by Fiedler
(1991a) as an initial grouping based mainly on field records of taxa from Europe and northwest
Africa
Category Features
0. Myrmecoxenous Not ant-associated
1. Weakly myrmecophilous Very few ant associations reported, stable ant association
formed only exceptionally
2. Moderately Varying proportion of larvae attended by ants, intermediate
myrmecophilous between categories 1 and 3
3. Steadily myrmecophilous Most, if not all, mature larvae ant-associated
4. Obligately myrmecophilous Larvae dependent on ants as commensals or parasites
Table 6.2 Synopsis of some key characteristics of myrmecophilous Lycaenidae as an aid to

categorising relationships (after Fiedler 1998)
Characteristic Facultative mutualism Obligate mutualism Parasitism
Specificity Low High High
Nature of association Stochastic Deterministic Deterministic
Proportion of Variable All All
ant-tended larvae
Importance for Variable, beneficial Indispensable Indispensable
survival
Costs Low High Low
Host plant range Narrow Amplified Narrow
Proportion of fauna Typically >50% Minority Minority
Interaction basis Nutrient supply, Nutrient supply, Manipulation
honest signalling manipulation
lous, at least towards the end of their larval life. Polyommatinae showed especially
high levels of facultative myrmecophily. It has been postulated widely that myr-
mecophily may be the ancestral habit for Lycaenidae (following Malicky 1969), so
that myrmecoxeny must then be a secondary condition, as discussed by Fiedler
(1991b, 1995). However, as Fiedler and Maschwitz (1989) reaffirmed, ‘the basic
pattern of lycaenid myrmecophily is indeed a mutualistic system’, with ants protect-
ing the larvae and gaining a food reward from them and, in the most specialised
cases, the lycaenids totally dependent on the ants.
Firm allocation of many lycaenids to functional category of ant association is
uncertain; the nature of intergrading features that occur (Table 6.2) was discussed
further by Fiedler (1998). Those syntheses led to designation of three classes of
parasitism, each exemplified by particular taxa, as (1) Maculinea-type with grada-
tion/change from larval herbivory to dependence on ants in later instars; (2)
Aphnaeine-type, with herbivorous larvae but with diet supplementation from ant
regurgitations, often from very specific associations; and (3) Miletinae-type, gener-
ally preying on Hemiptera and/or feeding on honeydew, with some species eliciting
regurgitation from the host ants.
Whatever categorisation is applied, the ecological and conservation implications
of ant associations with Lycaenidae are highly intertwined. For example, the warmer
microclimates on which some host ants depend may influence incidence and condi-
tion of the larval host plants needed by the butterfly. In such cases, such as for M.
arion, the butterfly depends indirectly on short swards of vegetation because only
there can the host Myrmica thrive. Distribution of trophobionts, whether Lepidoptera
or Hemiptera, can be influenced strongly by conditions needed by the attending
ants. Two rare species of aphids showed habitat specificity in that they only utilised
host plants near the ant nests – their conservation thus required not only host plants
but also suitably high density of nests of a suitable host ant (Hopkins and Thacker
1999).
Two, non-exclusive, hypotheses on the bases of costs and benefits amongst
lycaenid-ant associations have attracted considerable attention and comment. The
‘appeasement hypothesis’ is simply that ant-associated lycaenid larvae benefit

because they are not attacked by the ants; with the food rewards, they provide ‘a
kind of bribery’ (Pierce et al. 2002). The ‘protection hypothesis’ implies more
active involvement in which the attendant ants actively guard the lycaenids against
natural enemies and receive nutritious rewards for doing so.
Involvement of myrmecophily in speciation amongst the Lycaenidae is largely
speculative, but two possible mechanisms for a role in butterfly diversification have
been discussed, following Atsatt (1981) and Pierce (1984). These, both with rather
little conclusive support, are (1) host plant shifts as a consequence of ant-dependent
oviposition and (2) the patchy co-occurrence of suitable host plants and ants leading
to local fragmentation and isolation of populations. Both systems are likely to work
only in obligate mutualisms, typically only found in a small proportion of any local
lycaenid fauna. Likewise, whilst host plant selection by lycaenids may be strongly
influenced by their myrmecophily – with Pierce (1985) suggesting the contrast
between (1) myrmecophilous species preferring protein-rich host plants and (2)
myrmecoxenous species occurring on less nutritious hosts – this attractive scenario
has rarely been supported. Extensive review (Fiedler 1995) supported a higher inci-
dence of myrmecophily amongst legume feeders when compared with non-legume-
feeding species but did not support three other host plant preference traits suggested
by earlier studies. Neither parasitic plants (mistletoes and other Santalales) nor non-
leguminous nitrogen fixers supported higher proportions of myrmecophilous lycae-
nids than did normal non-leguminous hosts, and no preference by myrmecophilous
lycaenids was apparent for plants with extrafloral nectaries.
Amongst European lycaenids (103 species appraised by Stadler et al. 2003), lar-
val host plant specificity was associated with ant attendance – closely so in those
feeding on inflorescences or on Fabaceae, on which production of nitrogen-rich
(and more attractive) secretions is found (Pierce 1985). In contrast, living in forests
is associated with lower likelihood of myrmecophily. Different species of Lycaenidae
co-occurring on the same host plants may exhibit very different associations with
resident ants, as demonstrated for three closely related species found on Saraca
thaipingensis (Caesalpiniaceae), a common forest tree in Malaysia (Seufert and
Fiedler 1996a, b). All three are polyphagous across a host range of about 17 plant
species, all distributed patchily, so that host shifts are needed at times. Drupadia
theda has an obligate mutualism with Crematogaster ants; D. ravindra has a facul-
tative mutualistic relationship, with at least nine genera of ants involved; and
Cheritra freya is not attended by ants but tolerated by almost all the ant species
present. The three have thus developed markedly different associations, which have
been compared in considerable detail. D. theda uses ants as oviposition cues, and,
in experimental exclusion trials, larval mortality from predators and braconid wasp
parasitoids increased greatly when ants were excluded.
Partitioning of resources amongst the lycaenids exploiting S. thaipingensis
reflects complex patterns of their interactions with ants. Thus, on trees occupied by
aggressive weaver ants (Oecophylla smaragdina), only specialised mutualists that
can appease the ants can survive – larvae of others are killed and eaten. Seufert and
Fiedler (1996b) noted three of the ten Malaysian Lycaenidae (Hypolycaena erylus,
Table 6.3 The costs and benefits to the lycaenid Anthene emolus from its obligate association with
the weaver ant Oecophylla smaragdina (as suggested and discussed by Fiedler and Maschwitz
1989, who concluded that the benefits outweigh the costs)
Selective advantages to the butterfly
1. Larvae gain high level of protection from predators and parasitoids from vigorous defence by
Oecophylla
2. Rate of parasitisation lowered considerably as only second instar larvae are not attended
directly by ants, but this instar normally suffers only low parasitoid attack levels
3. Transport of larvae to plants that they could not otherwise exploit, possibly increasing
effectiveness of herbivory
Costs to the butterfly
1. Oviposition possible only in sites where Oecophylla lives, so that many potentially suitable
host plants cannot be utilised
2. Fatal ant attacks may take place during oviposition
3. Oviposition might also occur on plants occupied by Oecophylla but which are not suitable as
larval food
4. Survival of first instar larvae wholly dependent on the presence of Oecophylla – unless
‘adopted’ rapidly, larvae starve to death
5. Transport of larvae to unsuitable food plants
6. Adult Anthene highly vulnerable to ant predation at eclosion
Rapala pheretima, Anthene emolus) feeding on Saraca to be obligate myrmeco-

philes with Oecophylla. The first two species oviposit on ant-occupied plants with-
out any prior contact between butterfly and ant, and A. emolus oviposits with
nonaggressive contact, laying large batches of up to 150 eggs – an average number
of 59 eggs was reported by Fiedler and Maschwitz (1989). A. emolus uses Oecophylla
as an oviposition cue, and hatchling larvae are carried by the ants into the nests and
cannot survive outside (Fiedler and Maschwitz 1989).The balance between costs
and benefits is complex for the butterfly, whilst the ant can benefit consistently
(Table 6.3).
The contrast between these mutualists and the other (excluded) lycaenids ensures
that the host plants with colonies of weaver ants have a specialised and selective
association of lycaenids and cannot be exploited by other taxa. Few competitors or
natural enemies are likely to penetrate Oecophylla aggression, so that those special-
ists also gain enemy-free space and be actively defended by the ants. However, the
specialists that facilitate the mutualisms with O. smaragdina may also effectively
restrict their opportunities to exploit other habitats.
As in numerous other mutualism contexts, choosing the ‘right partner’ can be
crucial for the association to succeed and is a further scenario illuminated by studies
from lycaenids potentially linking with several ant species whose different abilities
and features may affect their roles in protecting the larvae. ‘Partner quality’ can differ
greatly between species, and some lycaenids differ in their behavioural responses to
ants in relation to features such as the number of attending ants and the level of con-
tacts between ants and larvae. Larvae can, for instance, vary the size and number of
droplets of nutrition they secrete as food for the ants. Ideally, an ant partner ‘selected’
Table 6.4 Summary of measurements on reactions of larvae of the lycaenid Glaucopsyche

lygdamus to tending ant species
Ant species Secreted droplets Tentacle eversions Attending ants
Formica obscuripes 2.0 +/− 0.4 50.8 +/− 5.9 1.64 +/− 0.07
F. podzolica 1.9 +/− 0.3 31.9 +/− 6.2 1.68 +/− 0.06
F. altipetens 1.8 +/− 0.3 40.8 +/− 6.5 1.82 +/− 0.03
F. neogagates 1.4 +/− 0.3 29.4 +/− 3.8 1.66 +/− 0.06
Myrmica incompleta 1.8 +/− 0.3 39.3 +/− 5.7 1.53 +/− 0.08
Tapinoma sessile 3.6 +/− 0.5 40.6 +/− 5.9 1.67 +/− 0.06
Sixteen larvae were exposed to two ants of each of the six species and number of secreted droplets,
tentacle eversions and mean number of attending ants recorded during 10 min interaction periods
(Axen 2000)
from a range of possible accompanyists would be one that provided effective protec-
tion to the larvae without incurring a high investment in food rewards (Axen 2000) –
with factors such as body size, colony size and aggressiveness amongst those that
may differ strongly amongst candidates. The facultatively myrmecophilous
Glaucopsyche lygdamus in North America is tended by at least ten ant species in
Colorado, where its responses to representatives of three species were compared by
Axen. The selected ants were (1) Formica obscuripes, large and readily attacking
objects that approach larvae; (2) F. altipetens, medium sized, moderately bold but
becoming defensive when disturbed; and (3) F. neogates, of similar size but less
aggressive. Other ants involved include Myrmica incompleta, timid but occasionally
tending G. lygdamus and Tapinoma sessile, described by Axen (2000) as a ‘quite
small … and timid ant species’, often the first to tend larvae but subsequently com-
monly replaced by other species. Larvae of Glaucopsyche were exposed to laboratory
colonies of ants and the number of droplets of ‘reward’ counted during an interaction,
together with number of larval tentacle eversions. The small T. sessile received nearly
twice as many droplets as any other species, although the mean number of ants
attending differed a little (Table 6.4). In tests involving more than one ant species in
sequence, changes in larval behaviour with different ants were rapid, occurring within
a few minutes.
Earlier, Pierce and Easteal (1986) established that G. lygdamus larvae on plants
from which ants had been excluded drop off those plants more readily than when
ants were present, but proportions taken by predators in the two treatments were
very similar. Overall, survival of ant-attended larvae was higher, with greatest mor-
tality in the later instars (Fig. 6.6). From the trials compared, the net selective
advantage of ants (Formica altipetens) attending the larvae was a 12-fold increase
(from 0.7 to 9.9%) in survival from egg to adult eclosion.
Ant density, tested in a separate set of trials, affected secretion rates for all three
species, but, following earlier work (Axen and Pierce 1998), the behaviour was
interpreted as regulating the number of attending ants, with need to attract more ants
of ineffective defenders in order to assure adequate protection, and undertaken
through proferring increased reward. Increased larval secretion with increased ant
Fig. 6.6 The mortality

percentage for each age
class (instar) of larvae of
Glaucopsyche lygdamus
with (solid line) or without
(dashed line) ants. For
instar three, mortality
shown is cumulative over
two instars (Pierce and
Easteal 1986)
density may be a safeguard even if those ants are poor protectors – simply because
their presence may increase the level of protection given.
Bringing together the ecological features that have been implicated in influenc-
ing ant attendance of aphids and lycaenids, Stadler et al. (2003) listed a series of
variables (Table 6.5) that have received varying levels of consideration and support
as correlates of mutualisms. The quantity and quality of honeydew or ‘nectar’ are
crucial in attracting ants. One functional contrast is that honeydew is a waste prod-
uct for aphids (p. 109) but nectar a specialised production from lycaenids – possibly
implying that the latter may be less flexible or opportunistic in fostering mutual-
isms, because the ‘cost’ is relatively high.
The concentration of interest and investigation of lycaenid-ant associations has
masked that, although rarely, some other Lepidoptera have also become
myrmecophiles. Larvae of a Malaysian tortricid moth, Semutophila saccharopa,
discharge sugar-rich solutions through the anus whilst they feed within silken shel-
ters on leaves of a giant bamboo (Gigantochloa scortechini) (Maschwitz et al.
1986). The shelters (about 1 cm in diameter) have a small opening through which
the anus can be protruded and liquid egested. This activity is stimulated by ant atten-
tion, and the liquid is imbibed by a variety of ant taxa.
The indirect effects between intertwined and co-occurring mutualist systems
have only rarely been explored. Kaminski et al. (2010), for example, noted the
absence of studies on the ecological consequences of co-occurrence of ant-tended
insects on a plant, on which survival effects on either trophobiont species are
obscure. Their example (Fig. 6.7) linked two classic mutualisms (ant-Hemiptera,
ant-Lycaenidae) and involved foliage-dwelling ants, herbivorous insects and natural
enemies on the host plant Schefflera vinosa (Araliaceae) in Brazil. There, the but-
terfly Parrhasius polibetes is facultatively attended by Camponotus ants which also
Table 6.5 Variables that have been associated in published literature with ant attendance of
myrmecophilous Hemiptera and Lycaenidae, as summarised by Stadler et al. (2003)
Variable Hypothesis
Size Larger insects produce more honeydew, likely to attract more ants
Feeding site on Aphids feeding on woody plant structures need more time to penetrate and
plants withdraw stylets and should use protective function of ants; lycaenids on
inflorescences are likely to attract more ants because of better nutrient
conditions
Colony structure Dense aggregations should be easier to sample for honeydew than solitary
individuals
Host specificity Polyphagous insects, occurring in different habitats, might come into
contact with ants; polyphagous lycaenids are often associated with
ant-dependent oviposition
Abundance Abundant species should be more likely to tolerate ant attendance and
provide more rewarding resource to ants
Mobility Mobile insects are less dependent on protection from ants
Other defence Species covered with wax or winged aphids should be less dependent on
structures protection from ants
Host alternation Host-alternating species are more likely to encounter ants and should
develop mechanisms to tolerate ant attendance
Exposure Species feeding in exposed sites are more likely to come into contact with
ants than concealed feeders
Distribution Species with wide distribution should be more likely to contact ants
Fabaceae Feeding on nitrogen-rich plants increases production of high quality nectar
and so facilitates ant attendance
Life form of host Accessibility for the ant partner might affect degree of ant attendance
Habitat Insects on plants in open/disturbed habitats have higher probability of ant
attendance, as those habitats may be more favourable for ants
Fig. 6.7 The indirect

interaction web for the host
plant Schefflera vinosa
(Araliaceae) and the
associated foliage-dwelling
ants, herbivorous insects
and natural enemies in
Brazil. Solid lines are
direct effects; broken lines
are indirect effects; plus
and minus signs indicate
positive (+) or negative (−)
signals from an initiator to
a receiver (arrowed)
species (After Kaminski
et al. 2010)
References 139
(with up to 20 other ant species also involved) tend the membracid treehopper
Guayaquila xiphias, a gregarious species which produces abundant honeydew. If
they are given a choice, the butterflies prefer to oviposit on branches with Guayaquila,
with ants tending the treehoppers and reducing natural enemy abundance, leading to
increased Parrhasius larval survival. The treehoppers thereby create enemy-free
space for the butterfly larvae on foliage, and this is exploited by the butterflies in
protecting their offspring. The ant-trophobiont association is an oviposition cue for
the butterfly. The associations are detected visually and show that behavioural adap-
tations to such facultative mutualisms may promote those interactions (Mota and
Oliveira 2016). The presence of ants alone was enough to elicit responses by the
female butterflies, but the presence of G. xiphias alone did not do so. Their impor-
tance, however, is to generate enemy-free space for the butterfly larvae by concen-
trating number of protective ants. In parallel, it had earlier been known that simple
aggregation of other lycaenid larvae may also achieve this by increasing attending
ant numbers (Axen and Pierce 1998).
References
Akino T, Knapp JJ, Thomas JA, Elmes GW (1999) Chemical mimicry and host specificity in the
butterfly Maculinea rebeli, a social parasite of Myrmica ant colonies. Proc R Soc Lond B
266:1419–1426
Als TD, Nash DR, Boomsma JJ (2002) Geographical variation in host-ant specificity of the para-
sitic butterfly Maculinea alcon in Denmark. Ecol Entomol 27:403–414
Als TD, Vila R, Kandul NP, Nash DR, Yen SH, Hsu YF (and 3 other authors) (2004) The evolution
of alternative parasitic life histories in large blue butterflies. Nature 432:386–390
Atsatt PR (1981) Lycaenid butterflies and ants: selection for enemy-free space. Am Nat
118:638–654
Axen AH (2000) Variation in behavior of lycaenid larvae when attended by different ant species.
Evol Ecol 14:611–625
Axen AH, Pierce NE (1998) Aggregation as a cost-reducing strategy for lycaenid larvae. Behav
Ecol 9:109–115
Bachtold A, Alves-Silva E, Kaminski LA, Del-Claro K (2014) The role of tending ants in host
plant selection and egg parasitism of two facultative myrmecophilous butterflies.
Naturwissenschaften 101:913–919
Bachtold A, Silva EA, Del-Claro K (2016) Ants, plant characteristics and habitat conservation
status affect the occurrence of myrmecophilous butterflies on an extrafloral nectaried
Malpighiaceae. Stud Neotropical Fauna Environ. doi:10.1080/01650521.2016.1198192
Barbero E, Patricelli D, Witek M, Belletto E, Casacci LP, Sala M, Bonelli S (2013) Myrmica ants
and their butterfly parasites with special focus on the acoustic communication. Psyche. Article
ID 725237. 11 p. doi:10.1155/2012/725237
Brakefield PM, Shreeve TG, Thomas JA (1992) Avoidance, concealment, and defence. In: Dennis
RLH (ed) The ecology of butterflies in Britain. Oxford University Press, Oxford, pp 93–119
Bronstein JL (1998) The contribution of ant-plant protection studies to our understanding of mutu-
alism. Biotropica 30:150–161
Cottrell CB (1984) Aphytophagy in butterflies: its relationship to myrmecophily. Zool J Linnean
Soc 79:1–57
Cushman JH, Rashbrook VK, Beattie AJ (1994) Assessing benefits to both participants in a
lycaenid-ant association. Ecology 75:1031–1041
De Vries PJ (1988) The larval ant-organs of Thisbe irenea (Lepidoptera: Riodinidae) and their
effects upon attending ants. Zool J Linnean Soc 94:379–393
De Vries PJ (1991) Evolutionary and ecological patterns in myrmecophilous riodinid butterflies.
In: Huxley CR, Cutler DF (eds) Ant-plant interactions. Oxford Unversity Press, Oxford,
pp 143–156
De Vries PJ, Cockroft RB, Thomas J (1993) Comparison of acoustical signals in Maculinea but-
terfly caterpillars and their obligate host Myrmica ants. Biol J Linn Soc 49:229–238
Downey JC (1962) Host-plant relations as data for butterfly classification. Syst Zool 11:150–159
Elmes GW, Thomas JA, Hammarstedt O, Munguira ML, Martin J, Van der Made JG (1994)
Differences in host-ant specificity between Spanish, Dutch and Swedish populations of the
endangered butterfly, Maculinea alcon (Denis et Schiff.) (Lepidoptera). Memor Zool
48:55–68
Fiedler K (1991a) European and North West African Lycaenidae (Lepidoptera) and their associa-
tions with ants. J Res Lepid 28:239–257
Fiedler K (1991b) Systematic, evolutionary and ecological implications of myrmecophily within
the Lycaenidae (Insecta: Lepidoptera: Papilionoidea). Bonn Zool Monogr 31:1–210
Fiedler K (1995) Lycaenid butterflies and plants: is myrmecophily associated with particular host-
plant preferences? Ethol Ecol Evol 7:107–132
Fiedler K (1998) Lycaenid-ant interactions of the Maculinea type: tracing their historical roots in
a comparative framework. J Insect Conserv 2:3–14
Fiedler K (2012) The host genera of ant-parasitic Lycaenidae butterflies: a review. Psyche 2012.
Article ID 153975. doi:10.1155/2012/153975
Fiedler K, Maschwitz U (1989) The symbiosis between the weaver ant, Oecophylla smaragdina,
and Anthene emolus, an obligate myrmecophilous lycaenid butterfly. J Nat Hist 23:833–846
Fiedler K, Saam C (1995) Ants benefit from attending facultatively myrmecophilous Lycaenidae
caterpillars: evidence from a survival study. Oecologia 104:316–322
Fiedler K, Holldobler B, Seufert P (1996) Butterflies and ants: the communicative domain.
Experientia 52:14–24
Forister ML, Gompert Z, Nice CC, Forister GW, Fordyce JA (2011) Ant association facilitates the
evolution of diet breadth in a lycaenid butterfly. Proc R Soc Lond B 278:1539–1547
Hayes MP (2015) The biology and ecology of the large blue butterfly Phengaris (Maculinea)
arion: a review. J Insect Conserv 19:1037–1051
Hinton HE (1951) Myrmecophilous Lycaenidae and other Lepidoptera – a summary. Proc Trans S
Lond Entomol Nat Hist Soc 1949–1950:111–175
Hojo MK, Wada-Katsumata A, Akino T, Yamaguchi S, Ozaki M, Yamaoka T (2009) Chemical
disguise as particular caste of host ants in the ant inquiline parasite Niphanda fusca (Lepidoptera:
Lycaenidae). Proc R Soc Lond B 276:551–558
Hojo MS, Yamamoto A, Akino T, Tsuji K, Yamaoka R (2014) Ants use partner specific odors to
learn to recognize a mutualistic partner. PLoS One 9(1):e86054. doi:10.1371/journal.
pone.0086054
Hojo MK, Pierce NE, Tsuki K (2015) Lycaenid caterpillar secretions manipulate attendant ant
behavior. Curr Biol 25:2260–2264
Hopkins GW, Thacker JI (1999) Ants and habitat specificity in aphids. J Insect Conserv 3:25–31
Inui Y, Shimizu-kaya U, Okubo T, Yamsaki E, Itioka T (2015) Various chemical strategies to
deceive ants in three Arhopala species (Lepidoptera: Lycaenidae) exploiting Macaranga myr-
mecophytes. PLoS One 10(4):e0120652. doi:10.1371/journal.pone.0120652
Jordano D, Thomas CD (1992) Specificity of an ant-lycaenid interaction. Oecologia 91:431–438
Kaminski LA, Freitas AVL, Oliveira PS (2010) Interaction between mutualisms: ant-tended but-
terflies exploit enemy-free space provided by ant-treehopper associations. Am Nat
176:322–334
Kitching RL, Luke B (1985) The myrmecophilous organs of the larvae of some British Lycaenidae
(Lepidoptera): a comparative study. J Nat Hist 19:259–276
Leimar O, Axen AH (1993) Strategic behaviour in an interspecific mutualism: interactions between
lycaenid larvae and ants. Anim Behav 46:1177–1182
References 141
Lohman DJ, Samarita VU (2009) The biology of carnivorous butterfly larvae (Lepidoptera:
Lycaenidae: Miletinae: Miletini) and their ant-tended hemipteran prey in Thailand and the
Philippines. J Nat Hist 43:569–581
Malicky H (1969) Versucheiner Analyse der okologischen Beziehungen zwischen Lycaenidae
(Lepidoptera) und Formiciden (Hymenoptera). Tidschr Entomol 112:213–298
Maschwitz U, Dumpert K, Tuck KR (1986) Ants feeding on anal exudate from tortricid larvae: a
new type of trophobiosis. J Nat Hist 20:1041–1050
Mota LL, Oliveira PS (2016) Myrmecophilous butterflies utilise ant-treehopper associations as
visual cues for oviposition. Ecol Entomol 41:338–343
Nash DR, Als TD, Maile R, Jones GR, Boomsma JJ (2008) A mosaic of chemical coevolution in a
large blue butterfly. Science 319:88–90
New TR (ed) (1993) Conservation biology of Lycaenidae (butterflies), Species Survival
Commission report, No. 8. IUCN, Gland
Pascale EG, Thiet RK (2016) The relationship between ants and Lycaeides melissa samuelis
(Lepidoptera: Lycaenidae) at Concord Pine Barrens, NH, USA. Environ Entomol 45:633–641
Patricelli D, Barbero S, Occipinti A, Bertea CM, Bonelli S (and seven other authors) (2015) Plant
defences against ants provide a pathway to social parasitism in butterflies. Proc R Soc B 282:
20151111. doi.org/10.1098/rspb.2015.1111
Pellissier L, Rasmanns S, Litsiong G, Fiedler K, Dubuis A, Pottier J, Guisan A (2012) High host-
plant nitrogen content: a prerequisite for the evolution of ant-caterpillar mutualisms? J Evol
Biol. doi:10.1111/j.1420-9101.2012.02555x
Pierce NE (1984) Amplified species diversity: a case study of an Australian lycaenid butterfly and
its attendant ants. In: Vane-Wright RI, Ackery PR (eds) The biology of butterflies. Academic,
London, pp 197–200
Pierce NE (1985) Lycaenid butterflies and ants: selection for nitrogen fixing and other protein rich
food plants. Am Nat 125:888–895
Pierce NE (1995) Predatory and parasitic Lepidoptera: carnivores living on plants. J Lepid Soc
49:412–453
Pierce NE, Easteal S (1986) The selective advantage of attendant ants for the larvae of a lycaenid
butterfly, Glaucopsyche lygdamus. J Anim Ecol 55:451–462
Pierce NE, Nash DR (1999) The imperial blue: Jalmenus evagoras (Lycaenidae). In: Kitching RL,
Scheermeyer E, Jones RE, Pierce NE (eds) The biology of Australian butterflies. CSIRO
Publications, Melbourne, pp 279–315
Pierce NE, Kitching RL, Buckley TC, Taylor MFJ, Benbow KF (1987) The costs and benefits of
cooperation between the Australian lycaenid butterfly, Jalmenus evagoras, and its attendant
ants. Behav Ecol Sociobiol 21:237–248
Pierce NE, Nash DR, Baylis M, Carper ER (1991) Variation in the attractiveness of lycaenid but-
terfly larvae to ants. In: Huxley CR, Cutler DF (eds) Ant-plant interactions. Oxford Unversity
Pierce NE, Braby MF, Heath A, Lohman DJ, Mathew J, Rand DB, Travassos MA (2002) The ecol-
ogy and evolution of ant association in the Lycaenidae (Lepidoptera). Annu Rev Entomol
47:733–771
Porter K, Steel CA, Thomas JA (1992) Butterflies and communities. In: Dennis RLH (ed) The
ecology of butterflies in Britain. Oxford University Press, Oxford, pp 139–177
Schonrogge K, Wardlaw JC, Thomas JA, Elmes GW (2000) Polymorphic growth rates in myrme-
cophilous insects. Proc R Soc Lond B 267:771–777
Seufert P, Fiedler K (1996a) The influence of ants on patterns of colonization and establishment
within a set of coexisting lycaenid butterflies in a south-east Asian tropical rain forest.
Seufert P, Fiedler K (1996b) Life-history diversity and local co-existence of three closely related
lycaenid butterflies (Lepidoptera: Lycaenidae) in Malaysian rainforests. Zool Anz
234:229–239
Shimizu-kaya U, Okubo T, Yago M, Inui Y, Itioka T (2013) Myrmecoxeny in Arhopala zylda
(Lepidoptera, Lycaenidae) larvae feeding on Macaranga myrmecophytes. Entomol News
123:63–70
Stadler B, Kindlmann P, Smilauer P, Fiedler K (2003) A comparative analysis of morphological

and ecological characters of European aphids and lycaenids in relation to ant attendance.
Tartally A, Nash DR, Lengyel S, Varga Z (2008) Patterns of host ant use by sympatric populations
of Maculinea alcon and M. ‘rebeli’ in the Carpathian Basin. Insect Soc 55:370–381
Thomas JA, Elmes GW (2001) Food-plant niche selection rather than the presence of ant nests
explains oviposition patterns in the myrmecophilous butterfly genus Maculinea. Proc R Soc
Lond B 268:471–477
Thomas JA, Elmes GW, Wardlaw JC, Woyciechowski M (1989) Host specificity among Maculinea
butterflies in Myrmica ant nests. Oecologia 79:452–457
Thomas JA, Simcox DJ, Wardlaw JC, Elmes GW, Hochberg ME, Clarke RT (1998) Effects of lati-
tude, altitude and climate on the habitat and conservation of an endangered butterfly Maculinea
arion and its Myrmica ant hosts. J Insect Conserv 2:39–46
Trager MD, Daniels JC (2009) Ant tending of Miami blue butterfly larvae (Lepidoptera:
Lycaenidae): partner diversity and effects on larval performance. Fla Entomol 92:474–482
Travassos MA, Pierce NE (2000) Acoustics, context and function of vibrational signaling in a
lycaenid butterfly-ant mutualism. Anim Behav 60:13–26
van Dyck H, Oostermeijer H, Talloen JGB, Feenstra V, Van der Hidde A, Wynhoff I (2000) Does
the presence of ants matter for oviposition to a specialized myrmecophilous Maculinea but-
terfly? Proc R Soc Lond B 267:861–866
Weeks JA (2003) Parasitism and ant protection alter the survival of the lycaenid Hemiargus isola.
Chapter 7
Other Insect Mutualisms
Keywords Beetles • Fungi • Ithomiinae • Macrotermitinae • Mimicry rings •

Mullerian mimicry • Scolytidae • Termites • Termitomyces
7.1 Introduction: The Remaining Variety
This short chapter emphasises the wider diversity of less prominent (or, in general,
less heralded in conservation) mutualisms that involve insects, and exemplifies
these by several rather different but widespread examples. Two insect groups dis-
cussed form complex associations with fungi, and the final theme (Mullerian mim-
icry) illustrates the functionally different structure of a mutualism whereby
convergence between different species facilitates their mutual protection from
predators.
7.2 Termites and Fungi
Parallels with ant-fungus associations are displayed by some other insects, notably
by macrotermitine termites, representing the other major group of social insects.
Together with attine ants, and reflecting their conspicuous nest structures, mac-
rotermitines have been described as ‘among the most impressive animal phenomena
in the world’ (Aanen et al. 2002). However, whilst the attine ants evolved primarily
as ‘farmers’ of relatively generalised fungi, these termites apparently specialised
early on a single group of fungi which has become isolated from any free-living
sister group and co-speciated with the termites as the variety of fungus comb sub-
strates increased. Macrotermitinae comprise about 11 genera and about 30 species,
accompanied by about 40 species of Termitomyces fungus (Aanen et al. 2002).
These fungus-growing termites are among the predominant decomposers in the Old
World tropics, with their greatest diversity in Africa. Macrotermitinae are the only
termites with this habit, and all the genera have mutualistic relationships with
Termitomyces fungi (Tricholomataceae) which are grown on a specialised ‘fungus
comb’ in the nests and continually supplied with predigested plant material as a

DOI 10.1007/978-3-319-58292-4_7
144 7 Other Insect Mutualisms
suitable substrate. Many of the fungus species occur with several termite species,
and horizontal transmission may be frequent both within and amongst extant taxa.
Phylogenetic analyses have shown the association to have a single African origin
(Aanen et al. 2002). The natural history of these associations, summarised suc-
cinctly by Wood and Thomas (1989), emphasises the long history of curiosity and
interest as mutualisms – as for leaf-cutting ants – and speculations on the relation-
ships within both.
The mutualism is based on enabling the termites to exploit the major resource of
dead vegetation, with the fungus degrading the lignin and cellulose components to
render the old (5–8 weeks) fungal comb material far more nutritious than the parent
vegetation around the termite nests, and favouring termite growth and reproduction.
The fungus is presumed to be the main decomposer of lignocelluloses, but may also
have additional roles that benefit the termites. For example, Bignell (2011) noted it
as an additional source of protein- and nitrogen-rich foods and/or an additional
digestive aid in providing enzymes that complement those of the termites them-
selves. Variations in function are likely with different termite taxa but, in general,
moderate levels of specificity occur between fungi and termites (Nobre et al. 2011),
and continual exchange of fungus appears to have taken place across host taxa and
regions. The fungus gains a suitable and competition-free environment. Continual
‘grooming’ by the termites suppresses germination and growth of other fungi on the
combs, and spread/transmission of the fungus occurs through combs in existing or
new termite colonies being inoculated. Winged reproductive adult termites (of
Microtermes, Macrotermes and perhaps other genera) ingest asexual fungal spores
before they disperse to found new colonies, and these are deposited on faeces of
young foraging workers at some later stage. Transmission of fungal spores by alate
termites is relatively rare in comparison with the more widespread gleaning of
Termitomyces spores from the wider environment by the first foraging workers of a
new colony and these being used to inoculate and establish the new fungus garden
(Korb and Aanen 2003). In many taxa, fungal combs are established only after for-
aging workers have carried or introduced fungal spores.
The obligate nature of the termite-fungus mutualism has been long-known
(Sands 1969), with termites unable to survive without the fungus comb present. The
dynamics of a fungus-termite association were exemplifed for the south-east Asian
Odontotermes formosanus, in which the faecal material is deposited as a fungus
comb (or fungus garden) that forms the substrate for the symbiotic Termitomyces
(Mathew et al. 2011). In addition, Bacillus microorganisms serve as mutualists in
this system, and can suppress growth of some fungi that exploit the fungus comb,
and also facilitate termite access to the lignocellulose components of the combs. In
this case, both Termitomyces and Bacillus have mutualistic associations with the
fungus comb containing the termites.
7.3 Beetles and Fungi 145
7.3 Beetles and Fungi
Many different groups of Coleoptera are associated with fungi, many of them sim-
ply as consumers, but the interactions between some tree-associated beetles and
fungi are intricate, and mutualistic. Most information on the intricate relationships
involving bark beetles and ambrosia beetles (Scolytidae, Platypodidae) with fungi
has come from studies on beetle species that are economically important, notably as
forestry pests. Thus, about 1400 species of ambrosia beetles (Xyleborini) have an
obligate mutualism with xylem-inhabiting specific ambrosia fungi, which they cul-
tivate and feed on. Fungiculture by these beetles has evolved on at least 11 occa-
sions, and the resulting symbioses are complex and, in many cases, difficult to
interpret (Freeman et al. 2016). Associations between the beetles and fungi are very
varied (Beaver 1989), but the beetles are generally associated with fungi in obligate
relationships, which have evolved independently on several occasions. Some stud-
ies of these associations have revealed little impact. The beetles occupy woody
plants and characteristically form branching ‘galleries’ in phloem (bark beetles) or
wood (ambrosia beetles). The fungi are carried by beetles in specialised cuticular
structures, mycangia, and physically inoculated into new sites. Collectively, these
beetles exhibit a wide range of feeding habits, but ambrosia beetles in particular use
fungi as a predominant larval and adult food which is essential to complete their life
cycle. Fungi, including the blue wood-staining taxa such as Ceratocystis spp., of
concern in commercial softwood quality, occur mutualistically with bark beetles,
with a variety of different fungus taxa associated with different beetle tribes, grow-
ing within the beetle galleries and being spread through beetle-tunnelling activities.
The mutualisms appear to be based on (1) the fungus growing in highly suitable
tended environments, with reliable dispersal by beetle vectors carrying them to new
suitable habitats, and (2) the fungus being the only, or predominant, source of food
for beetle larvae and young adults. Beetles may also benefit from the inoculated
fungus suppressing host tree defences around beetle attack sites on some conifers,
so increasing the chances of successful colonisation by genera such as Dendroctonus
and Scolytus, both genera that include major pest species. Indeed, Berryman (1989)
noted that some scolytids use fungi as ‘weapons to kill trees’. Most fungi exploited
by ambrosia beetles are polyphagous and breed in a variety of tree species, in con-
trast to the more specific nature of many bark beetle fungal associates.
In all these associations the host tree, whether living or dead, is an integral com-
ponent – if dead it plays no active role but may become a site for generating beetles
over substantial periods (of up to several decades) but if alive its defences to fungi,
beetles, or both render the association antagonistic. The overall impacts of the ophi-
ostomatoid fungi on their beetle vectors may thus vary with the condition of the host
trees, as proposed by Raffa (1995) in suggesting three possible (and non-exclusive)
mechanisms whereby the fungi could influence beetle populations: (1) fungi might
reduce host tree resistance to beetle attack; (2) some fungi may compete with beetle
larvae for host nutrients, or interfere with their development in some other way; and
(3) some fungi might compete with other fungi that are mutualists or antagonistic to
beetle larvae. Examples of each were cited by Eckhardt et al. (2004).
7.4 Mullerian Mimicry
The basis of Mullerian mimicry, one of the early classically defined forms of mim-
icry widespread amongst insects, is that a suite of co-occurring species unpalatable
to birds or other insectivorous predators each gain greater protection from predation
by their physical resemblance through convergence in colour pattern, form and
activity, and confuse predators into some collective avoidance following an initial
individual learning experience of distastefulness. The species can thereby all benefit
from this association, brought about through co-occurrence and apparency but with-
out any consistent functional interplay, and with the array of participating species
differing and varying in richness and composition in different places and habitats.
That mutual benefit accords those suites of species the status of mutualisms. The
species share the ‘cost’ (individual mortality to individual predators) of educating
predators, and may then benefit, with species mortality reduced in relation to their
relative frequency within the assemblage of mutual mimics.
Butterflies are by far the most thoroughly studied insect examples, especially in
the tropics where both diversity and predation pressure can be high. Patterns of the
brightly coloured wings of the species involved can converge in appearance within
local assemblages to comprise numerous local mimicry complexes. The most inten-
sively investigated complexes of Mullerian mimics involve Nymphalidae, amongst
which distinct complexes may co-occur in rainforest, sometimes separated by verti-
cal stratification (Beccaloni 1997a, b, Joron and Mallet 1998). Studies on Neotropical
Ithomiinae and Heliconius, in particular, demonstrate correlations among habitats,
behaviour and mutualisms (Mallet and Gilbert 1995). For Ithomiinae, Elias et al.
(2008) explored the relative roles of mutualistic interactions and competition in an
assemblage of 58 species in Ecuador, and showed that the community structure is
determined more by key mutualisms than by competition or phylogenetic links.
Co-occurring butterfly species can form relatively stable complexes within particu-
lar habitats in which their convergence to maximise ‘warning signals’ to predators
overlap and counteract competition within the complex.
For Ithomiinae in Ecuador, Beccaloni (1997b) noted that members of a mimicry
complex (often termed a ‘mimicry ring’, and defined as ‘a group of sympatric spe-
cies sharing a common warning pattern’: Joron and Mallet 1998) utilised larval food
plants growing at similar heights, in turn relating to height of adult flight, so that
assemblages are influenced by larval habits. The flight heights of other diurnal
Lepidoptera mimicking ithomiines are similar to those of their models, and analysis
of host plant records supported the reality of this mode of vertical separation. The
mimicry, therefore, may have evolved between species occupying the same vertical
interval, and reflect microhabitat-specific predation. Other studies on Ithomiinae
have found that mimicry affects species’ distributions across host plants and micro-
habitats, so influencing structure within communities.
Explored further by Chazot et al. (2014), the Andean ithomiine assemblages at
different elevations have similar phylogenetic compositions, and those communities
are shaped also by mutualistic interactions. This was suggested to act through two
References 147
possible paths, based on a study involving data from 155 species across 15 sites in
Ecuador and with evidence for mutualism taken as species sharing the same wing
patterns coexisting at larger scales more often than implied by chance. Those paths
were (1) different mimetic patterns dominating at different elevations, with co-
mimetic species occurring at similar elevations so that any colonisation of a differ-
ent elevational range should drive mimicry shifts – but with the corollary that
mimicry may have been an implicit part of historically shaping the ithomiine com-
munities; and (2) the sharing of particular elevations by co-mimetic species has
strong associations, leading to hypotheses that mutualistic interactions reinforce the
coexistence of those species along the elevational gradient (Chazot et al. 2014), pos-
sibly by initial convergence in colour pattern amongst the coexisting species fol-
lowed by convergence in the elevational range occupied. Positive outcomes for
mutualism can thereby help to structure the ecological communities of participating
species, as inferred by Elias et al. (2008).
Amongst species of Heliconius, massive diversification of colour patterns and
polymorphism produces a substantial variety of mimicry rings at any given locality
(Mallet and Gilbert 1995), and any separation of these by habitat or height would
increase the numbers of rings within an area. That study, based on associations
amongst nine species of Heliconius in Costa Rica, all of them feeding on Passiflora
shoots as larvae, revealed little difference in flight heights of mimicry rings, and all
participating species flew at levels from the ground to the canopy.
One of the most intensively studied Mullerian mimic pairs, Heliconius erato and
H. melpomene, both widely distributed in the neotropics, was for long believed to
represent a diverse coevolutionary mutualism. Recent mitochondrial DNA studies
revealed that the taxa did not share a biogeographical history and that at least some
of their strong resemblances are not derived from direct coevolution. Many
Mullerian mimic butterflies are monomorphic but, at the other extreme, of extensive
polymorphism, the distasteful Amazonian Heliconius numata has about 38 distinct
morphs across its range, and up to 11 of these in any area, each a mimic of a separate
species of Melinea (Ithomiinae) (Joron and Mallet 1998). This diversity of form was
attributed as a ‘bet-hedging’ strategy through which survival in an unpredictable
mimetic environment may be assured.
References
Aanen DK, Eggleton P, Rouland-Lefevre C, Guldberg-Froslev T, Rosendahl S, Boomsma JJ

(2002) The evolution of fungus-growing termites and their mutualistic fungal symbionts. Proc
Natl Acad Sci 99:14887–14892
Beaver RA (1989) Insect-fungus relationships in the bark and ambrosia beetles. In: Wilding N,
Collins NM, Hammond PM, Webber JF (eds) Insect-fungus interactions. Academic, London,
pp 121–143
Beccaloni GW (1997a) Ecology, behavior and natural history of ithomiinae butterflies (Lepidoptera:
Nymphalidae: Ithomiinae) and their mimics in Ecuador. Trop Lepid 8:103–124
Beccaloni GW (1997b) Vertical stratification of ithomiine butterfly (Nymphalidae: Ithomiinae)

mimicry complexes: the relationship between adult flight height and larval host-plant height.
Biol J Linn Soc 62:313–341
Berryman AA (1989) Adaptive pathways in scolytid-fungus associations. In: Wilding N, Collins
NM, Hammond PM, Webber JF (eds) Insect-fungus interactions. Academic, London,
pp 145–159
Bignell DE (2011) Morphology, physiology, biochemistry and functional design of the termite gut.
In: Bignell DE, Roisin Y, Lo N (eds) Biology of termites: a modern synthesis. Springer, Berlin,
pp 375–412
Chazot N, Willmott KR, Endara PGS, Toporov A, Hill RJ, Jiggins CD, Elias M (2014) Mutualistic
mimicry and filtering by altitude shape the structure of Andean butterfly communities. Am Nat
183:26–39
Eckhardt LG, Goyer RA, Klepzig KD, Jones JP (2004) Interactions of Hylastes species (Coleoptera:
Scolytidae) with Leptographium species associated with Loblolly pine decline. J Econ Entomol
97:468–474
Elias M, Gompert Z, Jiggins C, Willmott K (2008) Mutualistic interactions drive ecological niche
convergence in a diverse butterfly community. PLoS Biol 6(2):2642–2649. e300
Freeman S, Sharon M, Dori-Bachasg M, Maymoin M, Belausov E (and four other authors) (2016)
Symbiotic association of three fungal species throughout the life cycle of the ambrosia beetle
Euwallacea nr. fornicatus. Symbiosis 68: 115–128
Joron M, Mallet JLB (1998) Diversity in mimicry: paradox or paradigm? Trends Ecol Evol
13:461–466
Korb J, Aanen DK (2003) The evolution of uniparental transmission of fungal symbionts in
fungus-growing termites (Macrotermitinae). Behav Ecol Sociobiol 53:65–71
Mallet J, Gilbert LE (1995) Why are there so many mimicry rings? Correlations between habitat,
behaviour and mimicry in Heliconius butterflies. Biol J Linn Soc 55:159–180
Mathew GM, Ju Y-M, Lai C-Y, Mathew DC, Huang CC (2011) Microbial community analysis in
the termite gut and fungus comb of Odontotermes formosanus: the implication of Bacillus as
mutualists. FEMS Microbiol Ecol 79:504–517
Nobre T, Kone NA, Konate S, Linsenmaier KE, Aanen DK (2011) Dating the fungus-growing
termites’ mutualism shows a mixture between ancient codiversification and recent symbiotic
dispersal across divergent hosts. Mol Ecol 20:2619–2627
Raffa KF (1995) Bark beetles, fungi, trees and humans four perspectives, four agendas. In:
Christiansen E (ed) Bark beetles, blue-stain fungi, and conifer defence systems, no 6-95,
Norwegian Forest Research Institute AS, Norway, pp 7–9
Sands WA (1969) The association of termites and fungi. In: Krishna K, Weesner FM (eds) Biology
of termites, vol 1. Academic, New York, pp 495–524
Wood TG, Thomas RJ (1989) The mutualistic association between Macrotermitinae and
Termitomyces. In: Wilding N, Collins NM, Hammond PM, Webber JF (eds) Insect-fungus
interactions. Academic, London, pp 69–92
Part III
Environmental Impacts and Insect
Conservation: Lessons from Mutualisms
The array of mutualisms and allied interactions and associations discussed in the
previous part clearly present abundant opportunities for modification, loss and
renewal as their ecological contexts and environments change. Those changes,
whether they are natural or anthropogenic, can lead to serious conservation con-
cerns. All aspects of resource supply and other habitat changes – loss of vegetation,
invasive alien species impacts, climate and other physical changes, and others – that
are all priority concerns in wider insect conservation apply to the species participat-
ing in mutualisms both directly and in influencing the circumstances under which
they meet, interact and can persist together. That ‘species interactions are the funda-
mental components of ecological communities’ (de Freitas and Rossi 2015) is a
maxim emphasised repeatedly in such discussions. Changes to intricate interactions
have implications for many aspects of ecosystem functions and biodiversity conser-
vation, amongst which changes in diversity and community stability may both
occur.
The major threats to mutualisms and the numerous other interactions in which
insects participate – and which are the collective major drivers of concern in insect
conservation throughout the world – are habitat fragmentation, changes in land use
with transformations from ‘more natural’ to ‘more anthropogenic’ landscapes and
environments, invasive species impacts, and climate change. These overlap in many
ways, and ‘compartmentalising’ them for discussion of conservation of interactions
involves considerable thought to avoid confusion. Each has potential to disrupt eco-
logical associations, and many indeed do so, at scales from local to regional and
from individual site to landscape, and often reflecting local community structures
and the collective pressures on these. The three following chapters deal, in sequence,
with effects of alien invasive species, the direct intrusions of alien or native species
into established mutualisms, and the varied impacts of habitat loss and change.
Impacts of invasive alien species (Chap. 8) are a universal conservation concern.
Direct intrusions of alien or additional native species into mutualisms, some as
‘cheaters’ with evolutionary interests in sustaining those mutualisms as a resource
rather than destructive exploitation, have added to misunderstanding of how these
150 III Environmental Impacts and Insect Conservation: Lessons from Mutualisms
interactions function, and helped to clarify needs for their conservation (Chap. 9).
‘Habitat’, in the sense of an available suite of critical resources (Dennis et al. 2003,
2007), may be changed by direct alterations and the more insidious and gradual
impacts of climate change (Chap. 10). Whilst all these key themes are familiar as
threats and targets for remediation in wider insect conservation, the additional
emphasis on ‘interactions’ and their central importance in biodiversity conservation
provides a somewhat wider perspective for conservation of mutualisms. The status
of some parasites of well-established mutualisms can also furnish complex and
intricate scenarios for interpretation. Thus, the mutualisms between Piper myr-
mecophytes (p. 84) and the ant Pheidole bicornis are based on the widespread prin-
ciple of exchanging food and shelter for defence against herbivores and encroaching
vines – but Piper is highly unusual in that the plants produce food bodies only when
the ant is present, so production is presumed to be induced directly by the ants
(Risch and Rickson 1981). On ant-free plants, larvae of a clerid beetle (Phyllobaenus
sp.), representing a genus whose members are mostly predators of ants, can also
enable production of food bodies within the petiole chambers, as if ants were indeed
present. From Letourneau’s (1990) observations, the beetle larvae can exploit these
nest sites and food normally produced for mutualistic ants. The beetles may – should
they become available – also prey on ant brood, collectively depriving the plant of
both resources and protection by ants.
Historically, considerations of conserving interactions have lagged far behind the
prominence given to individual species conservation and, as Tylianakis et al. (2010)
remarked, ‘perhaps for simplicity’ have largely ignored the structure of interaction
networks, however essential those broader perspectives may be. Interactions,
including mutualisms, clearly necessitate conserving the participant species, and
many environmental changes have potential to decouple such partners by differen-
tially affecting them. Nevertheless, again from Tylianakis et al. ‘networks are more
than the sum of their component parts’. Many mutualisms are in principle simpler
than more diffuse networks with numerous participants, but their conservation must
still heed the ways in which each party (species) is affected by factors outside the
mutualism itself. Many specific pollinator-plant mutualisms comprise a core mod-
ule of the key players, with the association attended by a variety of other species that
vary considerably in their incidence, impacts and roles but collectively contribute to
stabilising or disrupting the core mutualism. The various interaction parameters dis-
cussed by Tylianakis et al. (2010) may each affect those central interactions but,
whatever approach to conservation is pursued, defining the interaction and the wider
factors that affect the major partners needs careful consideration. In short, practical
conservation of network structures, including mutualisms but also more widely, has
far to go before it becomes a routine or core activity equivalent to the more common
‘single species approach’. The latter, however, may necessarily be deficient without
that broader perspective.
However, much of the conservation attention and practice devolves on a few key
insect groups, highlighted in previous chapters as the major players in mutualisms
and across which the impetus for conservation varies considerably. They span a
range of public appeal and have garnered rather different levels of attention and
III Environmental Impacts and Insect Conservation: Lessons from Mutualisms 151
scientific understanding. Thus, butterflies have attracted more interest in their con-
servation than any other insect group, and that support has been a key influence in
leading to a wider appreciation of insect conservation and of the processes by which
this can be accomplished. Butterflies are a potent ‘flagship group’ in insect conser-
vation, for which a legacy of long hobbyist interest and documentation has led to
declines or losses of individual species, especially in parts of the northern temperate
zones, have been noticed and the threats to them well categorised as bases for sound
management. Other than for a few diurnal or spectacular moths, this majority of the
Lepidoptera have received far less attention.
Among the butterflies, Lycaenidae – including numerous myrmecophilous taxa –
dominate many regional lists of ‘Threatened butterflies’. Some conservation cam-
paigns (such as for Maculinea spp., p. 120) depend heavily on details of the
mutualisms with ants as functional augmentation of the more normal butterfly inter-
action of ‘insect-host plant’ alone. The need to consider host ants as conservation
targets flows largely from this context, and the numerous other mutualisms involv-
ing ants have received comparatively little direct conservation focus. Despite wide
appreciation of their varied and important ecological roles and their driving influ-
ences in many natural communities, public awareness of ants is dominated by the
relatively few widely dispersed ‘tramp species’ as amongst the most damaging of all
insect invaders. Reputations of ants as pests by far exceed any appreciation of their
positive or mutualistic ecological roles.
Conservation interests in Hymenoptera devolve largely on bees, and are driven
largely by concerns for pollination, both for crops and ornamental plants, but also
by realisation that many pollinator-plant associations are indeed highly specific and
groups such as orchids may suffer co-declines or co-extinctions should their specific
pollinators be lost.
In parallel with public awareness of ants, the roles of many Hemiptera as plant
pests, some of them demanding intensive management to safeguard crop plants and
becoming the foci of long-running biological control programmes, by far outweigh
interests in conservation. To many people ‘bugs are pests’, and involvement of
aphids, scale insects, mealybugs and their relatives in intricate or specific mutual-
isms on which other taxa depend has not widely increased their conservation
profile.
Appreciating the involvements of many insects in ecological interactions, includ-
ing the interdependencies implicit in mutualisms, and that these are vulnerable to
imposed changes – with those changes having cascade effects throughout wider
communities – is a key theme in promoting sound insect conservation.
References
de Freitas JD, Rossi MN (2015) Interaction between trophobiont insects and ants: the effect of
mutualism on the associated arthropod community. J Insect Conserv 19:627–638
Dennis RLH, Shreeve TG, Van Dyck H (2003) Towards a functional resource-based concept of
habitat: a butterfly biology viewpoint. Oikos 102:417–426
152 III Environmental Impacts and Insect Conservation: Lessons from Mutualisms
Dennis RLH, Shreeve TG, Sheppard DA (2007) Species conservation and landscape management:
a habitat perspective. In: Stewart AJA, New TR, Lewis OT (eds) Insect conservation biology.
CAB International, Wallingford, pp 92–126
Letourneau DK (1990) Code of ant-plant broken by parasite. Science 248:215–217
Risch SJ, Rickson FR (1981) Mutualism in which ants must be present before plants produce food
bodies. Nature 291:149–150
Tylianakis JM, Laliberte E, Nielsen A, Bascompte J (2010) Conservation of species interaction
networks. Biol Conserv 143:2270–2279
Chapter 8
Impacts of Alien Invasive Species
Keywords Dispersal • Figs • Islands • Mutualism • Novel ecological associations •

Pollinators • Tramp ants
8.1 Introduction
Many of the mutualisms discussed earlier are geographically localised and reflect
some form of coadaptation between partners within local environments. Those eco-
logical specialists or associates are moulded by the local ecological contexts, and
their persistence depends on continuity of the contexts in which those partnerships
have been developed. Plants and insects are amongst the predominant groups of
alien invasive species throughout the world’s terrestrial biomes, so that novel
encounters between species are widespread. These new interactions raise numerous
conservation concerns, including disruptions to long-evolved prior associations
between native species in the invaded environments.
Some specialised mutualisms are especially vulnerable. Impacts of invasive spe-
cies on naturally occurring mutualisms they encounter have been projected as ‘often
profound’ (Traveset and Richardson 2011), but have only recently been considered
specifically. Likewise, the needs of invasive species to encounter possible mutualist
partners to functionally compensate and replace those from which they have become
separated can be critical. For alien plants, pollination and seed dispersal mutualisms
may be important for successful invasion and, unless preoccurring mutualist part-
ners arrive with the plant, surrogates are necessary in the receiving environment.
Traveset and Richardson (2011) argued that plant invasions are facilitated by a large
array of native pollinators, seed dispersers and other potential associates in the new
areas and that, more broadly, invasibility could be expected to correlate positively
with native species richness in raising the chances for such opportunities to occur. A
schematic diagram (Fig. 8.1) summarises the major barriers that an introduced plant
must progressively negotiate, and the stages at which some key mutualisms most
affect the invasion dynamics.

DOI 10.1007/978-3-319-58292-4_8
154 8 Impacts of Alien Invasive Species
Fig. 8.1 Schematic representation of the continuum that includes various barriers that an intro-
duced plant may negotiate in order to progressively become invasive; bars show the various phases
and stages at which different categories of mutualisms are most influential in invasion dynamics
(Traveset and Richardson 2011)
8.2 Impacts on Mutualisms
Generalised (diffuse) mutualisms are often claimed to be resilient to changes from

invasive species. Countering this simplistic assumption, which is based largely on
the premise that ‘interaction frequency is a good measure of partner quality’ (Prior
et al. 2014), Prior and her colleagues noted that introduced alien species interact
with new partners or coevolved taxa in new communities, and those communities
may be changed in very complex ways. They concluded that ‘co-introduced mutual-
ists could be major drivers of ecological change’. The experimental context they
used was of artificially created communities of combinations of an invasive seed-
dispersing ant (Myrmica rubra) or a native parallel (Aphaenogaster rudis) with inva-
sive and native myrmecochorous plants. The invasive ant, co-introduced with an
invasive myrmecochorous plant, protected this alien species rather than native plants.
The invasive plant dominated in presence of M. rubra, whilst native plants domi-
nated with A. rudis, leading to strongly divergent effects on the plant communities.
The ability of specialist pollinators to track or follow their host plants into new
environments has received considerable attention, and is central to maintaining or
8.2 Impacts on Mutualisms 155
Table 8.1 The development of fig-fig wasp pollination mutualisms on the island of Anak Krakatau,
Sunda Strait, Indonesia: some key events and timing based on successive observations
Year Condition and changes
1979– No fruiting figs found; the vector fruit bat Cynopterus sphinx recorded in 1982
1983
1984 Fig seeds transported by C. sphinx; moderate-sized trees present but not yet
producing figs. Fig wasps found in aerial fallout as potential pollinators if figs were
available.
1985 Ficus fulva and F. septica observed fruiting for the first time; the obligatory
frugivore cuckoo dove (Macropygia phasianella) observed on the island.
1986 Both fig species pollinated successfully, and inquiline fig wasps present; further
fruit-eating pigeons and fruit bats observed
1992 Four dioecious Ficus species bearing figs found; postulated that wasps from
conspecific trees in the local population were now the major source of potential
pollinators.
After Compton et al. (1988, 1994)
reestablishing mutualisms in invaded areas. Fig wasps, for example, may be trans-
ported by wind over many kilometres, and track figs dispersed by frugivorous birds
(notably fruit pigeons) or bats to islands or other isolated environments, where their
continuity after such chance arrival depends entirely on the subsequent arrival of
those obligate mutualists. Figs have thus successfully colonised island environ-
ments in many parts of the world. One ‘strategy’ employed by some figs in such
areas is to inhibit the normal abortion of figs that remain unpollinated—such trees
may thereby gain more time to attract pollinators in environments where their very
existence is uncertain. Observations on Anak Krakatau, Indonesia, largely sup-
ported the sequence for island colonisation by figs discussed by Janzen (1979),
namely that seed dispersal by frugivorous vertebrates is followed by wasp colonisa-
tion. That sequence was traced on Anak Krakatau (Compton et al. 1988, 1994), as
in Table 8.1, with the fruit bat Cynopterus sphinx the likely main fruit vector. The
four species of Ficus found on Anak Krakatau in 1992 had lower pollination rates in
the two most abundant species, but these still had lower pollination rates than on the
older islands of the archipelago, only a few kilometres away (Compton et al. 1994).
The two fig species found in only low numbers had sufficient pollinators.
Although those distances from the older islands of the Krakatau archipelago
were small, elsewhere fig wasps have been reported to disperse much further—jour-
neys of at least 100 km were necessary in order to encounter figs on Kauai, Hawai’i,
for example (Corner 1958). This propensity can help rapid recovery of mutualisms
after catastrophic loss.
The Category 4 Hurricane Andrew in 1992 devastated Ficus aurea in Florida,
and many trees were denuded of foliage and figs (Bronstein and Hossaert-McKey
1995), with likely loss of all (or nearly all) wasp larvae present. Rapid recovery
ensued, with new foliage appearing within 3 weeks and largely complete after about
5 months. By that time, also, the fig wasp population on newly developed figs had
reached similar levels to prehurricane populations. Two interacting features were
suggested to enable this: (1) that the fig wasps can indeed disperse over longer
d istances than generally expected, as implied above, and (2) rapid re-establishment
of asynchronous fig phenologies, allowing wasps to persist through presence of dif-
ferent fig stages on the same trees. The latter is a normal feature of F. aurea, and
may speed recovery of pollinator populations. Whatever the mechanism(s) of such
recovery, it seems that pollinator loss in some fig populations, even when very
severe and seemingly dramatic, may be rather transient. More broadly, such obli-
gate interactions can be highly resilient to catastrophic losses. Re-establishment of
mutualisms, although sometimes difficult to monitor or detect, appears not to be
uncommon. More than 60 exotic species of Ficus have reached Florida, most of
them introduced as ornamental plants (Nadel et al. 1992). Some have become weedy
invaders of natural vegetation and, so, targeted for suppression. The three main
weedy species are pollinated by immigrant fig wasps and are naturalised, regularly
producing seeds or seedlings. Many other fig species there do not support perma-
nent pollinator populations, but have been noted as ‘sporadic seeders’ for which
regular pollinators have not been reported directly in Florida.
The re-establishment of fig-fig wasp mutualisms by invasions has attracted con-
siderable attention and, as McKey (1989) noted, the long–range dispersal of figs
poses severe problems in (1) reaching a minimum number of trees in a new area to
constitute a viable population amongst which pollinators can function, and (2) that
the pollinating wasps are very short-lived and, despite being produced in enormous
numbers, are unlikely to regularly survive long-distance dispersal. The Anak
Krakatau case, above, involves only short distances, but clearly any mutualism can-
not be re-established, or a new one formed, unless some parallel sequence of species
arrivals occurs. Historically, however, many figs have been introduced deliberately,
many of them as ornamentals, beyond their natural ranges so that a number of mutu-
alisms, with both partners introduced, now occur far beyond their original range and
are essentially naturalised (McKey 1989). The ‘more natural’ colonisation sequences
are exemplified by Anak Krakatau and the earlier records of figs and fig wasps from
the older islands of that archipelago (Thornton et al. 1996). Fig wasps were intro-
duced deliberately to the Hawaiian archipelago in conjunction with schemes to
incorporate alien figs in forest restoration from the 1920s on. In that context, with
Hawai’i lacking native figs, biological confusions from introductions to areas
already supporting native figs do not arise but, again as McKey (1989) commented,
subsequent intra-archipelago movements by both natural and human-aided means—
including dispersal from planes —have provided a series of contexts that are now
difficult to interpret, despite considerable contemporary historical documentation.
The spread of many invasive plants depends on pollinator activity, and any such
plant that is unable to recruit generalist or, more rarely, specialist pollinators from
the newly found native biota or retain its original pollinator(s) may become vulner-
able. Most are satisfactorily attended by newly encountered generalist pollinators
and, as Richardson et al. (2000) discussed, pollination limitation does not appear to
impede the spread of most alien plants.
If pollinators reach new areas, as accidental or deliberately introduced aliens,
novel interactions with species never previously encountered are highly likely.
Some may lead to modifications of preexisting mutualisms in the invaded region.

The resin-collecting orchid bee Euglossa viridissima, from Mexico and Central
America, is now abundant in Florida, following its invasion since 2003 (Pemberton
and Liu 2008), and visits many species of plants there. It also continues its earlier
associations (as in Mexico) as a pollinator of the newly naturalised Dalechampsia
scandens (Euphorbiaceae). In Florida this association may promote the plant’s fur-
ther naturalisation—and the bee has also formed regular, novel associations with
other plant species. The fate of specialised pollinators or plants that reach new areas,
where replacement resources may be limited or absent, is an increasing concern
because partners of an established mutualism may not stay together when this
occurs. In the above example, E. viridissima could also exploit ornamental resin-
forming species of Dalechampsia (from Peru) and Clusia (native to Brazil), and so
could maintain its needs on alien flora.
The independent or conjoint arrival of both partners of a well-established mutu-
alism into areas where neither occurred previously offers opportunities for ecologi-
cal changes affecting the native biota and changed dynamics of the association
through exposure to other possible partners. Especially for intricate asociations,
such joint arrivals may be more frequent than anticipated, but in many other cases
the partners are likely to be separated. The example of Cecropia (with well-known
ant mutualists, Azteca spp., in its native neotropics, p. 76) encountering ants in
Hawai’i (Wetterer 1997) indicates how such invasives may fare. Neither is native to
the archipelago, and Cecropia obtusifolia and about 40 ant species (none being
Azteca) had by then been introduced by human activity. The tree had spread on
several islands in low-elevation pastures and wet forests. Five species of ants were
collected during detailed searches (of all foliage and by splitting trunks to give total
inventory of species present) on Cecropia on the island of Hawai’i. All of them
were common tramp species and more than one species were found together on 13
of the 18 trees examined. No ants were found inside the hollow trunks and the two
most abundant species (Anoplolepis longipes, Technomyrmex albipes) were both
tending Hemiptera on foliage. Despite the presence of numerous Mullerian bodies
on the trees, Wetterer never observed ants eating these—but, because the cecropia
trees thrived in Hawai’i, confirmed that C. obtusifolia is not obligatorily dependent
on ant mutualists. Its success there might therefore reflect the absence of most
Neotropical herbivores and competitors, so that mutualist protection from these is
not needed in the newly invaded environment. Nevertheless, the tramp ants might
afford some protection from the herbivores in the new range area. In their native
neotropics, some nonmyrmecophytic species of Cecropia compensate for the
absence of mutualistic ants by increased chemical defence in young foliage and
physical toughness of older leaves (C. sciadophylla: Latteman et al. 2014). Earlier,
Faveri and Vasconcelos (2004) demonstrated that Cecropia can show very low lev-
els of herbivory when defensive ants are absent, but the mechanism was not inter-
preted in that account.
An invasive species can affect an established mutualism it encounters in either of

the following two ways:
1. It can displace a mutualist partner and assume that role to some extent, com-
monly as a relative generalist replacing an ecological specialist—and in more
extreme cases leading to breakdown of the mutualism because that partner can-
not be functionally replaced.
2. It can act as a competitor or predator of such a partner, perhaps altering the effec-
tiveness and outcomes of the mutualism.
Either impact can extend further within the local environment, or landscape.
Invasive species can have very disruptive effects on long-coevolved native mutual-
isms, whatever their partners may be. Novel competitive and trophic interactions are
legion, and the impacts of aggressively dominating generalist species are amongst
the most severe concerns for the well-being of numerous specialised insect species
and their restricted resource taxa.
Many specialist pollinators, for example, are competitively displaced by invasive
social Hymenoptera, imposing widely ramifying changes and impacts. The North
American Western yellowjacket wasp (Vespula pensylvanica) is a generalist preda-
tor and scavenger that can form enormous colonies that are active yearround. In
Hawai’i, the wasp is a nectar thief on the endemic tree Metrosideros polymorpha and
displaces native pollen vectors (notably endemic Hylaeus bees and birds, honeyeat-
ers [Drepanididae]) and other alien pollinators, leading to substantial community
restructuring (Hanna et al. 2014). Large-scale removal of V. pensylvanica led to sig-
nificant increase of effective bee pollination compared with control sites on which
wasps prevailed. Inflicting alien generalist pollinators onto specialised mutualisms is
a major conservation concern, not least through uncritical deliberate introductions of
honeybees and bumblebees for crop production into areas where they do not occur
naturally. In extreme cases, novel modes of pollination are introduced, such as the
buzz pollination and corolla-mutilating tactics of many bumblebees, that differ from
the more widespread ‘simple’ pollen transfer undertaken by many bees and others
and can lead to two major conservation concerns. First, they may threaten endemic
plants on which specific native pollinators depend (also affecting those pollinator
species through increased competition for resources and depleting the overall supply
of resources available) and, second, alien ‘sleeper weeds’ that could not thrive previ-
ously because of lack of effective pollinators, may come to the fore as competitors
in natural vegetation and displace plant species to which native specialist pollinators
may have become adapted. Novel interactions, as in numerous other ecological con-
texts, may variously be highly beneficial or seriously detrimental to mutualisms.
For the Metrosideros polymorpha example above, the alien Apis mellifera
appeared to act in the role of a ‘substitute pollinator’, but one whose effectiveness
was hampered by Vespula pensylvanica activity. Suppression of the wasps led to
effective replacement of native pollinators by honeybees (Hanna et al. 2013) which,
together with enhanced visitation rates by Hylaeus spp., led to increased fruit pro-
duction. Nevertheless, the wider ecological effects of this transformation from a
native pollinator-dependent community to one dominated by an alien pollinator
Fig. 8.2 Shift of a plant-pollinator mutualism to antagonism: (a), reciprocal benefit (mutualism)
with pollinator and plant both gaining from the association; (b), when relative abundances become
markedly unequal, one partner (e.g. pollinator) can overexploit the other (e.g. plant), so, whilst the
interaction remains profitable for the former, it can diminish or even be detrimental to the low-
abundance species, if interaction costs equal or exceed the interaction benefits; (c), a highly abun-
dant plant species can have nonreciprocal effects on a low-abundance animal partner, causing
negative indirect effects (dashed arrow) as competition for pollinators if, in its presence, a low-
abundance plant species incurs lower net interaction benefit (based on Aizen et al. 2014)
may also impose many adverse effects on native flora and accompanying pollinator
mutualisms. Such taxon substitution may indeed help to conserve some key interac-
tions, but deliberate restoration incorporating novel interactions may need very
careful consideration of the less conspicuous disruptions that may result.
A framework suggested by Aizen et al. (2014) predicted that when invaders
increase disproportionately in abundance compared with the interacting partners,
mutualisms can shift to antagonisms, as indicated in Fig. 8.2, and also increase
impacts of exploiters (Chap. 9). The combined impacts of excessively abundant
alien pollinators such as honeybees may both hamper reproduction of native plants
and promote reproduction of alien plants, so imposing influences at both population
and community levels.
The high densities reached by some invasive pollinators can hamper reproduc-
tion of native plants as their native pollinators are displaced and specialised associa-
tions are eroded, and also by promoting reproduction of alien plants, including
‘sleeper weeds’ as competitors. Invasive bees, in particular, are acknowledged
widely as likely to alter previously undisturbed plant-pollinator networks—but,
anomalously, may also increase pollinator abundance. Aizen et al. (2014) also pro-
posed that pollinator introductions intensify costs of mutualisms amongst native
plants, and that increased visitation rates by less diverse (invasive species) pollina-
tor assemblages may eventually lead to reduced reproduction by native plants.
Facilitation of invasive plant establishments by invasive generalist pollinators may
be especially frequent on isolated islands, where ecological ‘shifts’ from native
mutualisms may be induced amongst relatively undisturbed communities (Kaiser-
Bunbury et al. 2009).
‘The effects of invasive plant species on native pollinator species are largely
unknown’ (Mitchell et al. 2009). Different studies have yielded very disparate out-
comes—and components of differing contexts that influence these can include (1)
the relative abundance or density of the invasive species; (2) the regional abundance
of pollinators; and (3) the abundance and characteristics of co-flowering species.
Mitchell et al. noted the predominance of invasive alien plants, many of which must
share pollinators with native species in their newly occupied receiving environ-
ments. At least in theory, these could generate novel evolutionary changes—some
of which might in due course facilitate further range expansions. The influences of
invasive species on plant-pollinator interactions are affected by local context in
many ways. A plant’s ability to benefit from abundant efficient newly arrived polli-
nators (such as some alien bees) depends on the number of ovules presented for
fertilisation or the resources available for seed development. The capability of the
pollinators to exploit abundant floral resources provided by a single plant (such as a
massed monoculture crop) reflects the time available for foraging, the quality of
plant resources and the availability of other essential resources. Discussed by Aizen
et al. (2014), assessing the costs and benefits of the interaction, and how the balance
is affected by abundant invaders, can illustrate changes in pollination networks and
shift the original mutualism toward a more antagonistic interaction. In their exam-
ple, of invasive bees (Apis mellifera, Bombus terrestris), a combination of promot-
ing reproduction of alien plants and reducing reproduction of native plants can
occur. Invasive bees can also exclude native pollinators effectively, by direct com-
petition or aggressive behaviour. In some cases, this may lead to decline of the
native specialist pollinator taxa.
An unusual outcome occurred for two Andean orchids in the invasive range of
introduced Bombus spp. and A. mellifera, where these bees displaced their sole
known pollinator, the locally endemic Bombus dahlbomi (Sanguinetti and Singer
2014). They were thereby a tangible threat to the bee and potentially also to the
orchids (Brachystele unilateralis, Chloroea virescens). However, the introduced
bees were successful pollinators and were possibly an important contribution to
conservation of these orchids through the imposed novel interaction. Despite this,
Sanguinetti and Singer warned that any introduction of such bees should be under-
taken with great care because of the wider impacts they may have on local plant
communities.
Features such as flexible behaviour, efficient tracking of resources, effective

exploitation and indiscriminate foraging—all of which characterise some classic
invasive pollinator bees—may give those invaders the ability to ‘forcefully’ enter
complex native pollination webs and outcompete native species. Those invaders
may also have morphological differences from natives that can influence intrusions.
Discussing entry of alien bees into Tasmania, Goulson et al. (2002) suggested that
introduced honeybees (large, long-tongued) may be able to reach nectar from deeper
flowers and forage over longer daily periods than the numerous small, short-tongued
native bees, giving them competitive advantage. However, and as in some other
contexts, the historical record of impacts is highly incomplete because records have
not accrued from the time that honeybees first arrived, in the early nineteenth cen-
tury. The most susceptible native species are likely to have succumbed long before
any information was available. That others are still present but still considered vul-
nerable is testament to continuing and changing impacts.
High abundance of alien bees can impose various forms of exploitative competi-
tion on native species, reduce their visitation rates to flowers, decrease their partici-
pation in effective pollination, and induce long-term genetic changes in native flora
through traits such as promoting self-pollination or disturbing established patterns
of dispersal and gene flow (Dupont et al. 2004). In some contexts, those aliens are
relatively poor pollinators, but individual inefficiency is compensated by very high
numbers present. However, it is largely unknown whether such impositions by alien
species affect the reproductive output of the native bee taxa despite high likelihood
of this threat to their well-being. They are commonly implied, as by Sugden and
Pyke (1991) for declines of the Australian native bee Exoneura asimillima as the
density of hives of Apis mellifera increased locally.
As a general pattern, analyses by Aizen et al. (2008) suggested that arriving alien
insects are initially rare and participate in pollination networks through interactions
with native generalists, with wider interactions limited by their scarcity. Some of
these generalist invaders may be ‘stronger’ than the native species present, and dis-
place them, so that any ecologically specialised (usually rare) native species interact
exclusively with the invaders. Those invaders may then become ‘super-generalists’
and can modify the structure of pollination networks.
Some impacts of invasive species on pollinators are more dramatic. The alien
predatory green anole lizard (Anolis carolinensis) markedly reduced native flower-
visiting insects on some of the Ogasawara Islands of Japan, as a result of which
honeybees are the predominant flower visitors on the largest islands of the archi-
pelago (Abe et al. 2011). Smaller islands, where the anole is absent, have retained
the native endemic bees and other flower visitors—but the outcome from the anole
has been destruction of the endemic pollination system and its replacement by
forms of alien mutualism. The impact is attributed to a combination of loss of native
pollinators and increased reproductive success and invasiveness of alien plants
facilitated by the honeybee.
Perhaps most commonly, alien invasive insect nectar-robbers or seed predators
function as ‘exploiters’ (Chap. 9) of mutualisms, and Dunphy et al. (2016) pointed
out that exploitation may be severe because the native plants have had no o pportunity
to evolve defences against the invasives. Thus, in environments such as Hawai’i

with no native ants, invasive ‘tramp ants’ and others exploit native plant nectar to
the detriment of much of the specialised endemic flora as well as native pollinators.
Likewise, elaiosome robbing by invasive ants can disrupt some normal associations
in myrmecochory. Some ants, for example, rob elaiosomes without transporting the
seeds, leaving the propagules below the parent plants and (lacking any reward) unat-
tractive to other (native) ants for dispersal (Horvitz and Schemske 1986).
In Ontario, Canada, the native ant Aphaenogaster rudis (s.l.), is regarded as a
‘keystone mutualist’ because it disperses numerous myrmecochorous seeds. It was
the main dispersal agent for Canadian wild ginger (Asarum canadense,
Aristolochiaceae). The invasive European slug Arion subfuscus, in contrast, con-
sumed Asarum elaiosomes without dispersing seeds. Laboratory trials confirmed
that this robbing by slugs reduced seed dispersal considerably.
The Red imported fire ant, Solenopsis invicta, is a notoriously harmful invader
and has caused serious environmental concerns in many places (as examples from
the substantial literature, see Vinson 1994, Holway et al. 2002), but its direct impacts
on native mutualisms have only rarely been appraised. Those effects undoubtedly
occur, as part of a much wider range of impacts as it invades. In the eastern forests
of North America, S. invicta comes into contact with numerous herb species that
have ant-dispersed seeds. Experimental ‘cafeteria’ trials confirmed that fire ants
removed almost all presented seeds of six such plant species (Zettler et al. 2001).
This, in itself, could aid myrmecochory, but the ants scarify or destroy many seeds
and the invasion was assessed as a potential threat to spring ephemeral herbs in
rendering myrmecochorous plants increasingly susceptible to seed loss. As else-
where (such as South Africa: Bond and Slingsby 1984), invasive ants can have pro-
found effects on plant community structure through affecting seedling recruitment,
but the impacts on mutualisms can be very subtle. The Argentine ant (Linepithema
humile) in South African fynbos affects the mymecochorous Mimetes cucullatus
(Proteaceae) by traits such as moving seeds for shorter distances than native ant
species and leaving seeds exposed to predation on the soil surface rather than stor-
ing them in subterranean nests. Bond and Slingsby suggested that its continuing
invasion might lead to extinction of numerous endemic Proteaceae by ‘slow and
subtle attrition’ of their seed resources.
One mutualism-related resource sought by both fire ants and native species is the
high-carbohydrate honeydew produced by Hemiptera, with both mutualisms and
the accompanying resources important for dietary need and colony growth of inva-
sive ants (Lach et al. 2010), with the competitive and nutritional advances gained
increasing their wider impacts on native ants. However, Wilder et al. (2013) claimed
that the reality of introduced ants competing with native species for access to mutu-
alisms had not been tested experimentally—and, likewise, the impacts of loss of
carbohydrates provided by mutualism to the native ants that can also exploit insect
prey have been suggested to be relatively unimportant. Competition between S.
invicta and native ants (three predominant species of Dolichoderinae) was studied
in Texas and Alabama, to demonstrate that fire ants indeed excluded the native spe-
cies from the critical resource of mutualist honeydew. That resource contributed to
native ant colony growth even when unlimited alternative food (insect prey) was
provided. Key nutrients may be critical and mediate the effects of alien species on
native species. Easily digested liquid carbohydrates allow worker ants to maintain
the correct balance between nourishing themselves and providing resources for the
brood, and Wilder et al. suggested that this need may participate in the repeated
evolutions of relevant mutualisms where high-quality carbohydrate sources are
assured. Nutritional studies of such contexts may also contribute more widely to
interpreting the impacts of invasive species on natives. The flexibility amongst ants
of trophobiosis with Hemiptera is associated positively with invasiveness (Oliver
et al. 2008): it seems that invasive ants can easily adopt new partners, facilitated by
their preadaptations to mutualism with newly encountered native species, as dis-
cussed by Mondor and Addicott (2007). Thus, Argentine ants responded to the
alarm calls of the native poplar aphid (Chaitophorus populicola) in California, and
numbers of attending ants increased when such signals were detected, exemplifying
how such preadaptations may contribute to alien species displacing the native taxa
previously participating in such mutualisms.
The examples noted above simply display that the impacts of invasive ants on
native ant mutualisms are both complex and varied, with many of them poorly
known. Argentine ants enhance protection of the barrel cactus Ferocactus virides-
cens in California, but also lead to reduction of native ant diversity (Ludka et al.
2015) which, in turn, is positively associated with the seed mass produced. Those
positive and negative impacts, however, appeared to balance each other because
seed mass on cacti in invaded and uninvaded areas did not differ. More broadly, the
overall impact of Linepithema humile was considered likely to be negative in this
system, because the ant also disrupts pollination. Cactus bees (Diadasia spp.) spent
less time visiting flowers on cacti tended by Argentine ants than those with the
native Crematogaster californica, the most abundant native ant present.
Writing more generally on alien ant invasions, Rabitsch (2011) commented that
all facets of global change can contribute to driving invasion, at all stages of the
invasion process. Influences on native mutualisms are thus only one facet of their
diverse ecological impacts in invaded ecosystems, but contribute to wider scenarios
of biotic homogenisation. Changes to invaded communities may initially be subtle:
in south-eastern Australia, L. humile can virtually eliminate a keystone native seed-
dispersing ant (Rhytidoponera victoriae) (Rowles and O’Dowd 2009), but overall
levels of seed dispersal were maintained, albeit with changes in effectiveness across
different plant species. Variability in ant species, with L. humile selectively favour-
ing some diaspores and neglecting others, can indirectly reduce seed dispersal of
some native plants and also facilitate spread of alien plants through imposed changes
to preoccurring mutualisms.
References
Abe T, Waa K, Kato Y, Makino S, Okochi I (2011) Alien pollinator promotes invasive mutualism
in an insular pollination system. Biol Invasions 13:957–967
Aizen MA, Morales CL, Morales JM (2008) Invasive mutualists erode native pollination webs.
PLoS Biol 6(2):e31. doi:10.1371/journal.pbio.00600431
Aizen MA, Morales CL, Vazquez DP, Garibaldi LA, Daez A, Harder LD (2014) When mutualism
goes bad: density-dependent impacts of introduced bees on plant reproduction. New Phytol
204:322–328. doi:10.1111/nph.12924
Bond W, Slingsby P (1984) Collapse of an ant-plant mutualism: the Argentine ant (Iridomyrmex
humilis) and myrmecochorous Proteaceae. Ecology 65:1031–1037
Bronstein JL, Hossaert-McKey M (1995) Hurricane Andrew and a Florida fig pollination mutual-
ism: resilience of an obligate interaction. Biotropica 27:373–381
Compton SG, Thornton IWB, New TR, Underhill L (1988) The colonisation of the Krakatau
Islands by fig wasps and other chalcids (Hymenoptera: Chalcidoidea). Philos Trans R Soc
Lond B 322:459–470
Compton SG, Ross SJ, Thornton IWB (1994) Pollinator limitation of fig tree reproduction on the
island of Anak Krakatau (Indonesia). Biotropica 26:180–186
Corner EJH (1958) An introduction to the distribution of Ficus. Reinwardtia 4:325–355
Dunphy SAM, Prior KM, Frederickson ME (2016) An invasive slug exploits an ant-seed dispersal
mutualism. Oecologia 181:149–159
Dupont YL, Hansen DM, Valido A, Olesen JM (2004) Impact of introduced honeybees on native
pollination interactions of the endemic Echium wildpretii (Boraginaceae) on Tenerife, Canary
Islands. Biol Conserv 118:301–311
Faveri SB, Vasconcelos HL (2004) The Azteca-Cecropia association: are ants always necessary for
their host plants? Biotropica 36:641–646
Goulson D, Stout J, Kells AR (2002) Do alien bumblebees compete with native flower-visiting
insects in Tasmania? J Insect Conserv 6:179–189
Hanna C, Foote D, Kremen C (2013) Invasive species management restores a plant-pollinator
mutualism in Hawaii. J Appl Ecol 50:147–155
Hanna C, Foote D, Kremen C (2014) Competitive impacts of an invasive nectar thief on plant-
pollinator mutualisms. Ecology 95:1622–1632
Holway DA, Lach L, Suarez AV, Tsutsui ND, Case TD (2002) The causes and consequences of ant
invasions. Annu Rev Ecol Syst 33:181–233
Horvitz CC, Schemske DW (1986) Seed dispersal of a neotropical myrmecochore: variation in
removal rates and dispersal distance. Biotropica 18:319–323
Janzen DH (1979) How to be a fig. Annu Rev Ecol Syst 10:13–51
mutualisms on oceanic islands. Perspect Plant Ecol Evol Syst. doi:10.1016/j.ppees.2009.10.002
Lach L, Parr CL, Abbott KL (eds) (2010) Ant ecology. Oxford University Press, Oxford
Latteman TA, Mead JE, DuValt MA, Bunting CC, Bevington JM (2014) Differences in anti-
herbivore defences in non-myrmecophyte and myrmecophyte Cecropia trees. Biotropica
46:652–656
Ludka J, Levan KE, Holway DA (2015) Infiltration of a facultative ant-plant mutualism by the
introduced Argentine ant: effects on mutualist diversity and mutualism benefits. Ecol Entomol
40:437–443
McKey D (1989) Population biology of figs: applications for conservation. Experientia
45:661–673
Mitchell RJ, Flanagan RJ, Brown BJ, Waser NM, Karron JD (2009) New frontiers in competition
for pollination. Ann Bot 103:1403–1413
Mondor EB, Addicott JF (2007) Do exaptations facilitate mutualistic associations between inva-
sive and nat8uve species. Biol Invasions 9:623–628
References 165
Nadel H, Frank JH, Knight RJ (1992) Escapees and accomplices: the naturalization of exotic Ficus
and their associated faunas in Florida. Fla Entomol 75:29–38
involving ants. J Ecol Biol 21:1597–1608
Pemberton RW, Liu H (2008) Naturalized orchid bee pollinates resin-reward flowers in Florida:
novel and known mutualisms. Biotropica 40:714–718
Prior KM, Robinson JM, Dunphy SAM, Frederickson ME (2014) Mutualism between co-
introduced species facilitates invasion and alters plant community structure. Proc R Soc B
282:20142846 doi.org/10.1098/rspb.2014.2846
Rabitsch W (2011) The hitchhiker’s guide to alien ant invasions. BioControl 56:551–572
Richardson DM, Allsopp N, D’Antonio CM, Milton SJ (2000) Plant invasions- the role of mutual-
isms. Biol Rev 75:65–93
Rowles AD, O’Dowd DJ (2009) New mutualism for old: indirect disruption and direct facilitation
of seed dispersal following Argentine ant invasion. Oecologia 158:709–716
Sanguinetti A, Singer RB (2014) Invasive bees promote high reproductive success in Andean
orchids. Biol Conserv 175:10–20
Sugden EA, Pyke GH (1991) Effects of honeybees on colonies of Exoneura asimillima, an
Australian native bee. Aust J Ecol 16:171–181
Thornton IWB, Compton SG, Wilson CN (1996) The role of animals in the colonization of the
Krakatau Islands by fig trees (Ficus species). J Biogeogr 23:577–592
Traveset A, Richardson DM (2011) Mutualisms: key drivers of invasions … key casualties of inva-
sions. In: Richardson DM (ed) Fifty years of invasion ecology: the legacy of Charles Elton.
Blackwell, Oxford, pp 143–160
Vinson SB (1994) Impact of the invasion of Solenopsis invicta (Buren) on native food webs. In:
Wiliams D (ed) Exotic ants: biology, impact and control of introduced species. Westview Press,
Boulder, pp 240–258
Wetterer JK (1997) Ants on Cecropia in Hawaii. Biotropica 29:128–132
Wilder SM, Barnum TR, Holway DA, Suarez AV, Eubanks MD (2013) Introduced fire ants can
exclude native ants from critical mutualist-provided resources. Oecologia 172:197–205
Zettler JA, Spira TP, Allen CR (2001) Ant-seed mutualisms: can red imported fire ants sour the
relationship? Biol Conserv 101:249–253
Chapter 9
Exploiters of Mutualisms
Keywords Ants • Cheaters • Community structure • Ecological impacts • Nectar

robbing • Networks • Over-exploitation • Parasites of mutualisms • Pollen robbing
9.1 Introduction: Roles of Additional Species
All mutualisms are, in principle, prone to invasion and exploitation by other species
that seek reward without conferring any benefits in return. Those species are most
frequently native taxa, some with long coadapted relationships with the mutualists,
but additional impacts may occur from invasive alien species as a further aspect of
the disruptions they may cause. The impacts of additional species in a mutualism
raise important questions, perhaps the most relevant being to address the implica-
tions arising from them removing the reward that a mutualist offers to its partner as
a central component of the mutualistic interaction. The primary associations
between mutualists provide opportunities for many other species to either exploit
the interaction in some way, or to derive some benefits through less-defined, but
regular and predictable, co-occurrence.
The core module of mutualism formed by fig-fig wasp, yucca-yucca moth, or any
similar obligate association can become the realm invaded by numerous other native
taxa, many of which scarcely influence the mutualism but others having more dis-
ruptive impacts and leading also to changes in the encompassing community
through those additional or substitutive interactions. Thus, additional species in a
fig-fig wasp mutualism may influence the balance of benefits among the major part-
ners (Schatz et al. 2006). Presence of ants strongly reduced numbers of parasitoid
wasps on three Ficus species, and experimental exclusions of ants led to increased
abundance of these non-pollinating wasps on both F. racemosa and F. condensa,
whilst the complementary step of supplementing the number of ants led to decreased
non-pollinator wasp abundance on F. carica. Reduction of parasitoids benefits the
pollinators, in that in these cases ants were considered to be ‘indirect mutualists’ of
the plant-pollinator mutualism. As for many other mutualisms, these associations,
in Schatz et al.’s words are ‘enmeshed in a network of complex interactions’. Many
of these additional fig wasps can be effective pollinators. The host-specific parasit-
oid Diaziella spp. (Pteromalidae, Sycoecinae) of Ficus paracamptophylla, for

DOI 10.1007/978-3-319-58292-4_9
168 9 Exploiters of Mutualisms
Fig. 9.1 The interactions inferred around the mutualism between Ficus obtusifolia and the polli-
nating fig wasp Pegoscapus hoffmeyeri, from the accompanying non-pollinating wasps, herbivo-
rous moth larvae, and ants. Moth larvae and the competing Idarnes wasps affect pollinators and
seeds directly; gallers and parasitoids affect them more indirectly in having negative effects on the
fig fruit. Solid arrows are negative interactions; dashed arrows are weak negative interactions;
interactions between Ficus and Pegoscapus are positive (Jander 2015)
example, can transport more pollen than the tree’s primary pollinator (Jousselin
et al. 2001), the difference probably reflecting their larger size than the normal
Waterstoniella species.
A complementary need, of course, is for the mutualists to protect their associa-
tions from exploiters, in addition to primary roles such as protection of plants from
herbivory. Thus, Ficus obtusifolia and its two closely related cryptic pollinating fig
wasps (known as ‘Pegoscapus hoffmeyeri A and B’) constitute a mutualism that is
exploited by a variety of other small wasps (Fig. 9.1). Presence of ants conferred
several important benefits to that association (Jander 2015), as (1) reduced number
of figs aborted; (2) reduced herbivory of figs; and (3) reduced impacts of parasitoid
wasps, so collectively resulting in more pollination and greater seed numbers in ant-
protected figs. Some parasitoid wasps, however, have evolved behaviour to evade
ant attacks. Both pollinators and figs showed dramatically increased ‘fitness’, by
nearly fivefold and threefold, respectively, when ants were present. The structure of
the interaction web demonstrates the influences of the ant (Azteca trigona) on the
various groups of insects affecting the mutualism directly or indirectly. In addition
to benefiting the pollinators, the ants also benefit the tree through the variety of
associations – in Jander’s (2015) concluding phrase, the ants ‘seem to be both com-
mon and powerful protectors of the fig mutualism’.
Refined ‘signalling’ between the partners is essential for some obligate mutual-
isms to persist so that, for example, many lycaenids use the presence of their ant
9.1 Introduction: Roles of Additional Species 169
partners as a cue or signal for oviposition. Such signals may be open to usurpation
and be exploited by natural enemies to aid discovery of suitable prey or hosts. Thus,
for the Australian Jalmenus evagoras (p. 122), both spiders and a braconid wasp
parasitoid use the ant signals to increase their discovery rates (Elgar et al. 2016).
‘Choice-chamber’ experiments implied that chemical signals used by the ant
Iridomyrmex mayri also attracted the braconid, so using the host ant as a cue in find-
ing the lepidopteran hosts. In the same trials, the wasp was apparently repelled by a
non-tending ant, I. rufoniger. Web-building spiders were seemingly attracted by the
ant scents to trees supporting Jalmenus colonies and, by locating webs close to lar-
val aggregations, were able to increase captures of adult butterflies as they emerged.
Exploiters can develop intricate roles with or within an existing mutualism, and
those of facultative ant-plant mutualisms facilitated by extrafloral nectaries (p. 68)
as a major food focus for the ants may support both a variety of other generalist
nectar-seekers and a few far more specialised taxa with more unusual ecological
roles. The tree Qualea grandiflora (Vochysiaceae) in Brazil has abundant extrafloral
nectaries tended by a range of ants that are known to protect the plants from her-
bivory. However, the nectaries are used also as a specialised breeding site by the
drosophilid fly Rhinoleucophenga myrmecophaga, whose larvae are ‘sit-and-wait’
predators on the ants lured to the nectaries (Vidal et al. 2016). At least 15 species of
ants were recorded as prey of the sedentary fly larvae. The active potential prey
becomes stuck on the adhesive shelter constructed by the fly larva on top of the
nectary, but that structure does not prevent the nectary from functioning normally.
The overall interactions are thereby complex, with presence of the fly larvae induc-
ing cascade effects. Fly-inhabited foliage supported fewer ant foragers, increased
herbivore damage, and greater leaf damage. The ‘basic’ mutualistic benefits between
Qualea and the ants are thus jeopardised by the predatory fly larvae (Vidal et al.
2016).
Many other species occur regularly and predictably in some associations with
mutualisms. Some are clearly exploiters, leading Bronstein (2001) to distinguish
several categories of exploiters to exemplify the variety of contexts in which they
occur, as (1) individuals from outside the mutualism that return no benefit to either
mutualist; (2) species or individuals that have lost mutualistic roles over time but
retained their ability to obtain benefits from their former partners; and (3) individu-
als that, even when given the opportunity to behave mutualistically towards their
partners, do not do so. These rather different categories are usually not distinguished
but, as Bronstein emphasised, reflect that ‘exploitation’ is ‘a suite of phenomena
differing in both ecological and evolutionary meaning, unified simply by failing to
offer benefits in return for their activities’. Exploiters, therefore, are not
mutualistic.
However, some ambiguities are inevitable in such interpretations. Susceptibility
of a two-species mutualism to exploitation may arise from a conflict generated by
two trophic partners, a context explored for the myrmecophilous larvae of the lycae-
nid Thisbe irenea (p. 130) in which, in a situation paralleled by some other lycae-
nids, both the food plant and the larvae themselves use mutualistic ants for protection
(De Vries and Baker 1989). T. irenea larvae drink the extrafloral nectar of the food
Fig. 9.2 Occupation of Croton billbergianus in Panama by larvae of the butterfly Thisbe irenea,
based on 1378 observations in diurnal censuses. Information for each instar allocates larvae among
three positional categories: stem (open, larvae found resting on woody plant parts away from foli-
age); leaf (black, larvae found on leaf tissue); other (dotted, larvae found on or next to an extrafloral
nectary). Note that instars four and five rest on stems during the day but when active at night are
frequently found with their head over or adjacent to the nectaries (DeVries and Baker 1989)
plant (Croton billbergianus, Euphorbiaceae) as well as eat the foliage and nutri-
tional benefits for early instar larvae from nectar include accelerated growth to
reach the third instar (Fig. 9.2). At that stage, they begin to attract ants as mutualists
and for defence. The early instars pass much of their time on or near the extrafloral
nectaries, but later instars hide elsewhere during the day and move up stems to feed
(on both foliage and nectar) at night. Both ants and extrafloral nectar benefit the
larvae, and the suggestion that this might be an invasive association rather than a
true three-species mutualism was made because the substantial loss of foliage was
considered unlikely to benefit the plant.
9.2 Exploiters and Cheaters
Intruders into established mutualisms, whether habitual or more opportunistic, are

thus both taxonomically and ecologically varied. Some have negligible impacts,
whilst others pose significant disruptions. Exploiters show three major characteris-
tics, summarised by Sachs (2015) as (1) taking advantage of a resource or service
provided by mutualists that (2) increase the fitness of the exploiter, and (3) decrease
the fitness of the mutualist that is exploited.
One major complication is assessing the influences of additional exploiter spe-
cies that invade a mutualism – such as specialist inquilines or other herbivores that
appear only late in a plant’s flowering season, when they arrive far too late to con-
tribute to pollination (should they be able to do so) but perhaps adding considerably
to levels of seed predation or mortality. Even for mutualist pollinating species, late-
emerging individuals can function more as exploiters than mutualists, but almost all
9.2 Exploiters and Cheaters 171
obligate pollinator-seed predator mutualisms experience these additional influences.

Some exploiters are parasitoids, or otherwise directly affect mutualists. Others are
more benign, but the roles of most species remain somewhat unclear. They are usu-
ally considered as a ‘one-way flow’, by gaining in some way from the mutualism
but providing nothing in return.
Bronstein (2001) also framed three general predictions over the costs of exploita-
tion by additional species in mutualisms. These have only rarely been investigated
together in specific cases, but all were suggested in the social parasite Trachymyrmex
and were all discussed for that context (Adams et al. 2012). Those predictions are:
(1) if the exploiter prevents one mutualist partner from gaining benefit from the
other partner, and that benefit links to the success of that first partner, the cost of
exploitation will increase as the benefit decreases – with the extreme case being if
that mutualist is caused to avoid its partner. Adams et al. cited the case of fungus
gardens being abandoned by attine ants affected by exploiters; (2) the cost of exploi-
tation increases with the value of the exploited organism or commodity: in this same
example, the fungus garden may be lost if the mutualist ant workers are killed; and
(3) the overall ecological context of the mutualism clearly influences the costs/ben-
efits accruing – most obviously, perhaps, high abundance of exploiters may impose
far greater costs than if only a few individuals are present.
‘Parasites’ of mutualisms involving ants and plants have been described repeat-
edly, with non-defending ants occupying sites such as the thorns of swollen thorn
acacias (p. 64). Defensive and feeding behaviour and impacts on plant growth were
compared between the mutualist Pseudomyrmex ferrugineus and the putative para-
site ant P. gracilis (Clement et al. 2008), revealing that P. gracilis can both (1)
exclude the mutualist ant from colonisation and (2) negatively affect plant condition
and growth rate. When workers or queens of P. ferrugineus were encountered by P.
gracilis, they were attacked, so that the parasite’s presence could prevent the regular
mutualism from becoming established and so expose the host plant to increased
herbivory. The parasite also has smaller colonies than P. ferrugineus, and produces
more alates. It also has a far wider range of external food resources rather than being
confined to feeding on the acacia food bodies alone. The parasite’s pattern of repro-
ducing early and from small colonies was regarded as an ‘r-strategy’ by Clement
et al., who suggested that this might represent a more widespread condition for such
parasites of mutualisms. Defensive behaviour was compared by noting ant responses
to the stimulus of touching the plant surface with forceps – seemingly a very simple
act but one that readily elicited strong differential responses from the two ant spe-
cies. P. ferrugineus responded commonly by highly aggressive behaviour, but P.
gracilis workers, in smaller numbers, fled from the disturbance (Fig. 9.3).
Impacts of predators that invade obligate mutualisms may be complex, not least
in relation to their effects on the mutualists or on exploiters – with a stronger effect
on either of these components leading to greater relative negative or positive
impacts. Schatz et al. (2008) investigated the balance between predatory ants and fig
wasps in Brunei and showed that fig wasps could be a major prey resource for ant
colonies. They validated three relevant predictions as: (1) for dioecious Ficus spe-
cies, more ants were present on male trees than on female trees; (2) those ants are
Fig. 9.3 The behavioural

responses of two species of
myrmecophyte-associated
Pseudomyrmex ants to
simulated attack on
Acacia. Six levels of
defensiveness (A-F,
representing a gradient
from [A] workers
patrolling on >75% of
leaves, rapid recruitment,
and highly aggressive
behaviour to [F] no
workers on the plant
surface) are shown by
percent frequency of
responses for P.
ferrugineus (open,
N = 147) and P. gracilis
(black, N = 60) (Clement
et al. 2008)
Table 9.1 Numbers of ants observed on different dioecious fig species in Brunei
Fig species
Ant Total a b c d e f g
Crematogaster A 2894 738/91 0/1 53/18 216/24 467/9 91/8 162/16
Oecophylla 1041 813/94 0/0 0/0 56/0 55/0 11/0 12/0
smaragdina
Tetramorium spp. 139 80/0 44/0 0/0 0/0 0/0 0/0 12/3
Anoplolepis 132 3/33 0/0 19/0 18/3 49/1 0/0 6/0
gracilipes
Camponotus spp. 110 31/1 0/0 25/0 0/4 33/2 8/1 5/0
Tetraponera spp. 45 16/3 0/0 0/2 10/0 7/0 6/0 0/1
Dolichoderus spp. 44 36/0 0/0 0/0 0/0 0/0 2/0 6/0
(Others) 10 4/5 0/0 0/0 0/0 0/1 0/0 0/0
Numbers are ant individuals on male/female trees of (a) Ficus condensa, (b) F. megaleia, (c)
F. uncinata, (d) F. fulva, (e) F. aurata, (f) F. deltoidea arenaria, (g) F. deltoidea borneensis (Schatz
et al. 2008)
mostly dominant predatory species; and (3) the ants would nest in the trees. Ants,
indeed, were more than 10-fold more abundant on male than on female trees, with
similar trends across the seven host taxa examined (Table 9.1). Activity of the two
most abundant ants, both dominant predators (Oecophylla smaragdina,
Crematogaster sp.), focused on captures of fig wasps, and ant nests were present on
most male trees, where ants ‘ambushed’ wasps as they emerged from ostioles.
The term ‘cheater’ is used widely, and sometimes uncritically, but refers properly
to individuals of a normally or previously mutualistic species that utilise resources
from the partner without providing anything in return. This state contrasts with the
9.2 Exploiters and Cheaters 173
wide category termed ‘aprovechados’ (those taking advantage), referring to a non-

mutualistic species that utilises mutualistic resources or services – as do many
loosely defined ‘exploiters’. The key distinction, then, is that ‘cheaters’ have devel-
oped by changes from previous mutualists to become exploiters, whilst ‘aprovecha-
dos’ are species newly associated and which have not previously participated in the
mutualism they now exploit (Mainero and del Rio 1985). The importance of this
distinction, emphasised also by Frederickson (2013), reflects the origins and possi-
ble concerns over invasive species impacts. Distinguishing between the two catego-
ries thus necessitates information on whether the exploiter is derived evolutionarily
from the mutualists or is from ‘outside’. The term ‘parasites’ has sometimes been
preferred (Yu 2001) – but, whatever epithet(s) are applied, the principle of an inter-
loper usurping benefits that are normally exchanged between regular mutualists
applies. The concept thereby excludes taxa such as herbivores on fig trees or yuccas,
for which no participation in the true mutualisms on those plants is evident. However,
whatever categorisation is applied, ‘once a mutualistic relationship has arisen, the
appearance of cheaters becomes highly probable’ (Boucher et al. 1982) and may be
important in driving coevolution, because counteradaptations against cheating ben-
efit non-cheaters and so promote the mutualism itself.
The most common presumptions about roles of cheaters are that they are highly
harmful to the mutualisms they exploit. In reality, their significance is often unclear,
and recent attempts to define ‘cheating’ emphasise the need to incorporate the
impacts on fitness in a wider evolutionary context (Jones et al. 2015). In addition,
the term ‘cheat’ has been used in several different ways and contexts, and that vari-
ety itself can be confusing in discussions of interactions and evolutionary processes,
that extend well beyond their immediate relevance in mutualisms. A thoughtful dis-
cussion by Ghoul et al. (2013) led to definition based on four main distinctions: (1)
whether the species can include cooperation as an option in its behavioural reper-
toire – determining if the ‘cheater’ could cooperate under particular conditions and
whether it may have evolved from cooperation; (2) whether deception occurs – such
as through the cheater functioning by exploiting a signalling system; (3) whether the
cheater is conspecific with the co-operator – affecting whether the cheater can
replace competitors or simply reduce the relative benefits of the cooperation; and
(4) whether the cheater is facultative or obligate – determining the outcome of the
association, contrasting (a) cooperation being maintained with facultative cheating
but (b) lost if obligate cheaters predominate.
Two functionally distinct forms of aprovechados have been described (Mainero
and del Rio 1985), as (1) Type 1, that benefits both mutualists, and (2) Type 2, that
receives a negative impact from one or other mutualist. Examples they cited for
each include (Type 1) (a) Batesian mimics of mutualisms, that are palatable to pred-
ators, with cases across butterflies, bees, and hoverflies, in particular, and (b) repro-
ductive mimicry in plants, in which pollinators spend time visiting the mimic, which
is the aprovechado; and (Type 2) flower nectar robbers (below) that derive nectar
from the flowers whilst competing with pollinators for that resource and are likely
to cause negative effects by diminishing the nectar supply then available to
pollinators.
The categories were appraised for figs-fig wasps and yuccas-yucca moths by
Addicott et al. (1990), and the subtle ecological nuances present have drawn consid-
erable more recent attention. Thus, two species of Tegeticula yucca moths were
noted by Pellmyr (2003) as non-pollinating cheaters that lay eggs directly into fruits
at different stages of development. Their larvae eat seeds whilst coexisting with
larvae of the pollinator species. The presence of these cheater moths can substan-
tially increase seed losses, although they do not seem to compete with any resident
obligate species. Because pollinating yucca moth adults live for only a few days,
each species must gain access to its host during its individual short flowering
period – so that moth populations must be adapted locally to the flowering period of
the particular host yucca. If a single pollinator species was to utilise several host
species with largely non-overlapping flowering seasons as in the scenario discussed
by Pellmyr (2003), this would require intraspecific polymorphism in their emer-
gences, with a different peak to correspond with each host. Phenological specialisa-
tion relates to the high level of specificity (with >70% of the pollinating moth
species being monophagous). Because both the cheater species exploit several host
species, they may have greater phenological opportunity and reduced competitive
potential with individual pollinator species.
A further intriguing form of cheating within the yucca-yucca moth association
suite also emphasises the unique features of particular participant species. It exem-
plifies the issue discussed further by Pellmyr (2003), namely that cheating plants
can occur through evolving mechanisms that maintain pollination but prevent pol-
linator larvae from developing and destroying seeds. Yucca baccata has a highly
unusual system whereby all individuals are pollinated by a species of Tegeticula, but
many individual plants produce no moth larvae (Bao and Addicott 1998).
Surprisingly, a very high proportion of individuals (>70%) were involved. The
mechanism involves two different flower types, one having far fewer viable ovules
than the other. Larvae in flowers with few ovules perish. A high proportion of Y.
baccata individuals are thus obligate cheaters as the moth eggs and larvae undergo
exceptionally high mortality. Female moths lay in several inflorescences, so com-
plete reproductive failure for any individual is unlikely. Bao and Addicott added that
a remarkable feature of Y. baccata is that >70% of the population depends on the
other 30% to sustain the pollinator moth population.
As noted briefly earlier, species in mutualisms that are exploited by cheaters may
develop countermeasures that deter or thwart the cheaters, as so-called ‘exploitation
barriers’. In the context of pollinator thefts by cheaters, for example, ‘exploitation
barriers are traits that make access to floral structures, such as pollen or nectar, dif-
ficult’ (Santamaria and Rodriguez-Girones 2015). They include floral traits such as
tubular corollas and other structural idiosyncrasies in flower shape that exclude
most pollinators and promote specialisation, attractants of floral predators, red flow-
ers, and others. As discussed by Santamaria and Rodriguez-Girones, whilst such
barriers can foster specialisation, with risks of forgoing pollination when the spe-
cialist is rare or absent, this can be countered by the progressive accumulation of
resources at flowers eventually becoming available to other pollinators so that spe-
cialisation may be to different pollinators across different community contexts.
9.3 Protection from Over-Exploitation 175
Exploitation barriers have been implied to promote resource partitioning (Rodriguez-

Girones and Santamaria 2010).
One relevant and widespread context is theft of nectar by non-pollinating ants,
linked with the production of ant repellents by the plants and facilitating access by
‘proper’ pollinators. Two hypotheses, not mutually exclusive, have been advanced
to support the development of ant repellents and were defined and investigated by
Gonzalvez et al. (2015). The ‘nectar protection hypothesis’ is that repellents were
developed to deter ants from stealing nectar without pollinating the flowers, whilst
the ‘pollinator protection hypothesis’ has the rather different emphasis that repel-
lents were developed to discourage ants from disturbing or thwarting pollinator
activity. The hypotheses were tested together in China and Singapore using the
aggressive ant species Oecophylla smaragdina to assess the strength of repellence
on 32 bee-pollinated plants. For each plant, growth form, nectar production, and the
size of the bee pollinators were evaluated in relation to floral repellent levels. Large
bees (Xylocopa) were attacked by Oecophylla, but not driven away, whilst smaller
bees (Trigona, smaller Apis spp.) were more vulnerable to the attack. Their visita-
tion rates were reduced when ants were present. In this set of trials, flowers with
nectar and flowers pollinated by small bees produced more nectar than nectarless
flowers and flowers pollinated by Xylocopa, and flowers from trees had more ant
repellence than those on shrubs or vines. The study thus supported both of the above
hypotheses, adding the parameter that repellence level may also be related to plant
growth form. In particular contexts, ant repellents may be either direct or indirect
exploitation barriers.
9.3 Protection from Over-Exploitation
More generally, Gonzalvez et al. (2015) predicted that if the ants are parasites of the
plant-pollinator interaction (such as by being nectar thieves or predators of the pol-
linators), the flowers may develop repellents, but may develop alternatives if they
are mutualists (as defensive organisms that prey on herbivores or parasites, or as
pollinators). Effects of a plant’s ‘nectar barriers’ link with levels of pollinator spe-
cialisation and effectiveness within the pollination networks of the wider local com-
munity (Santamaria and Rodriguez-Girones 2015) in which some form of resource
partitioning amongst co-occurring pollinators may develop easily. Predatory ants
may respond to volatile chemicals from syconia (Ficus racemosa: Ranganathan and
Borges 2009), whilst in this case trophobiont-attending ants in the same environ-
ment did not respond to fig or fig wasp odours. The two predatory taxa involved
(Oecophylla smaragdina, Technomyrmex albipes) apparently use fig-associated
scents to facilitate encounters with prey wasps, using the same signals as those
intended for pollinators. Those same signals can also attract, and hone specificity of,
non-pollinating wasps (Proffit et al. 2009). Both the specific pollinator of Ficus
hispida and a parasite of this mutualism are enabled to locate this host among
Fig. 9.4 Outline map of the eastern United States, showing distribution of two pollinating species
of yucca moths (Tegeticula yuccasella, T. cassandra) and the cheater moth (T. intermedia) (Darwell
et al. 2016)
sympatric fig species. Dependence on the chemical could, in turn, constitute a bar-
rier to host shifts by both guilds of wasps.
Within a mutualism, exploiters could represent invasion by an additional species
or evolve from a mutualist by ‘reversal’, a situation that has apparently occurred at
least twice in yucca moths. Those moths then no longer pollinate but continue to
coexist with pollinator species. The species of Tegeticula that eat seeds of Yucca
filamentosa were studied by Marr et al. (2001): T. yuccasella is a pollinator and T.
intermedia a cheater. They appear to seldom compete for seeds and not to be limited
by seed availability. But, because cheaters can develop only in pollinated fruit, they
must coexist with the pollinator moth species, and Marr et al. showed that these two
species can coexist with little conflict. Over 3 years of survey, numbers of pollina-
tors and cheater larvae did not reach any threshold, whereby seed numbers became
limiting. Moth distributions were strongly influenced by climate. In Florida, two
species are present, but T. cassandra – a geographically restricted sister species of
T. intermedia, but a pollinator – is predominant (Darwell et al. 2016, Fig. 9.4).
Populations of both these pollinator species were heavily influenced by competi-
tion. In contrast, T. intermedia has expanded its range to much of eastern North
America and, as a cheater, is released from competition for oviposition sites.
Mechanisms thus exist that can prevent excessive exploitation of the yuccas by
the pollinating yucca moths, but few other parallel examples have so far been clari-
fied. One analogue involves Botanophila flies (Anthomyiidae) as vectors of fungal
spermatia (so functionally being the same as ‘pollinators’) of the ascomycete fun-
gus Epichloe elymi (Bultman et al. 2000). Fungal reproductive output tended to
increase as more fly eggs were deposited, and larval feeding was not strongly cor-
related with the number of those eggs. The interaction was interpreted as stable, a
condition attributed to high mortality of fly eggs and larvae, which increased as egg
9.4 Changed Communities 177
loads rose. Although considered a ‘balanced antagonism’ rather than a mutualism,

increased mortality of the flies clearly linked with decreased consumption of the
fungus.
However, not all species that co-occur regularly with mutualists are evident
exploiters. Both the yucca-yucca moths and fig-fig wasps systems, for example, are
also associated with consistently present and ecologically restricted and specialised
other taxa. Figs have suites of non-pollinating parasitoid wasps (West and Herre
1994), and many non-pollinating prodoxid moths (of Prodoxus, a sister genus of the
pollinating moths) occur on yuccas (Pellmyr et al. 2006).
9.4 Changed Communities
The parasitoid and other non-pollinating fig wasps are relatively poorly known but
are often very common, highly diverse, and can outnumber the more familiar pol-
linating taxa. Many parallel the pollinating wasps in being largely host-specific and
so form the foundation of complex, host-restricted communities. They fall into two
main ecological categories. Some are gall-formers, inducing development of the
ovules within which the larvae feed and may potentially compete directly with pol-
linators. Others use ovules already occupied by another wasp larva. Most species of
both categories are host-plant specific, so form definable community entities associ-
ated with different Ficus species. From studies on Idarnes wasps, some can have
detrimental impacts on the host fig’s reproduction, specifically through reducing
pollinator wasp production and so affecting the host’s potential to disperse pollen.
These wasps, however, are not direct parasitoids of the pollinators but compete for
the same pool of flowers in which the larvae of either group can develop. Despite
apparently long coevolutionary associations, the two wasp groups have developed
very different life cycles in order to exploit the figs: pollinating wasps usually enter
only one fruit to reproduce, whilst others oviposit in different fruits on individual
trees.
The dynamics among figs, pollinating wasps, and other fig wasps present are
often highly asymmetrical, reflecting how the interacting species variously cooper-
ate or compete with each other and leading to unpredictable oscillations in their
populations (Wang et al. 2014, 2015). Limited dispersal of some participants, such
as those imposed by fragmentation of forest patches and lack of any connecting cor-
ridors, may lead to creation of some metapopulation structure in which localised
extirpations and recolonisations replace a more predictable and stable population
structure. Ecological variety is imposed on any such local dynamics. In a study in
Cote d’Ivoire, Kerdelhue et al. (2000) distinguished three major ‘ecological clus-
ters’ on the timing of oviposition in the figs and the oviposition behaviour of the
wasps, as (1) wasps that form large galls through ovipositing early in fig develop-
ment, together with their specific parasitoids, and occurring in relatively small num-
bers of figs. Their reproductive strategy involves laying few eggs in several figs, and
several groups of wasps have independently adopted this pattern; (2) gall-makers
that oviposit in the fig cavity. After entering through the ostiole, again a habit that
has evolved independently in several different chalcidoid wasp groups, they occur
only in some figs but are often abundant where found, because the entrapped females
lay all their eggs in a single fig; (3) species that oviposit from outside and develop
in normal-sized galls of the same size as pollinator galls; many have very long ovi-
positors to facilitate this pattern. The survey confirmed that very complex wasp
communities occur and those ecological interactions can be strong, depending
largely on oviposition strategy and local relative abundance. The total number of
wasp species varies considerably across different Ficus species, even when the hosts
are closely related.
Ficus racemosa in southern China is pollinated by a single species of fig wasp,
Ceratosolen fusciceps, but five other, non-pollinating, wasps reproduce within the
fig (Wang et al. 2014). Elsewhere such acolyte communities can comprise up to 30
additional wasp species associated with a single fig species, representing the major
guilds of herbivores forming galls on or in the figs, and parasitoids of other wasps
present. That most non-pollinating wasps are themselves specific to a given species
of Ficus means that the ecological entities to which they contribute are themselves
well defined units comprising up to about 30 species spanning a variety of interac-
tions. Their tangible advantages for study (Anstett et al. 1997) encompass compari-
sons of communities with different structures and richness as tools for understanding
the evolution and susceptibility of pollinator traits and insights into the origins and
evolution of mutualisms.
In general, pollinating and cheating fig wasps only rarely occur together – but the
African Ficus sycomorus (often deemed a synonym of F. racemosa) sometimes sup-
ports different species of Ceratosolen with the two roles on the same fig (Compton
et al. 1991). The regular active pollinator is C. arabicus, whilst C. galili is only a
very occasional pollinator. Pollination by C. galili was simply through pollen car-
ried on the body surface, and the number of resulting seeds was negligible. Larvae
of this wasp commonly developed in figs without any pollinated flowers.
On Ficus racemosa (or F. sycomorus), both pollinating and non-pollinating fig
wasps are subject to predation by ants and other arthropods, but only one dominant
ant species occurred on each individual tree (Wang et al. 2014). The few ants
involved are all generalists. Observations of the Weaver ant (Oecophylla smarag-
dina) over time showed that the predators can benefit both the tree and mutualist
pollinators, in part because the ants reduced the density of a key non-pollinating
wasp (Platyneura mayri) that competes with the pollinator to lay eggs in fig flowers.
Although the finding that a predatory ant species can change the community struc-
ture of fig wasps appeared novel, Wang et al. noted that it may, in fact, prove to be
a common outcome. Their summary of the interactions (Fig. 9.5) indicated the
major effects projected.
Wider interactions between figs and fig wasps, whether pollinators or parasites,
are largely mediated chemically and, although the wasps may be preyed upon by
ants, other ants may be trophobionts.
Ants are indeed well-known to influence dynamics of fig-fig wasp mutualisms,
with predation on the pollinating wasps sometimes restricting fig male reproductive
Fig. 9.5 Interspecific relationships between ants, pollinating wasps and non-pollinating wasps
associated with Ficus racemosa, with interactions marked as beneficial (+) or antagonistic (−);
solid lines are direct interactions; broken lines are indirect interactions. The trophic cascade
impacts of ants include negative effects on non-pollinating wasps by both density-mediated indi-
rect interaction with primary productivity or changes to morphology, phenology, physiology or
behaviour of herbivores leading to cascading effects on plant growth, plant biomass, and related
ecosystem functions (after Wang et al. 2014)
function. Two guilds of ants occur on the Mediterranean Ficus carica. One guild
tends Hemiptera, and the other comprises two co-dominant ant species that prey
mainly on pollinating fig wasps (Schatz and Hossaert-McKey 2003) and, to a far
lesser extent, on parasitic wasps affecting the pollinators. The major predator,
Crematogaster scutellaris, was more abundant than Pheidole pallidula which,
together with several relatively uncommon species, also preyed on the wasps. The
heavier predation on pollinators than on parasitoids might negatively influence pol-
lination, but further investigation was needed to determine this.
Prodoxus moths coexist on yuccas with the better-known mutualist pollinators,
but do not participate in those mutualisms, and feed on plant parts other than seeds.
They are sometimes referred to as ‘bogus yucca moths’. Many of the 22 species
reviewed by Pellmyr et al. (2006) are relatively host-specific, and their larvae are
also site-specific as they feed specifically either on fruit, flowering stalk, or leaf tis-
sue. Stalk feeding is the ancestral habit, from which three separate origins of fru-
givory have flowed, together with one for leaf-mining. These moths are known also
for undergoing long larval/prepupal diapause that reflects the very sporadic flower-
ing habits of some yuccas. Powell (2001) noted one species diapausing for 30 years.
Following from other studies on Yucca filamentosa (above), the interaction between
this host and the moth Prodoxus decipiens, which feeds on four species of yucca,
proved difficult to quantify (Althoff et al. 2004). This state contrasted with those
Fig. 9.6 The geographical

patterns of coevolution
indicated by the
relationships between
proboscis lengths of the
pollinating fly Prosoeca
ganglbaueri and flower
depths of the primary plant
it attends, Zaluzianskya
microsiphon, across the
range of each in KwaZulu
Natal, South Africa
(Anderson and Johnson
2008)
involving conventional cheaters or pollinators, in which the costs or benefits can be

assessed by the numbers of seeds produced or eaten. Effects of Prodoxus may
include both direct (consumption of plant resources that otherwise could be used for
reproduction) and indirect (modifying the mutualism between pollinators and yuc-
cas) impacts on yucca reproduction. In this study, feeding by Prodoxus larvae
reduced the amount of plant nitrogen available for reproduction, but this did not
appear to affect seed output or quality.
The diverse and highly individualistic orchid-insect pollinator pollination asso-
ciations have attracted considerable attention in defining roles of exploiters in such
mutualisms, with many apparently unique contexts likely to represent more widely
distributed parallel circumstances as further examples accrue. Visual deceptions,
with the orchid flower developing features (notably of colour and shape) resembling
a potential mate for a specific insect pollinator have for long intrigued naturalists.
This kind of Batesian mimicry may sometimes be an exploitation of a mutualism
(Bronstein 2001) through the exploiter mimic attracting or diverting pollinators of
other (reward-producing) plants without providing any reward of their own
(Fig. 9.6). Such mimetic resemblances can have negative impacts on both the pol-
linators (which waste time and energy visiting flowers that provide no nectar reward)
and models (which can lose fitness when growing among mimics, because pollina-
tors are diverted from areas where no reward has been obtained, or develop prefer-
ences for alternative flowers with higher reward). Sexual deception in Orchidaceae
has evolved separately on a number of occasions, and has attracted wide attention.
Levels of specificity vary somewhat – some morphologically distinct species (for
example, of Ophrys: Gogler et al. 2009) can attract the same pollinator, even when
allopatric, and with the same attractant odour that essentially mimics the phero-
mone of the pollinating insect (Scopece et al. 2007).
The South African orchid Disa nivea is a rare floral mimic of the nectar-producing
Zaluzianskaya microsiphon (Scrophulariaceae), and the two plants grow together
on some sites. The nemestrinid fly Prosoeca ganglbaueri is apparently the exclusive
pollinator of Zaluzianskaya, which provides the predominant nectar source for the
fly over much of its flight season (Anderson et al. 2005). The orchid resembles its
model closely in appearance, and the fly apparently does not discriminate between
the two species whilst seeking nectar. It was the only insect observed carrying pol-
linia of Disa and is also known to pollinate several other flower species in the region.
The orchid and the putative model in this association are both adapted for pollina-
tion by long-tongued flies: the ‘tongue’ of Prosoeca ranges from 19–42 mm long
across different populations (Anderson et al. 2005). Disa here fulfils the basic crite-
ria that define a floral mimic, and the association between Zaluzianskaya and
Prosoeca was considered a true pollination mutualism, with the single pollinator
exploited by the orchid.
This association has been studied across the geographical range of both mutual-
istic species, to determine whether regional differences in flower corolla size and fly
proboscis length showed any correlated variation. Both features were highly vari-
able among populations (Anderson and Johnson 2008), but the population means
were significantly correlated (Fig. 9.6). Further, experimental movements of short-
corolla plants to areas with normally long-corolla flowers and long-tongued flies
resulted in them having lowered mean seed set/flower. It seemed likely that local
adaptations within different populations had occurred, suggesting that trait diversi-
fication across populations can be related to coevolution.
Floral mimicry is widespread amongst orchids, but few examples have been
appraised in the above detail. Visual and chemical attractants for pollinators are
widespread, and mimics of nectar-hosts are frequent. At one study site in Panama,
at least nine of the 30 orchid species visited by euglossine bees deceive their polli-
nators by such means (Roubik and Ackerman 1987). Many orchids were not
extremely specialised and had up to four visiting bee species. Sexual deception
amongst orchids is also notable, whereby the flowers are adapted to attract specific
male insects (almost always of solitary groups of Hymenoptera) that attempt to
mate with the flowers and so act as pollinators. These include many cases where no
floral reward is given (Gaskett 2011). Many sexually deceptive orchids are largely
pollinator-specific, and exploit the visual and chemical cues used by the male insect
to discover a mate. Many orchids are ‘food-deceptive’, resembling rewarding flow-
ers but not providing any reward in return for pollination. Food deception is more
common than sexual deception (Gaskett 2011).
Approximately a third of all orchid species produce no nectar and, because pol-
len is amalgamated into pollinia, this is also not generally available as a trophic
reward. The origins of ‘empty flowers’, discussed by Smithson and Gigord (2003),
imply that such orchids may benefit from mimicking reward-bearing plants, but
many are pollinated by naïve pollinators seeking a reward.
Deceptive flowers are diverse, and many insect pollinators will commonly
encounter flowers in which either (1) the nectar reward has been depleted temporar-
ily by previous visitors or (2) no nectar is provided because the flowers are perma-
nently nectarless. However, as Renner (2007) noted, rewards for pollinators can
take forms other than nutrition – the plants may, for examples, provide oviposition
sites, larval feeding sites, nest-building materials, or simply rendezvous for territo-
rial or courtship display. Bees and other pollinators can assess flowers on the basis
of appearance and scent, often in a general way to, in Renner’s words, ‘develop
expectations about floral syndromes and their typical rewards’.
9.5 Nectar and Pollen Robbers
The roles of nectar robbers (p. 67), to some ecologists classical universal ‘cheaters’,
were reappraised by Maloof and Inouye (2000) Published studies on the impacts of
nectar robbers on plant seed set gave equal numbers (six of each category) for nega-
tive, neutral, and positive outcomes. For their primary focal group (Bombus spp.,
bumblebees) it seemed that either harmful or beneficial effects might occur: pollina-
tion may ensue whilst the robber collects nectar. Robbers may influence plant fit-
ness by affecting behaviour or the usual pollinators, with some of the changes
summarised in Table 9.2. Again, diverse possibilities exist and Maloof and Inouye
concluded that it should not be assumed initially that a given nectar robber is either
a cheat or a mutualist: indirect effects are easily overlooked. Effects of nectar rob-
bers depend in part on four factors and lead to realisation that those robbers are, in
fact, often mutualists. The four factors signalled for relevance are (1) the identity of
the legitimate pollinator; (2) the growth form of the plant and its influence on forag-
ing patterns – for example, whether geitonogamy is possible or usual; (3) how much
nectar the robbers remove – if all, the legitimate pollinators may not function, if
some, these may remain; and (4) the resources available in the wider environment,
especially whether alternative accessible nectar resources occur that allow pollina-
tors to remain. Debate over the functional roles and status of species in such asso-
ciations continues, reflecting the reality that the terms ‘mutualists’ and ‘robbers’
represent the extremes of a complex continuum along which many nuances and
Table 9.2 The potential changes in behaviour of pollinators that can be caused by nectar reduction
by nectar robbers, with the ‘probable result’, assuming that all other behaviour remains unchanged
on robbed and unrobbed flowers (Maloof and Inouye 2000)
Behavioural change in pollinator Probable result for plant
Flight distances increase Pollen flow distance increases
Fewer flowers visited on each Reduces geitonogamy in self- fertile flowers
inflorescence
Visit more flowers per unit time Pollen flow rate increases
Reduced time spent in each flower Less pollen transfer
Change foraging tactic to pollen Depends on morphology of plant
collection only
Avoidance of robbed flowers Could increase pollen flow distance if pollinator does not
switch to different species
Switching to a different plant No pollen transfer
species
9.5 Nectar and Pollen Robbers 183
shifts in balance occur (Fedriani et al. 2012). They, and others, have warned against
relying on the simplistic generality of this often-implied dichotomy.
Nectar robbing is widespread and has been reported to occur in at least 59 plant
families. In contrast, it seems that plant defences against nectar robbers are not as
common, although examples proposed for this role and listed by Richardson (2004)
included cases of toxic nectar, production of sap from damaged flowers, thickened
corollas to deter robbers from chewing entrance holes, clustered flowers, and
benefits from ant defenders. Nevertheless, plants have developed many ways to
deter nectar robbers (Irwin et al. 2010), and the phenomenon is experienced by most
flowering plants with tubular corollas or other specific forms for nectar supply. The
impacts are thus both direct (physical damage to plant reproductive structures and
removal of nectar and/or pollen) and indirect (through changes to pollinator behav-
iour and visitations). One evolutionary ‘dilemma’ that arises is that features that
deter nectar-robbing may also discourage pollinators and render such traits difficult
to interpret. Lack of knowledge of, for example, how chemical defence effective
against robbers also affects key pollinators and pollen is one of several critical unan-
swered questions listed by Irwin et al., and in need of further elucidation.
Some plants, indeed, can counter the potential impacts of nectar robbers in favour
of their roles as pollinators, by features such as protecting the nectar chamber from
external piercing or necessitating mechanical force to retrieve the nectar. The car-
penter bee Xylocopa californica arizonensis is an effective pollinator of Ocotillo
(Fouquieria splendens, Fouquieraceae) in Texas (Scott et al. 1993) and, rather than
being primarily a nectar robber on this plant, has a more mutualistic relationship
based on both pollen and nectar availability during the bee’s breeding season. The
association, however, is localised, occurring only when Ocotillo grows close to the
bee’s nesting sites, mainly in the stalks of Agavaceae. This association between the
bee and Fouquieria was not considered to have developed through any reciprocal
coevolution or specific adaptation of Xylocopa because the bee visits a variety of
flowers and can obtain nectar from flowers with a variety of structures.
The ambiguity between ‘mutualists’ and ‘antagonists’ as terms applied to nectar
robbers has been investigated in some elegant experiments involving the behaviour
and impacts of flower visitors. For example, Richardson (2004) used fluorescent-
coloured powders to mimic pollen removed from plants of desert willow (Chilopsis
linearis) by the nectar-robbing bee Xylocopa californica in Arizona, and employed
black light surveys at night to detect different coloured ‘pollen’ from robbed and
unrobbed flowers on source trees when it was present on other trees. Different
source trees and colour codes were used on each of 8 days, with 50–72 flowers
treated each morning. The values of this system for studying nectar-robbing included
(1) Chilopsis is pollen-limited, so pollination by such vectors is important, and (2)
the plant is self-incompatible and produces massed flowers simultaneously. Benefits
to the plant are thus greater if pollinators leave the plant after visiting only a few
flowers, but pollinators also have the benefit of reduced dispersal costs if they forage
on many flowers on the same tree. Further, (3) Chilopsis has no obvious way to
deter or exclude nectar robbers, unlike some other Bignoniaceae. Effects of nectar-
robbing are only evident through changes in pollinator behaviour on Chilopsis, and
Richardson concluded that the nectar-robbing often cost the plants little and, in
some cases, was beneficial.
Interactions between the plant Corydalis caseana (Papaveraceae), its natural and
the only known pollinator (Bombus appositus), and a related nectar robber bumble-
bee species (B. occidentalis) in Colorado indicated that the nectar robbers had no
effect on fruit or seed set. Maloof (2001) suggested that they might cause increased
pollen flow between plants, by changing the behaviour of the pollinator.
Pollen theft is an allied syndrome, also diverse but that has been studied to a
lesser extent than nectar theft and which may affect plant reproduction more directly
(Hargreaves et al. 2009), through contributing to low pollination efficiency. Pollen,
taken by a considerable variety of insect ‘thieves’, is a major – in some cases, sole –
source of nutrition for them or their larvae, and is actively sought by many insects.
Some also function as pollinators, but many others disrupt normal pollination sys-
tems. Hargreaves et al. recognised the two categories of (1) habitual thieves, which
always remove pollen but never cross-pollinate, so always reduce potential cross-
pollination of the species involved, and (2) conditional thieves that pollinate but
only inefficiently, even casually, and in functional terms act as thieves if other flower
visitors provide more effective pollination, but pollinators when they are the plant’s
only or predominant visitors. Pollen theft is suspected to occur much more widely
than at present documented and also to have greater impacts on plant reproduction.
Likewise, the community context of pollen theft is still poorly understood: thieves
sustained by one plant species may be important pollinators of co-occurring taxa,
for example, as proposed by Baker et al. (1971).
Rather than a ‘pollinator versus robber’ polarisation, these two activities com-
monly intermingle, in that at least some pollinators of many plants also rob, and
some robbers can act as pollinators. Honeybees and some bumblebees, for example,
can fulfil different roles on the same plants, as pollen collectors, nectar collectors,
and secondary nectar robbers. Richardson and Bronstein (2012) expanded this
range of ‘mixed behaviours’ to include also some Lepidoptera and Diptera and
noted two rather different observations that might be involved in development of
such behaviour. These were (1) that some generalist pollinators cannot obtain nectar
from particular plant species by their usual foraging techniques and may switch to
nectar-robbing rather than change to forage on other species, and (2) amongst
insects that cannot chew holes in corollas, through which to obtain nectar, the pres-
ence of true nectar robbers to ‘prepare the way’ can lead to increased secondary
nectar-robbing. Again, honeybees sometimes shift to nectar-robbing when bumble-
bees are also present.
Richardson and Bronstein suggested that switching to secondary nectar-robbing
may become more common at plants which are attended by diverse arrays of gener-
alist foragers that include primary nectar robbers. Mathematical modelling of the
roles of nectar robbers has demonstrated some scenarios in which robbers are essen-
tial for the maintenance of mutualists. In some model outcomes (Wang et al. 2015),
invasions by robbers can lead to extinction of pollination mutualisms and, in turn,
extinction of the robbers themselves.
9.6 Defences Against Exploiters 185
Considerable variations and impacts of both nectar-robbing and pollen-robbing

in time and space are widespread. In studying the pattern of nectar-robbing by the
bumblebee Bombus occidentalis on four different host plants growing together in
Colorado over up to 7 years, Irwin and Maloof (2002) noted finding ‘an amazing
level of complexity in plant-robber association’ which ‘varied through time, and
through the landscape, at every hierarchical scale examined’, with inter-year, inter-
site, and inter-species variations sometimes dramatic, and substantial heterogeneity
also within a site. They suggested that such variation may have substantial
implications for the population dynamics of host plants, as well as on selection for
floral and flowering traits.
Increased threat to rare plants may also occur, as shown from a study of nectar-
robbing by short-tongued bumblebees on the narrowly endemic Polygala vayredae
(Polygalaceae) in the Spanish Pyrenees (Castro et al. 2008). Substantial reduction
in nectar supply from intensive robbing reduced that available for legitimate polli-
nators, including the long-tongued Bombus pascuorum. Non-robbed flowers had
greater reproductive outcome and, in this study, changed foraging behaviour of pol-
linators was associated clearly with nectar-robbing and decreased plant fitness.
Interactions between different visitor species on a plant may, likewise, be intri-
cate. The bumblebee Bombus occidentalis is a non-pollinating nectar robber on
Scarlet gilia (Ipomopsis aggregata, Polemoniacae) in Colorado. Ipomopsis is also
attacked by a fly, Hylemya sp. (Anthomyiidae), whose larvae feed on developing
seeds and can substantially reduce seed production (Brody et al. 2008). Experiments
that involved manipulating robbery of nectar and correlation with levels of seed
predation, as in other examples, gave very variable outcomes. However, they col-
lectively supported that seed predation was not independent of nectar robbery, with
the general principle that pattern of host use by nectar robbers can elicit changes in
subsequent use by seed predators exploiting or sharing the same plant species.
Mymecochory is not immune to cheating, by production of non-rewarding dia-
spores that chemically mimic elaiosomes and so attract ants for dispersal without
providing any reward. Chemical mimicry was demonstrated in Germany by field
tests with Formica polyctena and its responses to different diaspores, seed ‘dum-
mies’, and scent-standardised seeds (Pfeiffer et al. 2010), and those tests confirmed
that parasitic myrmecochorous dispersal of non-rewarding diaspores indeed
occurred. Removal rates were less than for rewarding diaspores, and in some cases
the chemical attractant signal may persist (rather than be lost with an eaten elaio-
some) so that additional transport may be facilitated.
9.6 Defences Against Exploiters
Protection of key resources from exploiters is a clear need for mutualisms to persist
and thrive, but the mechanisms involved have been studied only in relatively few
cases. Thus, the food bodies that nourish Pseudomyrmex ants on some myrmeco-
phytic acacias (Chap. 4) are protected in part by the balance of different protease
inhibitors – with the effect that consumption of food bodies by congeneric exploiter
ants may inhibit their digestive capacity (Orona-Tamayo et al. 2013). As well as
direct protection of food bodies by the mutualist ants, further chemical-mediated
protection can thus occur, with specific protease inhibitors and ant proteases form-
ing specific ‘lock and key systems’ that render a potential reward less attractive or
accessible to generalist potential exploiters. Nectar from extrafloral nectaries of
Acacia myrmecophytes, in a closely parallel example, may contain a soluble inver-
tase that renders the nectar free from sucrose – and so undesirable to potential
exploiters, whilst workers of the mutualist Pseudomyrmex ant lack that enzyme and
prefer the sucrose-free nectar (Kautz et al. 2009).
Specific adaptation to a partner in this way can prevent the mutualist from
becoming a cheater and helps to protect the mutualism from exploitation whilst
simultaneously enhancing the link between mutualist and plant host. In this study,
Kautz et al. found that worker ants of all three mutualist species lacked the invertase
activity and depended on the sucrose-free diet from their hosts. Phylogenetic analy-
ses of the various ant species involved confirmed that the exploiters were not cheat-
ers, but parasites of the mutualism.
Attendants of an ant-scale (or other hemipteran) mutualism can create complex
direct and indirect interactions exacerbated by the increased abundance and high
density of the Hemiptera and that can contribute to effective protection against natu-
ral enemies. Predators or parasitoids seeking ant-tended prey or hosts must have
mechanisms (variously chemical, behavioural, or morphological) to avoid ant
attacks, must be able to leave offspring sufficiently close to the ant–attended food
supply, and their eggs must be protected from destruction by the ants. Some preda-
tory Coccinellidae with those traits are thereby ‘myrmecophilous’, and selective
predators of Hemiptera. Their defences are varied. The waxy filaments produced by
larvae of Azyga orbigera, an important ladybird predator of the Coffee green scale
(Coccus viridis, Coccidae) in Mexico, are an excellent defence against the mutualist
ant Azteca instabilis (Liere and Perfecto 2008). The mandibles of attending ants
become clogged with the sticky waxes, an effective inhibitor of subsequent attacks
both by the individual ant and also other members of the same colony. Beetle larvae
from which those filaments were removed artificially in laboratory conditions were
rapidly attacked and killed.
Some other myrmecophilous ladybirds are protected by chemical traits. Further,
being able to live safely near potentially aggressive ants can give such predators
effective ‘enemy-free space’, where their parasitoids, hyperparasitoids, or other
predators may be deterred or thwarted from attack – either directly or by distur-
bances whilst searching for or handling their quarry. In Liere and Perfecto’s labora-
tory trials, decreased parasitisation rates by the encyrtid wasp Homalotylus
shuvakhinae were attributed to ant interference with searching and oviposition
behaviour of the wasps.
The status of ‘social parasite’ amongst exploiters, by which one social species
parasitises another, is a common phenomenon in Hymenoptera, and those invaders
may affect mutualisms by specific activities. Some species of socially parasitic
Megalomyrmex ants are associated with fungus-growing ants and, although the
9.6 Defences Against Exploiters 187
intensity of interaction differs across species, some appear to be obligate associates

of Attini. Thus, M. adamsae affects both parties in the mutualism between
Trachymyrmex zeteki and its fungus (Adams et al. 2012) by (1) preventing reproduc-
tion by the host reproductive ants; (2) affecting host worker ant behaviour; and (3)
reducing the colony’s resources. In young Trachymyrmex colonies, impacts on the
two parties were equally negative.
Persistence of the various associations of interacting species, including parasites,
within a mutualism was explored through a ‘metacommunity approach’ involving
two major ant species associated mutualistically with the myrmecophyte Leonardoxa
africana africana (p. 66), in which Cataulacus mckeyi is a parasite of the mutualism
between Petalomyrmex phylax and this plant in Cameroon (Debout et al. 2009).
Coexistence of these ant species was related to a suite of five different life history
traits that differed strongly in the two taxa. Those traits and the principle differences
are (1) winged individuals of Cataulacus disperse relatively further than those of
Petalomyrmex: ‘dispersal’; (2) colonies of Cataulacus reach reproductive maturity
at smaller size than those of Petalomyrmex: ‘growth’; (3) Petalomyrmex has greater
fecundity: ‘fecundity’; (4) Petalomyrmex also has greater survival rates of adult
colonies: ‘adult survival’; and (5) Petalomyrmex also has greater competitive advan-
tage through rapid worker production in incipient colonies: ‘competitive ability’.
Interactions between these are summarised in Fig. 9.7, in which the six between-
species potential ‘trade-offs’ are (1) dispersal-fecundity, (2) dispersal-adult sur-
vival, (3) dispersal-competitive ability, (4) growth-fecundity, (5) growth-adult
survival, and (6) growth-competitive ability.
The key sources of environmental heterogeneity here were tree density, tree size,
and local disturbance. The last of these explained most of the variance in ant species
frequency among patches. High disturbance rates, for example, reduced colony
lifespan. However, patterns in how the two species use a single resource were clearly
involved in maintaining the association and indicating several different ways in
which a parasite may persist within a mutualism.
Some level of incidence of parasitism within mutualisms is very widespread,
often aided by very subtle forms of signalling to promote the association. Thus,
parasitism of lycaenid-ant mutualisms by non-rewarding lycaenid species emitting
signals that mimic those of the core nectar-rewarding species have been reported.
Larvae of Lycaena (Gaeides) xanthoides in North and Central America can influ-
ence ant behaviour to increase protection from predators (Oliver and Stein 2011) in
a form of aggressive chemical mimicry, tested in trials involving the ant Formica
francoeuri, which associates with a number of rewarding lycaenid species.
Promoting the coexistence of potentially competing parasites within mutualisms
includes a series of themes of varying tractability, some very difficult to study and
interpret but collectively, as concluded by Palmer et al. (2003), offering ‘excellent
opportunities for research into mechanisms of competition and species coexistence’.
Palmer et al. listed and discussed a series of six themes that could be explored profit-
ably by focused research. These, listed below, are noted to indicate the complex
questions involved, as (1) partner choice – can, and to what extent can, mutualists
influence the identity and effectiveness of their partners in both ecological and evo-
Fig. 9.7 The dynamics of interaction and coexistence between two ant species (Cataulacus mck-
eyi, Petalomyrmex phylax) mutualistically associated with the myrmecophyte Leonardoxa afri-
cana africana in Cameroon. The two central traits (dark boxes) and other three life history traits
amongst which between-species hierarchies were found (open boxes) provide trade-offs between
these two categories, as shown by the arrows. For each trait hierarchy, the ant species with relative
advantage is shown in bold. Trade-offs, see text, are numbered as: 1, dispersal-fecundity; 2,
dispersal-adult survival; 3, dispersal-competitive ability; 4, growth-fecundity; 5, growth-adult sur-
vival; 6, growth-competitive ability. Major sources of heterogeneity in the local environment are
also shown (dotted boxes) (Debout et al. 2009)
lutionary time scales? (2) How does the effectiveness of different partners or poten-
tial partners vary, and impose pressures to select beneficial partners that reduce
associations with less beneficial partners? (3) Does the visitation rate interact with
other mechanisms of coexistence and, if so, how? For example, are competitively
inferior species more or less likely to be poor mutualists or antagonists than com-
References 189
petitively dominant species? (4) What are the spatial dynamics of mutualisms
involving multiple species? For example, are there different benefits from a partner
interacting only at the edge of its range and one overlapping broadly with the
mutualist? (5) How reliable, both in space and time, are different mutualists, and
does that affect the fitness of a partner? (6) Are alternative mechanisms for coexis-
tence affected differently by evolution of partner specificity? A consistent partner in
a mutualism in which other partners are potentially available must be the most
effective of those over time and, if abundance or availability changes, its unreliabil-
ity might increase either transiently or permanently. In Debout et al.’s (2009) discus-
sion, a single failed year for either partner of a plant-ant mutualism could doom the
other, whilst impacts of many pollinator declines may be more transient, or easily
substituted.
An allied complexity occurs when adjustments within a mutualism lead to dis-
tinction between (1) a species’ ‘proximate response’, as the short-term effect of
adding or removing a partner and defined by de Mazancourt et al. (2005) as each
partner having decreased performance as a consequence and (2) a species’ ‘ultimate
response’, as a long-term effect in which the focal species can adapt to addition or
removal of a partner. The difficulty of measuring the latter, in de Mazancourt et al.’s
opinion, renders the proximate response the only worthwhile criterion to consider
practically.
References
Adams RMM, Shah K, Antonov LD, Mueller UG (2012) Fitness consequences of nest infiltration
by the mutualist-exploiter Megalomyrmex adamsae. Ecol Entomol 37:453–462
Addicott JF, Bronstein JL, Kjellberg F (1990) Evolution of mutualistic life-cycles: yucca moths
and fig wasps. In: Gilbert F (ed) Genetics, evolution and coordination of insect life cycles.
Springer, Berlin, pp 143–161
Althoff DM, Segraves KA, Sparks JP (2004) Characterizing the interaction between the bogus
yucca moth and yuccas: do bogus yucca moths impact yucca reproductive success? Oecologia
140:321–327
Anderson B, Johnson SD (2008) The geographical mosaic of coevolution in a plant-pollinator
mutualism. Evolution 62:220–225
Anderson B, Johnson SD, Carbutt C (2005) Exploitation of a specialized mutualism by a deceptive
orchid. Am J Bot 92:1342–1349
Anstett MC, Hossaert-McKey M, Kjellberg F (1997) Figs and fig pollinators: evolutionary con-
flicts in a coevolved mutualism. Trends Ecol Evol 12:94–99
Baker HG, Cruden RW, Baker I (1971) Minor parasitism in pollination biology and its community
function: the case of Ceiba acuminata. Bioscience 21:1127–1129
Bao T, Addicott JF (1998) Cheating in mutualism: defection of Yucca baccata against its yucca
moths. Ecol Lett 1:155–159
Boucher DH, James S, Keeler KH (1982) The ecology of mutualism. Annu Rev Ecol Syst
13:315–347
Brody AK, Irwin RE, McCutcheon ML, Parsons EC (2008) Interactions between nectar robbers
and seed predators mediated by a shared host plant, Ipomopsis aggregata. Oecologia
155:75–84
Bronstein JL (2001) The exploitation of mutualisms. Ecol Lett 4:277–287
Bultman TL, Welch AM, Boning RA, Bowdish TI (2000) The cost of mutualism in a fly-fungus
interaction. Oecologia 124:85–90
Castro S, Silveira P, Navarro L (2008) Consequences of nectar robbing for the fitness of a threat-
ened plant species. Plant Ecol 199:201–208
Clement LW, Koppen SCW, Brand W, Heil M (2008) Strategies of a parasite of the ant-Acacia
mutualism. Behav Ecol Sociobiol 62:953–962
Compton SG, Holton KC, Rashbrook VK, van Noort S, Vincent SL (1991) Studies of Ceratosolen
galili, a non-pollinating agaonid fig wasp. Biotropica 23:188–194
Darwell CT, Segraves KA, Althoff DM (2016) The role of abiotic and biotic factors in determining
coexistence of multiple pollinators in the yucca-yucca moth mutualism. Ecography 39: 001-
010 (on line)
Debout GDG, Dalecky A, Ngpomi AN, McKey DB (2009) Dynamics of species coexistence:
maintenance of a plant-ant competitive metacommunity. Oikos 118:873–884
de Mazancourt C, Loreau M, Dieckmann U (2005) Understanding mutualism when there is adap-
tation to the partner. J Ecol 93:305–314
De Vries PJ, Baker I (1989) Butterfly exploitation of an ant-plant mutualism: adding insult to her-
bivory. J N Y Entomol Soc 97:332–340
Elgar MA, Nash DR, Pierce NE (2016) Eavesdropping on cooperative communication within an
ant-butterfly mutualism. Sci Nat 103:84. doi:10.1007/s00114-016-1409-5
Fedriani JM, Zywiec M, Delibes M (2012) Thieves or mutualists? Pulp feeders enhance endo-
zoochore local recruitment. Ecology 93:575–587
Frederickson ME (2013) Rethinking mutualism stability: cheaters and the evolution of sanctions.
Q Rev Biol 88:269–295
Gaskett AC (2011) Orchid pollination by sexual deception: pollinator perspectives. Biol Rev
86:33–75
Ghoul M, Griffin AS, West SA (2013) Toward an evolutionary definition of cheating. Evolution
68:318–331
Gogler J, Stokl J, Sramkova A, Twele R, Francke W (and four other authors) (2009) Menage a
trois – two endemic species of deceptive orchids and one pollinator species. Evolution
63:2222–2234
Gonzalvez FG, Chen J, Rodriguez-Girones MA (2015) The function of ant repellence by flowers:
testing the “nectar protection” and “pollinator protection” hypotheses. Evol Ecol 29:391–403
Hargreaves AL, Harder LD, Johnson SD (2009) Consumptive emasculation: the ecological and
evolutionary consequences of pollen theft. Biol Rev 84:259–276
Irwin RE, Maloof JE (2002) Variation in nectar robbing over time, space, and species. Oecologia
133:525–533
Irwin RE, Bronstein JL, Manson JS, Richardson L (2010) Nectar robbing: ecological and evolu-
tionary perspectives. Annu Rev Ecol Evol Syst 41:271–292
Jander KC (2015) Indirect mutualism: ants protect fig seeds and pollen dispersers from parasites.
Jones EI, Afkhami ME, Akcay E, Bronstein JL, Schary R (and 9 other authors) (2015) Cheaters
must prosper: reconciling theoretical and empirical perspectives on cheating in mutualism.
Ecol Lett 18:1270–1284
Jousselin E, Rasplus J-Y, Kjellberg F (2001) Shift to mutualism in parasitic lineages of the fig/fig
wasp interaction. Oikos 94:287–294
Kautz S, Lumbsch HT, Ward PS, Heil M (2009) How to prevent cheating: a digestive specialisation
ties mutualistic plant-ants to their ant-plant partners. Evolution 63:839–853
Kerdelhue C, Rossi J-P, Rasplus J-Y (2000) Comparative community ecology studies on old world
figs and fig wasps. Ecology 81:2832–2849
Liere H, Perfecto I (2008) Cheating on a mutualism: indirect benefits of ant attendance to a coc-
cidophagous coccinellid. Environ Entomol 37:143–149
Mainero JS, del Rio CM (1985) Cheating and taking advantage in mutualistic associations. In:
Boucher DH (ed) The biology of mutualism. Croom Helm, London/Sydney, pp 192–216
References 191
Maloof JE (2001) The effects of a bumble bee nectar robber on plant reproductive success and
pollinator behavior. Am J Bot 88:1960–1965
Maloof JE, Inouye DW (2000) Are nectar robbers cheaters or mutualists? Ecology 81:2651–2661
Marr DL, Brock MT, Pellmyr O (2001) Coexistence of mutualists and antagonists: exploring the
impact of cheaters on the yucca-yucca moth mutualism. Oecologia 128:454–463
Oliver JC, Stein LR (2011) Evolution of influence: signaling in a lycaenid-ant interaction. Evol
Ecol 25:1205–1216
Orona-Tamayo D, Wielsch N, Blanco-Labra A, Svatos A, Farias-Rodriguez R, Heil M (2013)
Exclusive rewards in mutualisms: ant proteases and plant protease inhibitors create a lock-key
system to protect Acacia food bodies from exploitation. Mol Ecol 22:4087–4100
Palmer TM, Stanton ML, Young TP (2003) Competition and coexistence: exploring mechanisms
that restrict and maintain diversity within mutualist guilds. Am Nat 162:S63–S69
Pellmyr O (2003) Yuccas, yucca moths, and coevolution: a review. Ann Missouri Bot Gard
90:35–55
Pellmyr O, Balcazar-Lara M, Althoff DM, Segraves KA, Leebens-Mack J (2006) Phylogeny and
life history evolution of Prodoxus yucca moths (Lepidoptera: Prodoxidae). Syst Entomol
31:1–20
Pfeiffer M, Huttenlocher H, Ayasse M (2010) Myrmecochorous plants use chemical mimicry to
cheat seed-dispersing ants. Funct Ecol 24:545–555
Powell JA (2001) Longest insect dormancy: yucca moth larvae (Lepidoptera: Prodoxidae) meta-
morphose after 20, 25 and 30 years in diapause. Ann Entomol Soc Am 94:697–680
Proffit M, Chen C, Soler C, Bessiere J-M, Schatz B, Hossaert-McKey M (2009) Can chemical
signals, responsible for mutualistc partner encounter, promote the specific exploitation of nurs-
ery pollination mutualisms? – the case of figs and fig wasps. Entomol Exp Appl 131:46–57
Ranganathan Y, Borges RM (2009) Predatory and trophobiont-tending ants respond differently to
fig and fig wasp volatiles. Anim Behav 77:1539–1545
Renner SS (2007) Rewardless flowers in the angiosperms and the role of insect cognition in their
evolution. In: Waser NM, Ollerton J (eds) Plant-pollinator interactions; from specialization to
generalization. University of Chicago Press, Chicago, pp 123–144
Richardson SC (2004) Are nectar-robbers mutualists or antagonists? Oecologia 139:246–254
Richardson LL, Bronstein JL (2012) Reproductive biology of pointleaf manzanita (Arctostaphylos
pungens) and the pollinator-nectar robber spectrum. J Pollin Ecol 9:115–123
Rodriguez-Girones MA, Santamaria L (2010) How foraging behavior and resource partitioning
can drive the evolution of flowers and the structure of pollination networks. Open Ecol
J 3:1–11
Roubik DW, Ackerman JD (1987) Long-term ecology of euglossine orchid-bees (Apidae:
Euglossini) in Panama. Oecologia 73:321–333
Sachs JL (2015) Exploitation of mutualism. In: Bronstein JL (ed) Mutualism. Oxford University
Santamaria L, Rodriguez-Girones MA (2015) Are flowers red in teeth and claw? Exploitation bar-
riers and the antagonist nature of mutualisms. Evol Ecol 29:311–322
Schatz B, Hossaert-McKey M (2003) Interactions of the ant Crematogaster scutellaris with the fig/
fig wasp mutualism. Ecol Entomol 28:359–368
Schatz B, Proffit M, Rakhi BV, Borges RM, Hossaert-McKey M (2006) Complex interactions on
fig trees: ants capturing parasitic wasps as possible indirect mutualists of the fig-fig wasp inter-
action. Oikos 113:344–352
Schatz B, Kjellberg F, Nyawa S, Hossaert-McKey M (2008) Fig wasps: a staple food for ants on
Ficus. Biotropica 40:190–195
Scopece G, Musacchio A, Widmer A, Cozzolino S (2007) Patterns of reproductive isolation in
Mediterranean deceptive orchids. Evolution 61:2623–2642
Scott PE, Buchmann SL, O’Rourke MK (1993) Evidence for mutualism between a flower-piercing
carpenter bee and ocotillo: use of pollen and nectar by nesting bees. Ecol Entomol
18:234–240
Smithson A, Gigord LDB (2003) The evolution of empty flowers revisited. Am Nat 161:537–552
Vidal MC, Sendova SF, Oliveira PS (2016) Mutualism exploitation: predatory drosophilid larvae
sugar-trap ants and jeopardize facultative ant-plant mutualism. Ecology 97:1650–1657
Wang RW, Yang Y, Wiggins NL (2014) Asymmetric or diffusive co-evolution generates meta-
populations in fig-fig wasp mutualisms. Sci China Life Sci 57:596–602
Wang Y, Wu H, Wang S (2015) Invasibility of nectar robbers in pollination-mutualisms. Appl Math
Comput 250:908–919
West SA, Herre EA (1994) The ecology of the new world fig-parasitizing wasp Idarnes and impli-
cations for the evolution of the fig-pollination mutualism. Proc R Soc Lond B 258:67–72
Yu DW (2001) Parasites of mutualisms. Biol J Linn Soc 72:529–546
Chapter 10
Habitat Losses and Conservation
of Mutualisms
Keywords Climate change • Community structure • Conservation • Disturbance •

Ecosystem services • Habitat fragmentation • Habitat management • Pollination •
Reintroduction • Translocation
10.1 Introduction: Mutualisms in Practical Conservation
Examples discussed in the previous chapter help to emphasise that knowledge of

mutualisms has two rather different roles in practical conservation, but in both those
areas of interest, the importance of fundamental understanding of the interactions’
functions can guide the purposes and practices of any needed management. Threats
to, and conservation of, mutualisms are discussed in this chapter
The fields of relevance are:
1. The conservation of ecosystem processes and ‘services’ related directly to the
restricted interactions, most notoriously pollination, and with this motivated com-
monly by needs to counter losses of native pollinators by increased use of relative
generalists. The latter are commonly widespread or deliberately introduced spe-
cies (such as the honeybee, Apis mellifera, and bumblebees, Bombus spp.) and – as
for many alien species – introducing and deliberately manipulating these raises
concerns over their potential invasions of native trophic networks and interference
with local specialised mutualisms. Plant-pollinator interactions focusing on seed
dispersal systems and pollinator visitation rates and subsequent reproductive per-
formance of the plants may provide a useful monitoring tool for some ecosystem
functions in disturbed ecosystems (Kaiser-Bunbury et al. 2009). Conservation of
native pollinators, many of them relatively specialised and involving numerous
participants in localised mutualisms, has received much attention, not least because
of perceptions of threat from introduced pollinators such as bumblebees (p. 67).
2. The conservation of individual threatened species, for which obligate mutual-
isms or related parasitism and feeding specificity are clear and intricate compo-
nents of any successful species recovery or other management plan. Whilst
lycaenid butterflies have contributed most awareness to the importance of con-
serving mutualisms, the lessons learned apply widely in considering the true
needs of any species selected for conservation attention.

DOI 10.1007/978-3-319-58292-4_10
194 10 Habitat Losses and Conservation of Mutualisms
The two themes intermingle, emphasise the importance and complexity of con-
serving ecological interactions involving insects and transcend the two major focal
levels of insect conservation – the species themselves and the wider communities in
which they dwell, in environments sustained by the ecological processes to which
the species contribute.
Changes to natural environments have many consequences, some leading to
losses or declines of native species through changing the networks of interactions
that sustain them and the structure and integrity of the wider communities in which
they live. However, many relatively generalist mutualisms may persist in novel eco-
systems even when many native species have succumbed. This persistence may take
the form of ‘by-product mutualisms’ (in which each of the partners acts to optimise
its own fitness, with a by-product of this being increased fitness of the other partner
(Leimar and Connor 2003), and in which the partners each remain non-specific).
The interaction might then persist if either (1) a subset of the remaining partners
continue to interact or (2) invading species adopt the functional roles previously
held by now-extinct or effectively displaced native species. Invasions and anthropo-
genic transfers of species inevitably create opportunities for novel interactions
(Chap. 8) and are commonly linked with changes to vegetation composition or
structure. Changed species composition may not restrict ecological function if inva-
sive or other alien species have potential to be functionally important and substitute
adequately for the native species they have usurped. Associations between ants and
epiphytic ferns in Borneo are regarded as a ‘two-way by-product mutualism’, in
which the ants live within the fern root masses and the ferns are protected from
herbivores (Fayle et al. 2015). Selective logging of the forest and conversion to oil-
palm plantations led to higher turnover of ant species, but those ants continued to
protect ferns from herbivory – and Fayle et al. found that the persistence of the
associations had no obvious shifts in benefits to the partners. Oil-palm ferns, how-
ever, supported fewer ants than ferns on native trees, and this was attributed to the
hotter, dryer microclimates of the plantations, associated with lower volumes of
occupiable ferns. Persistence of the mutualisms in plantations relied heavily on non-
native ant species, with the abundance of several ‘tramp’ ants enabling these to
adopt the roles of now-lost native species.
Many of the intricate partnerships noted in earlier chapters appear to be both
highly specific and their disruption to have immediate and wider cascade impacts
within the hosting community. Consequences of physical and biological changes to
their environments may be severe. The precise causes of many of those conse-
quences, manifesting as losses or declines of species and their associations, must
necessarily be inferred rather than proven but have led widely to recommendations
that those partnerships are a fundamental component of practical conservation.
However, as Bond (1994) commented, the by then already vast amount of literature
dealing with pollination and plant propagule dispersal contained only a small pro-
portion directly addressing reproductive failure and plant extinctions. This led him
to ask the wider question ‘Do mutualisms matter?’ In that context, of pollinator-
plant associations, the three key conditions affecting whether plant extinctions
eventuated are (1) the probability of the mutualism failing, (2) the level of the plant’s
reproductive dependence on the mutualism and (3) the importance of seeds in the
10.1 Introduction: Mutualisms in Practical Conservation 195
Table 10.1 Extinction risk of plant species in relation to pollination and seed dispersal as key
interactions that may involve mutualisms
Dependence on process
Risk of process
Rank failure Pollination Dispersal Dependence on seeds
High Single species Dioecious Dispersal obligatory Seed propagation only.
dependence self-incompatible for germination and No seedbank, sparse
reaching seedlings
Specialist Self-compatible
Generalist
Low Wind Self-pollinated Dispersal not needed Vegetative propagation.
apomicts for germination, Persistent seed bank
recruitment or dense seedlings
survival
Abbreviated after Bond (1994)
plant’s demography. Comparative ranking of these features implied that many plants
compensate for high risk in one of these categories by low risk in another.
Declines of native pollinators, a key component of the ‘global pollination crisis’,
have been detected by both decreases of species richness and the more specific
losses of ecologically specialised rare species (Ghazoul 2005), both generating
needs for wide landscape measures to foster diversity as well as finely honed species
management. In general, concerns over pollinator losses encompass three main
themes, as (1) security and future of global food production, as the topic that garners
most urgent need and political/public awareness and support; (2) the stability of
wider pollination networks in natural vegetation systems, implicitly including
mutualisms; and (3) related but more specific disruptions of intricate plant-pollinator
associations, many of them in some way mutualistic. The last of these includes two
rather different conservation themes, as summarised by New (2012), as (1) loss of
any known specific pollinators associated with plants for which no other pollinators
are known or postulated clearly implicates chances of loss of that plant and (2) the
wellbeing of wider vegetation systems may depend on relatively generalist pollina-
tors with links to numerous plant species, so that ‘diversity’ of pollinators is a key
to community persistence.
Whether mutualisms or not, many plants indeed depend on restricted sets of pol-
linators, most evidently bees, for reproduction. ‘Orchid bees’, Euglossini, for exam-
ple, are the sole pollinators of about 700 species of Neotropical orchids, but,
although the orchids depend wholly on those bees, at least some of the bees are able
to utilise and pollinate other plants (Pemberton and Wheeler 2006).
Any capability for flexibility – such as the role filled by one mutualist being
replaced from another partner should the primary partner be lost – needs to be
assessed in planning management, so that the extreme pollinator specificity of some
obligate mutualisms is indeed far rarer than plants being pollinated by several to
many species. Bond noted that plants with several mutualists may still be vulnerable
if these are closely related, but, in general, the dependence of plants on mutualisms
for reproductive success varies greatly with plant breeding systems categorised by
relative dependence on pollination for seed set (Table 10.1). Pollinator networks,
with the number and nature of possible interactions increasing with the number of
pollinator species involved (Jordano 1987), typically include (1) a small number of
highly specialised rare species, (2) many moderately specialised species and (3) a
few common generalist species (Menz et al. 2011). The last are sometimes the pre-
dominant pollinators.
In addition to changing vegetation, conservation of pollinating bees depends on
availability of suitable nesting sites. Depending on the particular taxa involved, suit-
able ground for excavating nest holes, or cavities in stems or wood may be needed,
and the latter can be supplied easily by managers as ‘bee hotels’ and the like. Whilst
focus here is on the insects, any conservation programmes for orchids or other flow-
ering plants with specific pollinators must incorporate the needs of those mutualists
effectively. For conserving wider diversity, ‘refuge areas’ of undisturbed vegeta-
tion – from which aggressive alien invaders should ideally, but usually impractica-
bly, be excluded or removed – are widely recommended. For bees in agricultural
landscapes, Banaszak (1992) recommended that at least a quarter of the landscape
should be such refuge area, with potential to approach this justified through numer-
ous ‘agrienvironment schemes’ in which the importance of pollinators in assuring
food security is acknowledged. Appreciating the roles of natural and seminatural
vegetation in providing key resources is an obvious, but sometimes overlooked,
need.
Any changes to the local biological environment in which a mutualism thrives
can strongly influence the balance between ‘costs’ and ‘benefits’, by factors such as
absolute and relative density of the partners, diversity and community composition
of associated species (such as parasitoids), the abundance of all these, seasonality
and many others that influence, for example, fitness of the host plant and foraging
tactics of ants, herbivores or other marauders. The quality and quantity of honeydew
or nectar in relation to ants’ nutritional needs, for example, may influence whether
the ants tend trophobionts such as Hemiptera, abandon and ignore them, or prey on
them (Zhou et al. 2015). Extrafloral nectar quality, notably concentrations of amino
acids and sugars, may induce preferences for ants, in turn affecting their abundance
and protective roles. Both parameters influenced preference by Oecophylla smarag-
dina, for example (Bluthgen and Fiedler 2004), whilst laboratory bioassays revealed
a range of ant species to consistently prefer more concentrated sugar solutions over
more dilute ones, and with the frequent ranking preferences of ‘sucrose > glucose >
fructose’ as key components. Interspecific variability in preferences amongst sugar
solutions with different amino acids was common, but most ants tested still pre-
ferred sugar solutions which contained amino acids over sugar solutions alone.
Relatively small changes to the abiotic environment can influence the intensity of
an ant-plant-herbivore mutualism, as Kersch and Fonseca (2005) demonstrated for
the interaction between the Neotropical Inga vera (Mimosaceae) and its associated
ant species, notably Camponotus fastigatus and Crematogaster quadriformis.
Experimental comparisons showed that (1) sunny plants had stronger mutualistic
relationships, whilst (2) shade eliminated the protective benefit of the ants, with the
ants becoming commensal. Inga plants from which ants were excluded were shorter
and had far less foliage than those with ants, so that the protective benefits from
10.1 Introduction: Mutualisms in Practical Conservation 197
ants – here against the depredations of the thrips Selenothrips rubrocinctus, which
showed clear preference for plants growing in the shade – were very obvious.
In general, interactions may be more susceptible to imposed changes than are
many individual taxa. Interpretations are rendered both difficult and uncertain
because the various drivers of environmental change occur in concert and the rela-
tive effects of, for example, climate change, resource loss and fragmentation and
changing species community composition on any interaction are usually unclear.
The main drivers of ecological change noted by Tylianakis et al. (2008), namely,
increased atmospheric CO2 levels and associated climate changes, deposition of
nitrogen from anthropogenic activities, loss and fragmentation of natural habitats
and biotic invasions, all cause reduced interaction strengths for mutualisms (polli-
nation, seed dispersal) involving plants. Those impacts are in addition to those on
other interactions.
The ‘community context’ of the interaction can change rapidly, and an apparent
mutualism may be transformed to commensalism or, even, antagonism. The moth
Greya politella (Prodoxidae, one of 16–20 species in the genus and closely related
to the yucca moths, p. 50) is generally host specific to the Prairie star, Lithophragma
parviflorum (Saxifragaceae), in western North America. The moths are passive pol-
linators and oviposit within the flowers, where larvae feed on the developing seeds
(Thompson and Fernandez 2006). However, the outcome of this interaction varies
in different locations. Whilst the moths depend on the single host species, the plant
populations vary in the extent to which each depends on Greya for pollination.
Many parallels have been inferred.
Preliminary investigations in Canada (Taki and Kevan 2007) indicated that
insects associated with plant-pollination interactions might be more vulnerable to
habitat loss than are their partner plants, possibly implying that those insects might
be more specialised than the plants. Separate analyses of insects and plants exempli-
fied asymmetric interaction at the wider community level. The extent of a species’
relative specialisation/generalisation was assessed by using the number of its part-
ner species – but Taki and Kevan noted the bias due to low frequency of observing
specialist visits and interactions and that bias must be taken into account in such
appraisals.
In general, many insect conservation programmes have tended to underestimate
the importance of mutualisms, but understanding of these associations is needed at
all levels – in programmes spanning species to communities and landscapes in
which the interactions occur.
This need reflects that mutualisms are widespread – as Vasquez et al. (2009) put
it, ‘Mutualisms between plants and animals pervade nature’. Many of the examples
discussed earlier also endorse the comment that ‘Mutualisms clearly are not isolated
from the community in which they occur’ (Bronstein et al. 2003), with the outcomes
of paired species associations influenced strongly by factors well beyond that imme-
diate interaction. The factors influencing mutualisms, and the roles of mutualisms in
wider conservation management issues, are both complex and difficult to address
fully. Bronstein et al. (2004) commented that ‘our understanding of how they [mutu-
alisms] might respond evolutionarily to anthropogenic change remains rudimen-
tary’, and regarded this situation as ‘alarming’. The interactions that occur between
species, including plant-pollinator, ant-butterfly and in many other mutualisms, are
frequently not ‘symmetrical’ in that – other than the relatively few obligate mutual-
isms whose study has contributed so much to how those interactions evolved and are
sustained – the level of specialisation or generalisation differs between the partners.
In these contexts, either the plant or the pollinator – or either the ant or the butterfly -
may be more specialised (or more generalised) than its partner, so that the relative
vulnerability of each to changes in their natural communities or local environments
may differ. Responses to habitat fragmentation, for example, may differ according
to impacts on the partners – any changes in pollinator species abundance, composi-
tion and foraging activity due to patch dispersion and size changes may have impli-
cations for mutualist plant fitness (Ashworth et al. 2004). Widespread concerns over
pollinator declines include that some appear to be especially sensitive to habitat
fragmentation, a trend commonly considered to be more severe for specialist polli-
nators than for generalists that can exploit a wide range of plants. However, the
extent to which changes in pollinator or other mutualist assemblages affect any
plants reflects the plant’s dependence on the mutualism. Testing the various hypoth-
eses on how disturbances can be predicted to affect mutualisms would require sev-
eral key aspects that, from Prinzing et al.’s (2007) survey, had not then been
integrated into any single study of impacts. They noted needs for (1) a clearly
defined disturbance gradient that can be replicated, and encompassing the levels of
natural disturbances, (2) information on the richness and abundance of both partners
in the mutualism and (3) information on the parallel changes of non-mutualist mem-
bers of the community, extending to interactions between guilds of mutualists – the
last of these recognising that many mutualists are facultative or diffuse rather than
obligatorily specific.
10.2 Mutualisms and Disturbance
As one important context, disturbance to tropical forests by clearing, fire, conver-

sion to shifting agriculture or other development clearly causes massive changes
and losses to endemic and wider native diversity and community structures.
Mutualisms are no exception, although only rarely highlighted amidst such changes
that, in principle, apply to almost all natural vegetation regimes. The considerable
variety of ant-plant mutualisms in lowland forest in Borneo, for example, is ‘drasti-
cally reduced’ by slash-and-burn deforestation (Tanaka et al. 2007). Frequency and
composition of mutualisms (hemipteran-ant-plant) were compared across forest
stands of different ages, as ‘time since last slash-and-burn’. Plots of primary forest
in the Lambir Hills National Park, Sarawak, were compared with secondary plots of
1 year, 5 years and >19 years old. Collectively, 17 species of trees with extrafloral
nectaries yielded 31 ant species by direct census from collecting them from the
lower foliage layers. More widely, 28 ant species were associated with hemipterans
across at least 23 tree species. Drastic declines in diversity of species involved in
10.2 Mutualisms and Disturbance 199
Fig. 10.1 Specialisations
in plant-pollinator
interactions. Two scenarios
are shown for each of
(upper, a, b) symmetrical
and (lower, c, d)
asymmetrical interactions.
In symmetrical
specialisations, (a)
generalist plants are
pollinated by many
generalist pollinators (G)
and specialist plants by one
or few specialist
pollinators (S) – in that
case, habitat fragmentation
will more strongly affect
specialist plants compared
with generalist plants (b);
under asymmetrical
specialisation (c) generalist
plants are pollinated by
species of both categories
(G, S) and specialist plants
mostly by one or few
generalist pollinators – in
this case susceptibility to
habitat fragmentation is
similar for both plant
categories (d) (Ashworth
et al. 2004)
mutualisms were observed, together with reduced frequency of myrmecophytic

trees in secondary forest. The accumulated data suggested that impacts of deforesta-
tion could persist for at least 20 years.
Ashworth et al. (2004) considered aspects of the level of specialisation in such
interactions for intact and fragmented habitats, devising the scheme summarised in
Fig. 10.1. In symmetrical interactions many different generalist pollinators associ-
ate with generalist plants, and few specialist pollinators pollinate specialist plants.
Habitat fragmentation is likely to more strongly and adversely affect the specialist
plants as their pollinators are progressively lost. Under asymmetrical conditions, a
generalist plant may be pollinated by an array comprising both generalist and spe-
cialist pollinators and a specialist plant by one or few such generalists. In this con-
text, generalist and specialist plants are influenced in similar ways in fragmented
habitats. This scheme brings out the scenario validated by others studying plant-
pollinator networks, to demonstrate that most mutualism webs are indeed highly
asymmetrical, especially in that many specialist plants are pollinated mostly by
generalist pollinators and generalist plants by a combination of specialists and gen-
eralists. The susceptibility of a plant to habitat fragmentation may depend heavily
on the specificity of its pollinator(s); if even a generalist pollinator has broad eco-
logical tolerances, this may help to buffer impacts of fragmentation.
Forest fragmentation can dramatically affect the community structure amongst
mutualistic assemblages such as figs and their attendant wasps. In China, Ficus
racemosa (p. 167) had lower abundance of pollinators in fragmented areas than in
primary forest (Wang and Sun 2009), with differences in interactions also influ-
enced by seasonal changes. Following Bronstein (1989), difficulties for pollinators
arise when (1) fig trees flower only rarely in winter and (2) those trees that do flower
are less easily detectable and more difficult to reach during winter and the dry sea-
son. Fig syconia take longer to develop in the dry season, providing longer periods
for non-pollinators to oviposit – so that increased abundance of non-pollinating
wasps over that time may exacerbate pollinator supply and create a ‘more fragile’
interaction between figs and pollinators that might be especially severe in frag-
mented forest patches (Wang and Sun 2009). More broadly, variations in the fig-fig
wasp mutualism over the geographical range of the interactions can thus make it
more difficult for pollinator populations to persist. Bronstein (1989) emphasised
that geographical comparisons of such systems have considerable value, although
they are often only correlative studies, and can add enormously to understanding
how the interactions have evolved.
A widespread consequence of habitat patch fragmentation is the increase of
‘edge effects’, widely acknowledged but only rarely investigated specifically in
relation to their impacts on mutualisms. Those impacts could include altering pat-
terns of herbivory or the levels of protection of plants that can occur. Evidence
reviewed by Evans et al. (2013), together with further experimental study, led them
to suggest (p 1744) that ‘edge effects on the distribution and encounter frequencies
of honeydew-producing insects, ants, and other herbivores may determine levels of
herbivory and the amount of protection ants provide plants in fragmented habitats’.
Their investigation was of the myrmecophyte Solanum americanum (Solanaceae)
that hosts aphids and so attracts a variety of ants on which the primary herbivores
are grasshoppers. Temperature-correlated grasshopper activity typically renders
these insects more abundant in the centre of patches and less abundant along more
exposed edges. Manipulative trials, including removals of ants and aphids from
plants introduced into field plots at different distances from the patch edge, were
undertaken in conjunction with direct counts of aphids and ants on all plants (paired
comparisons of insecticide-sprayed and water-sprayed plants) and visual surveys
for grasshoppers as well as direct assessments of herbivory. Major conclusions after
8 weeks were as follows: (1) edge effects on herbivore abundance can strongly
influence the level of herbivory and the extent of protection given by ants, but also
(2) the mechanisms for this can be complex. They may, for example, reflect that it
10.2 Mutualisms and Disturbance 201
is only far from edges that ant protection is high because recruitment of ants in
response to increased herbivory is greatest there: the number of ants per plant, a key
aspect of protection level, may be influenced by trophobionts but may also be highly
variable.
Most studies on habitat fragmentation and mutualisms have focussed on plant-
pollinator interactions, with the underlying premise that disruption of mutualisms
could lead to disruption and possible local extinctions of intertwined partners. As
Bruna et al. (2005) noted, however, few of those studies assessed simultaneously the
impacts of fragmentation on both plant and pollinator populations, so that the pre-
cise causes and extents of disruptions remain unclear. The extremely low incidence
of hawk moth pollination of Oxyanthus in South Africa (Johnson et al. 2004, and
noted on p. 41), for example, was attributed largely to habitat fragmentation and the
consequent disruption of the pollination mutualism. Oxyanthus pyriformis pyrifor-
mis is a very rare tree, now confined to two small habitat patches isolated by urban-
isation near Durban, and losses of regular pollinators may well be hastening the
tree’s extinction. Whilst the very low pollination rates recorded could be natural,
impacts of urbanisation have probably accelerated decline, and Johnson et al. noted
the possible uses of hand-pollination and outplantings of saplings as measures that
could perhaps help to counter that trend.
More generally, plant-pollinator networks can be compared between ‘reference’
and restoration-managed sites as a measure of success of restoring ecological func-
tion. Comparison of ancient and restored (after 11 or 14 years from starting man-
agement) heathland sites in southern England showed that the major pollinators
found on restored sites were widespread generalists (Forup et al. 2008). The net-
works were also significantly less complex on the restored sites, which were
regarded as likely to be more susceptible to further disturbances than the reference
sites. The focus on pollination services in this example demonstrated the values of
common insect species, but more specialised or obligate mutualist interactions were
not addressed.
Numerous plant-pollinator interactions exemplify the even wider range of mutu-
alisms in which the partners have very different mobility, so implying that dispersal
capability of any mobile partner and likely to be affected by processes such as land-
scape change and habitat fragmentation decreasing connectivity may influence
regional persistence of the association. Discussed by Amarasekare (2004), species
that have their local reproduction unaffected by dispersal may benefit from manage-
ment that includes linking habitat fragments (such as by corridors) to endorse both
local and regional persistence. In contrast, if local reproduction is reduced by dis-
persal, dispersal rates should be sufficiently low to sustain the growth rate of source
communities and not increase their vulnerability by any losses. Higher risks of
larger-scale extinctions may arise as habitat destruction induces mass dispersal, and
source communities are lost. From Amarasekare’s models, preservation of large
habitat fragments, irrespective of connectivity, may be a better option for preserving
mutualisms. Spatial dynamics can thus hinder or enhance persistence of mutual-
isms, depending on effects of dispersal on local source communities.
Comparison of ant-partner communities in intact forest and forest fragments in

Brazil incorporated 12 plant taxa and 33 ant morphospecies. The ants comprised 11
obligate associates and 22 opportunistic species on the myrmecophytes, and, as
elsewhere, many of the opportunists provide rather little defence against herbivores.
Despite most of the mutualists being rare, with only small populations in fragments,
the associations seemed likely to persist. Densities of ant plants in fragments iso-
lated for 20–25 years were only about half those in continuous forest, but only lim-
ited direct effects on partner species richness were found (Bruna et al. 2005), and
vulnerability is due more to rarity of many myrmecophytes than to direct loss of the
mutualistic ants.
The five European species of large blue butterflies (Maculinea spp.) have been
described as ‘among the most endangered invertebrates in the world’ (Hochberg
et al. 1992) (p. 124). Conservation management has depended on knowledge of
their mutualisms or parasitic/predatory interactions with Myrmica ants, lessons
from which have contributed enormously to more general awareness of ecological
specialisations amongst such associations and the wider practical importance of
elucidating these and integrating their needs and relationships to habitat condition
and extent. Local or national losses of these butterflies have been frequent in Europe,
and the extinction of M. arion in Britain in 1979, followed by its successful reintro-
duction from Swedish stock (Thomas et al. 2009), is a key success story in insect
conservation. A central contribution to that success was the discovery that grazing
regimes on the patches of chalk grassland supporting the larval food plant (Thymus
praecox) of M. arion severely affected the wellbeing of the predominant host ant,
Myrmica sabuleti. The ant thrived only on closely grazed grassland with the warm-
est ground temperatures. Slight reductions of grazing led to its replacement by the
less suitable M. scabrinodis, and cessation of grazing (as a well-intentioned site
preservation strategy undertaken before the microclimate needs of the host ants
became clear) eliminated both of these ants. Thomas (1980) noted that this vital
information came too late to save the British large blue from extinction. More
widely, conservation plans for these butterflies (such as that by Munguira and
Martin 1999) emphasise needs to understand myrmecophilous relationships and the
factors that facilitate optimal associations to predominate. If ants are rare, it may
become difficult for ovipositing butterflies to find them and increase chances of
larval survival by laying close to the ant colonies. Plans to conserve myrmecophi-
lous butterflies must incorporate measures to sustain all three members of the mutu-
alistic partnerships (butterfly, food plant, host ant). The programmes are thus
inevitably more complex than for the ‘simpler’ normal bipartite relationship of but-
terfly and larval food plant alone. Both categories, of course, also need to incorpo-
rate adult food needs, usually as nectar.
10.3 Community Contexts 203
10.3 Community Contexts
Although studies of pairwise interactions between species are attractive because of

their apparent simplicity, to ignore the wider community context is, again in the
words of Bronstein et al. (2003), ‘unrealistic at best and misleading at worst’. They
listed major complications as:
1. Each species in a mutualism interacts with a suite of co-mutualists, with interac-
tions between those altering pairwise outcomes.
2. Some associations are only beneficial when another, non-mutualistic species is
also present – for example, in protection mutualisms, the benefits to the pro-
tected species occur only when natural enemies are present: at other times, the
‘protector’ may have neutral or negative effects.
3. Natural enemies and other non-mutualistic species can change the success of one
mutualistic partner and so, potentially, also the other.
4. The presence or absence of species involved with one of the partners can alter the
success of the mutualism.
5. Some of these, as ‘exploiters’, obtain the rewards offered by one mutualistic
partner to the other whilst giving nothing in return.
The long-held traditional view that many mutualisms comprise tightly coevolved
interactions that benefit both parties has been challenged repeatedly by studies that
demonstrate how the interactions may be influenced by many external factors – so
that the outcome may be beneficial, neutral or harmful to one or other partner. As
above, this variety causes conservation concerns in posing problems for predicting
impacts of environmental changes. Indeed, as Stanton (2003) noted, highly specific
passive mutualisms are relatively rare. More diffuse mutualistic interactions reflect
quality of alternative partners, for example, and such alternatives may each both
influence and be influenced by the selection of (or by) partners with differing levels
of ‘fitness’. Those trends may be revealed, at least tentatively, by exploring interac-
tion networks, as Stanton showed.
Whilst losses and changes to the embedding environment are paramount in many
conservation evaluations, other threats are commonly interwoven with habitat qual-
ity issues and influence the management needed. Thus, orchids, as a highly valued
flagship group of flowers, are vulnerable to excessive commercial collecting as well
as widespread loss of habitats. Their conservation planning necessitates knowledge
of their pollination syndromes, which are extremely varied, and pollinator specific-
ity. One example is Habenaria radiata in eastern Asia and Japan (Ikouchi et al.
2015), for which the diurnal hesperiid butterfly Pelopidas mathias was the only
insect visitor detected during surveys. It can be an effective pollinator and seems to
be at least a predominant pollinator – but the study also raised an uncertainty that is
not uncommon in assessments of pollination, namely, the lack of nocturnal observa-
tions that could not exclude the possibility of nocturnal moths, recorded as pollina-
tors of some congeneric orchids, also contributing. Nevertheless, conservation of
Table 10.2 Possible changes in plant-moth pollination networks that can result from artificial
lighting and predicted effects for local flower visitation activity
Effect Potential impacts on flower visitation
No lighting (None)
Concentration effect – moths attracted to lit Increased moth abundance in lit region;
region decreased moth abundance in unlit region
Ecological trap effect – moths attracted up to Increased moth activity but decreased flower-
lights visiting activity in lit region; decreased moth
abundance in unlit region
Disruption effect - normal behaviour Decreased flower-visiting activity in lit region;
disrupted around lights no effect in unlit region
Preferential disruption effect – behaviour of Decreased flower-visiting activity by some
particular subset/s of moths disrupted around species in lit region; no effect in unlit region
lights
After Macgregor et al. (2015)
the known orchid pollinator was recommended, with measures to also conserve the
surrounding wetlands that host the larval food plants of Pelopidas.
Concerns over declines of many moths (such as documented convincingly in
Britain by Conrad et al. 2004 and Fox 2013) are linked with their roles as predomi-
nant nocturnal pollinators. Proliferation of artificial lighting (‘light pollution’) may
be an important contributor to pollination declines (Macgregor et al. 2015). That
review implied that pollination could become seriously disrupted by the impacts of
increasing light pollution on moths, with potential changes in moth assemblages
inferring the kind and intensity of pollinator-plant interactions. Several different
scenarios arise (Table 10.2).
10.4 Fragmentation and Loss of Habitats
This universal threat and concern in insect conservation affects all partners in mutu-
alisms and all members of natural communities, as critical resources disappear or
their supply becomes less assured, and many aspects of the physical and biotic
environments change. Population structures and normal dispersal patterns may be
disrupted severely and ecological and evolutionary consequences for taxa and eco-
logical processes placed at risk. Modelled decline of mutualistic networks with
habitat loss (Fortuna and Bascompte 2006) implied strongly that structure affects
rates of loss. Commonly, some specialists are the first to become extinct, whilst the
central ‘core’ of the most generalised species are the last to succumb, and are more
usually resilient to habitat loss. Because those species, commonly the most con-
spicuous and obvious to observers, persist, earlier losses of specialists may not be
detected until ecological impacts occur, as implied for plant-pollinator networks
and habitat fragmentation (Harris and Johnson 2004).
10.4 Fragmentation and Loss of Habitats 205
Amongst mutualisms, perhaps pollination systems have received most attention

from habitat changes at local and landscape scales. Many angiosperms threatened
with extinction occur in fragmented or highly restricted biotopes and in increasingly
small and scattered populations. These features can amplify pollination failure
through reduced numbers of pollinators and the increased complexity of their forag-
ing environments. Plants without any ‘reserve capability’ (such as by self-pollination
or asexual reproduction) may be especially vulnerable and, as a widespread recom-
mendation, conservation must consider the wellbeing of key or specific pollinators
as well as of the plant’s environment. However, and as Menz et al. (2011) empha-
sised, ‘ecological restoration of plant-pollinator interactions has received surpris-
ingly little attention’. Restoring these in degraded environments benefits both
individual plant species conservation and the wider resilience of plant communities.
Restoration projects can have markedly differing objectives (Menz et al. 2011) –
restoration of the original pollinator community and their key plants may be very
different from, and far more complex than, restoring functional pollinator groups to
enable persistence of broad native plant communities. Pollinator assemblages often
comprise three major functional components: (1) a few abundant generalist species,
(2) many moderately specialised species and (3) a few highly specialised and usu-
ally rare species. The last of these is commonly signalled as having conservation
priority, reflecting the predominance of individual ‘threatened species’ in designat-
ing those priorities. Restoration must provide the critical resources needed by pol-
linators either ‘on site’ or within realistic foraging distances. As well as nesting sites
(for bees), nectar and honeydew are often key needs and in restoration are com-
monly provided by establishments of ‘framework plantings’ of species used specifi-
cally to provide these resources, with considerations of using different plant species
to complement the pollinator array. They may include ‘bridging plants’ that provide
nectar and pollen at times they are not otherwise available easily and are especially
important for specialised pollinators with otherwise restricted food supply. Creation
of ‘stepping stones’ of restored plant resources may also provide connectivity and
reduce fragmentation effects on remnant plant populations in the wider landscape.
Recent declines of many native pollinators have led to global concerns over how
this vital process can be sustained. Much pollination of crops relies heavily on a few
species of domesticated bees, most notably Apis mellifera and a few species of
Bombus bumblebees. Continued range expansions and international translocations
of these to accomplish pollination raise worldwide concerns for native species as
these generalist intruders invade more natural environments and encounter native
pollinators involved in mutualisms with local and endemic flora. Many authors have
discussed, or alluded to, impacts on specialised mutualisms in which usurpation of
specialist pollinators by invasive generalists raises concerns for plant conservation.
Other concerns, such as hybridisation between introduced and native populations
(summary in New 2012), may also affect mutualisms through changing capability
or resilience of partners.
Much has been written about declines in pollination services, but the patterns of
change in pollinator assemblages and mutualisms have received far less attention
and are difficult to generalise. Thus, comparisons of bee and hoverfly assemblage
changes in the Netherlands and Britain gave rather different implications for the two
groups (Biesmeijer et al. 2006). Examining incidences of bees (ca one million
records) and hoverflies in 10 × 10 Km grid cells for both regions showed similar
trends of declining bee richness (comparing data for before and after 1980), but
those trends were more varied for Syrphidae. In both groups, species with narrow
habitat needs had declined most, together with univoltine species, with clear ten-
dency for specialists to be the most generally susceptible.
That widespread trend, of specialists declining whilst generalists prosper, leads
to two major concerns: (1) losses of crop yields and its implications for human food
supply and (2) losses of species that play critical roles in natural ecosystems, many
as specialists associated with equally restricted plant species and which are collec-
tively critical in sustaining wild plant diversity. As Senapathi et al. (2015) noted,
management for conservation of ‘pollination services’ thus has very different
approaches from managing for natural pollinator diversity, not least because it can
emphasise the wellbeing of widespread polyphagous generalists and diminish the
significance of the far more numerous ecological specialists. The contrast in man-
agement approaches is reflected in the trade-offs that can occur between these
(Fig. 10.2), with awareness that most pollination mutualisms would fall to the right-
hand side of this diagram, in conjunction with most rare or locally endemic taxa.
Senapathi et al. also noted that many of the latter groups play no direct roles in crop
pollination but may become susceptible to competition, reflecting the specialised
nature of the insects involved.
Specialisation and mutualism are, of course, not an ‘automatic’ correlation but
may often appear to be so – or very closely linked. Careful study may be necessary
to confirm that mutualism does not occur in an association. Senapathi et al. (2015)
cited the European subalpine bumble bee Bombus (Megabombus) gerstaeckeri that
forages only on Aconitum spp. (monkshoods) and is one of the longest-tongued spe-
cies of this subgenus, in which most species are extreme specialists (Huang et al.
2015). Virtually all such species have no economic values, and an important infer-
ence from Senapathi et al.’s essay is that management practices that aim to conserve
only a limited number of economically significant pollinator species also need to
consider the array of ecological partners needed for these and the wider diversity of
species not directly targeted by those endeavours. Despite the occurrence of special-
ised pollination mutualisms, the great majority of pollinator-plant associations are
more generalised, and wider webs of more complex interactions prevail (Memmott
1999).
Reversal of pollinator declines has depended largely on restoration and enhance-
ment of floral resources within vegetational diversity, largely following principles
pioneered for use within agricultural landscapes. Measures such as planting along
field edges, and allied management to reduce loss and fragmentation of natural veg-
etation and to promote connectivity, are widespread, but tailoring plant mixes for
optimal combinations of pollinator preferences and phenological variety of species
can become complex. As M’Gonigle et al. (2016) noted, ‘selecting a subset is
challenging’.
10.4 Fragmentation and Loss of Habitats 207
Fig. 10.2 A scheme illustrating examples of trade-offs between management for pollinator con-
servation and management for pollination services; a, b, c, d each represent a particular scenario of
(a) low diversity/low services (‘loss-loss’), (b) low diversity/high services (‘loss-win’), (c) high
diversity/low service (‘win-loss’) and (d) high diversity/high service (‘win-win’). The arrows show
pathways and management measures that facilitate movements from one scenario to another. The
curves represent possible expansion of the outcomes for successful management toward both pro-
ductivity and conservation (Senapathi et al. 2015)
Recognition that native bees and other pollinators can enhance crop pollination
whether or not honeybees are present has led to affirmations of needs for integrated
crop management measures that include conservation of native pollinators. Many
wider measures apply easily to those species involved in mutualisms, as well as to
the more generalised species that may predominate in more economically applied
contexts. Those measures (Garibaldi et al. 2013) include (1) restoration and conser-
vation of natural vegetation within and near cropping areas, (2) addition of diverse
nesting and floral resources, (3) promoting mosaic land use patterns to increase
patchiness and (4) assuring pollinator protection from pesticide applications.
‘Pollination failure’ involves a number of different processes (Wilcock and

Neyland 2002) over the sequence of pre-dispersal pollen consumption, pollinator
restrictions, and post-dispersal failures and incompatibilities. Specialist pollen vec-
tors in restricted mutualisms (such as amongst yuccas and figs, Chap. 4) represent
the highest risk of failure from pollinator losses. They contrast with those plants
with a limited range of (often related) pollinators and intermediate risk of failure
and plants with generalist pollinators undergoing only much lower risks.
Nevertheless, risk of failure is likely to increase with some changes in environmen-
tal conditions (such as cooler temperatures reducing pollinator activity) as well as
by their direct losses.
10.5 Climate Change
Studies on plant-pollinator mutualisms, in particular, have indicated some of the

likely more general effects of climate changes, most notably of increased tempera-
tures affecting insect and plant life histories and seasonality, as well as inducing
distributional changes. Evidence of impacts of climate changes on plant-pollinator
interactions is accumulating through changes in flowering phenology and the earlier
dates of appearance and changed activity regimes of butterflies and other pollinators
(Hegland et al. 2009). Distributions of both parties are also influenced – with many
insects extending ranges or moving to higher latitudes and elevations as warming
allows. Many possible outcomes are possible for mutualists (Fig. 10.3), but in gen-
eral, knowledge of the effects of climate warming on mutualisms is still highly
incomplete. Nevertheless, it seems inevitable – albeit speculatively in many cases –
that the parties in many such associations will be differentially affected by changes
in temperature and humidity regimes, as well as by changes in their encompassing
communities and landscapes.
Future projections remain largely speculative even though some general trends
have been indicated strongly over recent decades. Flowering duration, as well as
phenology, may be affected, but the complex influences of climate on such param-
eters render specific predictions very uncertain. Abundance and distribution of flow-
ers are affected also by edaphic factors, precipitation and humidity, amongst other
factors, and any of these may in turn affect availability of pollinators. Studies of
relevant groups of insects along latitudinal and elevational gradients over time can
perhaps help in forecasting the assemblage compositions that will remain possible,
or the most likely changes in assemblages. Some critically important species may
be lost. As Fig. 10.3 demonstrates, spatial and temporal mismatches can influence
availability of mutualist partners – in particular, some specialist pollinators may be
left with diminished or no foods following phenological shifts, whilst opportunist
generalists may benefit from increased food availability. Demographic conse-
quences of mismatches are a significant concern (Hegland et al. 2009), with altered
supply of nectar (carbohydrate) and pollen (protein) influencing pollinator survival
and reproduction. Nevertheless, studies on the broad structure of pollination net-
10.5 Climate Change 209
Fig. 10.3 Some effects of climate warming on the phenology and distribution of plants (left) and
pollinators (right) in creating mismatches in time and space and disrupting mutualisms. The lower
box on each side indicates some of the key factors affecting the demography of the partners, with
the subsequent effects of the mismatches largely unknown (Hegland et al. 2009)
works suggested that they may be resilient to climate warming. Because both pol-
linator activity and flowering appear to be influenced by temperature, their responses
may – despite individual variations and rates of changes – occur in parallel and
partially counter some potential disruptions.
The roles of keystone mutualists may change with elevation, probably related to
the gradient of differences in climatic factors, notably temperature and rainfall, and
indicate a pattern possibly labile to future climate changes. In New South Wales,
Australia, the myrmecochorous relationships of Acacia terminalis differ at higher
and lower elevations (Thomson et al. 2016). The major vector groups differed
(lower, ca 200 m a.s.l, ants; higher, ca 700 m a.s.l., vertebrates), but the keystone ant
genera also changed from Rhytidoponera (lower elevations) to Aphaenogaster
(higher), collectively resulting in different dispersal patterns for the acacia at
different elevations. Both ant genera are functional opportunists characteristic of
disturbed habitats. More generally, climatic changes might alter myrmecochorous
interactions along elevational or other gradients and lead to implications for chang-
ing vegetation structure.
Differential effects of temperature change on the partner species may have com-
plex outcomes, especially so in interacting with other disturbances such as habitat
loss. In some cases, insect mutualisms may even buffer the effects of warming on
specific herbivores, and when disrupted, the direct effects of warming on predators
and generalist herbivores can have strong indirect effects on plants (Marquis et al.
2014). Manipulations of temperature and the presence of aphids (Chaitophorus
populicola) on Populus tremuloides in Massachusetts revealed that aphid persis-
tence was higher when they were tended by ants, whether or not warming occurred,
whilst the continuation of the mutualism led to decreased stress on the plants. Three
ant species (Camponotus pennsylvanicus, Formica subsericea, Lasius umbratus)
were involved, with somewhat different outcomes. Aphid colonies tended by F. sub-
sericea persisted for considerably longer than when other ants were present, and
Marquis et al. suggested that such differences were likely to be magnified under
warmer conditions. More broadly, temperature affects the strength of many trophic
interactions - for example, effects on arthropod predators may have significant flow-
on consequences for herbivore prey. Published accounts variously report elevated or
reduced impacts, as well as increased cascade effects on plant communities (Drieu
and Rusch 2016), not least through increased combined antagonistic effects of mul-
tiple predator species. These trends urged the need to conserve species-rich assem-
blages of higher level consumers to buffer some effects of climate warming.
The difficulties of attributing changes to climate or other causes have been illus-
trated for the mutualisms between Maculinea butterflies and Myrmica host ants (pp.
120–125), amongst which discordant responses can occur. Filz and Schmitt (2015)
emphasised the importance of habitat loss as a major threat to the butterflies and, in
conjunction with range shifts induced by climate change, could lead to spatial
restriction of the species by restricted availability of their ant host. Flexibility of
host choice might enable the butterfly to shift amongst hosts, leading to co-
occurrence of several Maculinea species and maintenance of viable populations. In
that context, the extent of harmful impacts from climate change depends on the level
of specificity involved: amongst Maculinea species, each large blue clearly has a
single preferred host ant (p. 126), but adoption of others is believed widely to be
much more flexible than earlier presumed. Development of models implied that the
butterflies might be able to occupy larger ranges in Eurasia based on their climate
tolerances, but each is spatially limited by availability of their host ants.
In any obligatory one-to-one mutualism, persistence clearly depends on both
partners surviving climatic or other changes – loss of either one essentially means
loss of the mutualism. For fig-fig wasp mutualisms, thermal tolerances of female fig
wasps of four species from Singapore were assessed to demonstrate that a rise of
only 3 °C (as projected to occur by 2080–2099 by some models) could markedly
lessen their already short lifespan (Jevanandam et al. 2013). Wasps kept in a growth
chamber at a range of temperatures clearly succumbed more rapidly as temperatures
were increased (Fig. 10.4). In nature, this would markedly reduce the searching time
available to locate suitable syconia and for pollination or oviposition, with likely
consequences including (1) declines of fig wasp populations, (2) reduced pollina-
tion levels that lead to declines of fig populations and (3) reduced fig production
associated with decreased food supply for a range of frugivores. Such projections do
not, of course, incorporate or reflect any counter-adaptations by the wasps to gradu-
Fig. 10.4 The median

lifespan, from emergence,
of four species of fig wasps
from Singapore maintained
at different temperatures
until all had died. Mean
daily temperature is shown
as baseline information.
The species are Valisia
malayana (solid circles),
Ceratosolen constrictus
hewitti (open circles), C.
appendiculatus (solid
triangles) and Eupristina
verticillata (open
triangles) (Jevanandam
et al. 2013)
ally increasing temperatures and that would decrease or eliminate severity of the
outcomes – but such adaptations cannot be universally assumed to occur. Jevanandam
et al. noted that species with wider latitudinal ranges may have historically experi-
enced wider thermal extremes than those species wholly restricted to the more asea-
sonal tropics. Likewise, variations amongst syconia on any single fig tree may
provide some opportunities for thermal acclimation, and this possibility is favoured
by the wasps’ very short generation times. Nevertheless, it seemed that many fig
populations within the tropics could indeed become vulnerable to global warming,
with cascade impacts from the mutualism breakdowns posing serious conservation
concerns.
In general, and reflecting the presumed high levels of specificity involved (p. 48),
many fig wasps are presumed or expected to span the entire range of their host
Ficus. Because adult fig wasps do not feed, their starvation tolerance may also limit
longevity and function – this, also, may mirror local humidity and temperature
regimes, as shown by comparison of two species of Ceratosolen associated with
Ficus sycomorus in Africa (Warren et al. 2010). C. arabicus survives for longer than
C. galili, which has narrower thermal tolerances and is more susceptible to
desiccation. The former is more likely to disperse to suitable trees beyond the natal
site, and the greater rarity of C. galili was in part attributed to lower incidence in the
colder and drier months.
In many mutualisms, warming in spring can be a signal to synchronise develop-
ment or behaviour – so that pollinators are synchronised with flower opening or
herbivores with flush growth as optimal quality food – but with the corollary that
any differential change in such cues can disrupt the associations so that the partners
may then respond in different ways. Explored for synchrony of seed-dispersing ants
(Aphaenogaster spp.) in North America on early and later-flowering plants along an
elevational gradient, reflecting temperature differences, Warren and Bradford (2014)
noted that the warm-adapted Aphaenogaster rudis had progressively moved upward
and replaced cool-adapted A. picea in many places. Their trials also involved trans-
plantations of plants to other climate regimes. Failures in seed dispersal reflected
not simply the absence of a corresponding mutualist partner, but the seasonal syn-
chrony between the partners was disrupted. The early-blooming plants could not
synchronise with the warm-adapted ant mutualist, leaving the plant populations
clumped. In contrast, the mutualisms persisted in the later-blooming plants because
seed production then coincided with ant foraging activity, and dispersal was not
interrupted. Stuble et al. (2014) used a similar context to investigate whether ant
seed dispersal mutualisms were affected by changing temperature, pointing out that,
if a keystone myrmecochore was affected, plants relying on that ant might become
at increased risk. They noted that Aphaenogaster rudis was responsible for >90% of
ant-mediated seed dispersal in some North American forests and that any impacts
on this ant could become widespread. Those trials investigated the rates of seed
removal from ‘warming chambers’ placed in forests, these chambers using hydronic
radiators to warm the enclosed 5 m diameter plots by measured amounts. Seeds
were removed by seven ant species, with, unexpectedly, A. rudis removing <50% of
all seeds taken. Such multiplicity of myrmecochores may increase resilience to
warming, by the mixed responses of the contributing species. Further, as implied
above, major natural periods of seed removal have normal temperatures far below
the critical thermal maximum for A. rudis – however, that ant species with high
thermal tolerance are most active under the warmest temperatures experienced
below that limit suggests also that climate warming has some potential to increase
rates of seed dispersal.
Information on the tolerances and behaviour of individual species remains a key
element in estimating if, and how, a mutualism may respond to climate changes.
Species-specific information feeds evaluation of vulnerability, with the major driv-
ers of vulnerability being persistence and the ability to shift range. The latter
involves both the availability of climatically suitable areas and species associations
and dispersal capability. If either partner lacks capability to withstand change in the
affected region, some form of range shift is a fundamental need. Referring to indi-
vidual species, Arribas et al. (2012) designed the decision framework summarised
in Fig. 10.5 to help determine the major conservation needs. Harmonising a strategy
for partners in mutualism might benefit from this approach in seeking ways in which
their mutual interests might be given priority.
Decoupling of mutualisms by climatic influences may take a considerable period,
over which other factors or counter-adaptations are also likely to occur. Thus, Sagata
and Gibb (2016) modelled the possible future trends of an Australian ant-hemipteran-
plant interaction (involving Iridomyrmex rufoniger, Eriococcus coriaceus,
Eucalyptus camaldulensis) under three predicted Melbourne (Victoria) mean sum-
mer temperature regimes – namely, 23 °C (2011), 25 °C (2050) and 29 °C (2100) -
Fig. 10.5 Conservation
strategies: a decision
framework to guide
strategies, derived from the
different drivers of species’
vulnerability to climate
change (Arribas et al.
2012)
to indicate possible scenarios of plant growth and ant foraging activity increases,
but these linked with scale insect growth, abundance, size and honeydew standing
crop on trees declining by the highest temperature. Implied impacts of ant harvest-
ing were most evident under the 29 °C projection, by which the combined impacts
of temperature and ants reduced the honeydew standing crop to barely detectable
levels. Such inexorable but subtle and difficult-to-project changes may become
numerous and collectively add considerable uncertainty to forecasting the continu-
ation of many intricate associations. It is almost universal that the projected adaptive
capability of the partners in a mutualism over many generations is unknown, but
greater information on dispersal capability and current climatic optima is also more
likely to be available.
The need to identify the abiotic factors that determine the vulnerability of species
to climate change in planning conservation has often been expressed (Klockmann
et al. 2016). The differing capabilities and tolerances of partners in a mutualism add
to the complexity of this task. As in many other contexts of evaluating potential
impacts of climate changes, comparative studies of traits along elevational or
latitudinal gradients have helped to indicate likely changes in mutualisms and toler-
ances by revealing differences between the partners’ capabilities to respond.
A study of 24 species of myrmecophilous Lycaenidae: Polyommatinae (of a total
fauna of 29 lycaenid species found in the study region) and their ant associates over
an elevational range of 1000–3200 m in the Swiss alps (Pellissier et al. 2012a, b)
suggested that absence of myrmecophily was a secondary state that represented loss
of mutualism from either facultative or obligate forerunners. Low degree of myr-
mecophily in species, and tolerance of lower temperature conditions, indicated loss

of mutualism, or its less specialised nature, in cold climate conditions. Lower tem-
perature communities comprised more generalised species with less myrmecophily
and were interpreted as a probable ‘environmental filter’ effect. Pellissier et al. sug-
gested that the severe (high elevation) environments could affect myrmecophily in
three ways, which were not mutually exclusive. These were (1) increase in the net
direct costs of providing resources for the ants, (2) reduction in availability of ants
and increase in indirect costs and (3) changed predation pressures influencing the
benefits of myrmecophily. For each of these in turn, tentative reasoning was:
1. The direct costs of myrmecophily/mutualism may simply become too difficult to
meet under conditions in which temperature, resources and suitable develop-
mental periods are all reduced, with shortened developmental time influencing
chances of parasitisation and the moister conditions possibly enhancing risks of
disease.
2. The cost-to-benefit ratio of mutualism may suffer if the partners become more
difficult to encounter: under conditions with decreased ant diversity, any special-
ised butterfly may find it more difficult to interact with a given ant species –
absence of some butterflies from the highest elevations apparently reflected their
host ants being rarer there.
3. Under conditions with reduced pressure from natural enemies, benefits to larvae
from protection by ants are diminished, and the decreased importance of this
primary driver of the mutualism becomes less relevant.
However, the costs of producing the attractant secretions would continue, so the
pressures to abandon myrmecophily are enhanced. In this survey, the colder condi-
tions of higher elevations tended to exclude the lycaenids that were associated
strongly with ants, so promoting loss of mutualisms. Some lycaenids living in such
extreme habitats where ants are scarce have evolved reductions in their ‘nectar’-
producing organs, a parallel to changes that occur in endophytic feeders.
In many ways, severe or changing climatic conditions ‘filter’ assemblages of
related species and, thus, the associations in which they participate. Many studies
exemplify changes within higher insect taxa, but most do not directly include con-
siderations of mutualists. Thus, richness and functional diversity of ants decreased
with decreasing temperatures along an elevational gradient in the Swiss alps
(Reymond et al. 2013), with likely consequences for ant-butterfly mutualisms.
Microevolutionary adaptations to climate changes and changes in trophic interac-
tions are widespread, and needs to adopt different larval host plants may also affect
ant-butterfly partnerships and be vital to prevent local extirpations (Descombes
et al. 2016). Cooler temperatures at higher elevations or latitudes prevent colonisa-
tion of many insect lineages, whilst facilitating occupancy by a restricted range of,
sometimes closely related, species that can tolerate those more severe conditions
(Pellissier et al. 2013). Thus, in general, Lycaenidae are relatively rare at higher
elevations, a feature attributed by Pellissier et al. to changing balance in mutualisms
or absence/limitation of specific larval food plants. The intricacies of biotic associa-
tions amongst Lycaenidae render them collectively a ‘sensitive clade’ (Pellissier

et al. 2013) likely to be susceptible to changing conditions.
Local abiotic environmental conditions – whether or not related to climate
changes but many (such as sward height and vigour, shade, site aspect in relation to
insolation and others) difficult to exclude from climatic influences - may thus
change the cost-benefit balance within a mutualism continuously, and the mutual-
ism may break down completely if a reasonable balance is not sustained (Bronstein
1994, Leigh 2010). As examples of the influences on aphid-ant mutualisms, Yao
(2014) listed (1) influences of the host plant, (2) parasitoid wasps, (3) ant predation
on aphids, (4) competition for mutualistic ants and (5) the physiological costs of ant
attendance. Any such list for a mutualistic context could be expanded to incorporate
more specific details, such as the roles of cuticular hydrocarbons in aphid-ant asso-
ciations (as in lycaenid-ant associations). The roles of parasitoid wasps of aphids,
for example, render the associations even more intricate: the wasps may be thwarted
by ants, but after they oviposit, the parasitoidised hosts are also protected by the
ants, providing circumstances that may lead to higher emergence rates for the wasps.
It could thus be claimed that the normally beneficial protection of aphids by the ants
might increase the costs of that attendance (Yao 2014).
Simply shading potted plants of Tripsacum dactyloides (Poaceae) supporting
populations of the trophobiont five-spotted gama leafhopper (Dalbulus quinque-
notata, Cicadellidae) by installing a 45% shade mesh 3 m overhead led to increased
numbers of its attending obligatory mutualist ant, Brachymyrmex obscurior (Moya-
Raygoza and Martinez 2014). The mutualism, developing largely on the basal leaves
of the grass, persisted in both shaded and unshaded treatments – but significantly
higher numbers of leafhopper nymphs on shaded plants were attended by signifi-
cantly greater numbers of Brachymyrmex, and the mutualism was thus ‘more
intense’ there, with likely consequences including different pressures from natural
enemies on the treatments, and consequent different needs for protection.
Other studies have also demonstrated that shaded grasses are superior food
sources for phloem-feeding Hemiptera, through their sap having higher nitrogen
concentrations. Trials for the Dalbulus-Brachymyrmex associations are summarised
in Fig. 10.6. Interestingly, D. quinquenotata produces more honeydew than con-
genors that are not attended by ants and is also more sessile (Larsen et al. 1992),
both features that could foster mutualism.
In a greenhouse environment, D. quinquenotata produced larger (by about 23%)
and more frequent (two to four times as many) droplets of nectar, to produce three
to six times the volume produced by two non-attended species, D. maidis and D.
gelbus. Non-attended leafhoppers walk, jump or fly away when approached by ants,
in contrast to D. quinquenotata which responds positively to them by extruding
honeydew droplets and retaining these for direct removal from the anal tube. The
other species simply expel honeydew rapidly.
Persistence of mutualisms depends both on the rate of climate change and the
densities and seasonal activities of the species involved – many of which may char-
acteristically be highly seasonal in appearance. Gilman et al. (2011) explored the
circumstances under which this might occur, with the initial premises that faculta-
Fig. 10.6 The effects of

attendant ants on
honeydew production by
Dalbulus leafhoppers. D.
quinquenotatus is a
myrmecophilous species.
Data are shown for three
species; the other two not
myrmecophilous and are
for (a) droplet diameter
(μm) and (b) time intervals
(seconds) between droplet
productions (Larsen et al.
1992)
tive mutualists should be more resilient to disturbances than obligate mutualists and
that many mutualistic pollinators were indeed more likely to persist when abundant
alternative food sources were available to counter declines of any mutualist flower.
In parallel, evolving new phenological patterns in response to climate change may
be feasible if mutualist partners can be obtained outside the current flowering or
foraging periods. Gilman et al.’s models, however, included the considerable com-
plexities of generalising to predict which populations may be vulnerable to climate
changes by breakdown of mutualisms. The critical field tests, of determining
whether each partner population can persist if the other is absent, are very difficult
to undertake.
10.6 Reintroductions
Increasingly, reintroductions and other forms of translocations are considered for

insect species conservation, commonly associated with measures to restore or
enrich receptor sites so that the key resources needed are readily accessible. Assisted
colonisation, although still somewhat controversial in conservation management, is
also an important consideration in conserving biological associations and interac-
tions. Success may depend on re-establishing past mutualisms, and failure to recog-
nise the intricacies of these has sometimes led to failures of ambitious programmes.
For many plants, photoperiod and temperature constraints imposed by latitudinal
10.6 Reintroductions 217
transfers may affect the variety of their responses and compatibility with other taxa
(Wadygmar et al. 2015), including mutualists.
With increased interest in measures such as assisted migration to anticipate cli-
mate change impacts, knowledge of mutualisms and the tolerances of the various
partners becomes of critical importance – together with knowledge of the identity
(and, increasingly, of genetic structure) of the populations involved. Thus, the suc-
cessful reintroduction of Maculinea arion (p. 125) to Britain was made from but-
terflies from a Swedish (Oland) stock (Thomas et al. 2009), but ‘only educated
guesses could be made as to whether this founder population was genetically repre-
sentative’ for north-western Europe (Ugelvig et al. 2012). Populations of M. arion
are highly structured genetically, but this notable success could not – at that time –
incorporate that knowledge. The complexities of defining needs of mutualistic spe-
cies in undertaking reintroductions were illustrated well by introducing two other
species of Maculinea to the same site in the Netherlands (Wynhoff 2001), following
their loss from the country. M. nausithous and M. teleius were both brought to the
Moerputten nature reserve, selected because it harboured the native food plant
(Sanguisorba officinalis) and both the major Myrmica host species. Adult butterflies
of both species, from Polish stock, were released on the same day. Both established
successfully, but differences in preferred habitat structure and butterfly dispersal
patterns led to somewhat different trajectories. M. nausithous occurred in ‘rougher’
vegetation and M. teleius more in ‘open’ vegetation areas, and the latter also had
wider nectar sources as the butterflies visited many flower species, rather than
Sanguisorba alone. Different needs of, and dependences on, the major host ants of
each species also contributed to demonstrating the minutiae of ecological details
that attend increasing the chances of success of introducing ecologically intricate
taxa. Progress in conservation of myrmecophilous Lycaenidae has effectively drawn
attention to the complexities of considering mutualisms and their requirements.
Many other cases may have failed or succeeded fortuitously and without full
awareness of restrictions linked with that specificity and variety. They transcend the
categories of (1) introductions to new sites, (2) reintroductions to sites historically
occupied by the species, (3) translocations from one site to others and (4) augment-
ing and supplementing populations (butterflies: Schultz et al. 2008). Many of those
exercises are conjoint with, or follow, restoration measures to improve site or
resource conditions, but most such efforts have concentrated on vegetation and key
food plant regimes. However, primary emphasis on habitat restoration for butterflies
has differed somewhat between the United States and Britain (the areas reviewed by
Schultz et al.), as (1) emphasis on eradication of alien plants, and promoting native
species in conjunction with natural disturbance regimes (United States), and (2) re-
establishing traditional management practices such as coppicing and grazing
(United Kingdom). Many ecological subtleties are thereby approached or incorpo-
rated indirectly into management. Direct considerations of mutualisms have tradi-
tionally been relatively rare, but knowledge of their significance and intricacies has
been enhanced by moves to conserve myrmecophilous Lycaenidae, in particular.
Management of chalk grassland to conserve British Lycaenidae (BUTT 1986) indi-
cates the mosaic of regimes that may be necessary to conserve co-occurring butter-
flies and their host ants. There, as elsewhere, inappropriate habitat management
could lead to conditions in which a less suitable host ant predominates over the most
desirable associate and in which butterfly larval survival is reduced. Whilst this
context is now well understood through long-term deliberate investigation, many
parallel efforts are far less informed. For whatever reason, failure to consider mutu-
alisms may increase risk of failure in practical species-focused conservation.
Any such translocations, often expensive and complex exercises to undertake
and monitor, must accept the risk of coextinctions occurring if suitable mutualistic
partners are not available in the receiving environment. In many other cases, reduced
‘performance’ is likely as partners re-associate or novel mutualisms develop. The
most traditional practice of planning translocations almost entirely for each single
species, even when multiple species are moved to the same site(s), has minimised
the practical importance of sustaining interactions. The need to consider deliberate
co-translocations of mutualistic partners is urgent (Plein et al. 2016).
10.7 Prospects
Aslan et al. (2013) used the term ‘widow species’ for mutualists whose partners
have become extinct and, from reviewing threats of vertebrate extinctions on
vertebrate-pollinated plants, suggested that many such plants may be affected, with
likely cascading effects to create widespread impacts on ‘biodiversity’. They pro-
posed that conservation measures focusing explicitly on mutualisms could be vital
to both (1) conserve widow species and (2) maintain ecological functions. That
review, despite the considerable uncertainties revealed, is likely to reflect the gen-
eral scenario for losses of insect mutualists which have attracted relatively little
specific attention beyond pollination networks. Consequences vary widely across
species, regions and ecological contexts, and individual studies on insect mutual-
isms each contribute to awareness of their vulnerability and ecological subtleties.
Not least, determining whether novel or alternative partnerships are likely or feasi-
ble is important in assessing the scope of conservation measures needed. Most such
measures will be associated with some changes in fitness from the former balance
between the partners.
The loss of ecological interactions is often treated as secondary to a primary
focus on losses of species per se (Valiente-Banuet et al. 2014). Species losses inevi-
tably cause changes in the interactions in which they participate directly. Interaction
losses may precede complete loss of the participating species as the latter decline to
levels below those at which they can fulfil their normal ecological roles – for exam-
ple, particular specialised pollinators may decline to levels that result in reduced
reproduction of their plant partners. However, the impact of any interaction loss on
wider ecological functions varies, leading Valiente-Banuet et al. to identify three
categories of functionality. These were termed:
10.7 Prospects 219
1. Additive interactions, when the loss of the interaction affects at least one of the
other species involved but has no direct effects on the other links of the interac-
tion network – exemplified by fig-fig wasp interactions as the pollination service
depends on the contributions of every species-specific fig-fig wasp interaction.
2. Keystone interactions, when the interaction controls the ecosystem or commu-
nity fitness to a disproportionately high extent – it is expected to occur when
there are strong differences in contributions of interspecific interactions to eco-
logical function and exemplified by removal of generalised mutualists from
plant-pollinator or seed dispersal networks.
3. Redundant interactions, occurring when any contribution to wider community
function is replaceable – such as many cases of pollination in which pollinators
are essentially interchangeable and, at other than extreme levels of anthropo-
genic change, ecological functions are not affected.
Valiente-Banuet et al. (2014) noted ‘interaction losses are early warning signals
of critical transitions’, and using these may help detect the levels of habitat loss for
which loss of ecosystem functions is likely to arise.
The resources required by any species are difficult to both define and manage
fully, and the local environment (in a broad sense of providing both context and
conditions needed by any species selected for individual conservation attention) can
only rarely be appraised comprehensively for its ability to sustain the supply of
those critical resources. The most obvious resources are those linked with food
needs, as ‘consumables’ that with the broader array of ‘utilities’ emphasised by
Dennis (2010) largely define habitat suitability. They, with the embedding ‘condi-
tioners’ such as suitable climate, provide the template for conservation need and
likely success. Historically, many conservation attempts, even for relatively well-
known insects such as butterflies in the northern temperate regions, have failed,
whether in situ or following apparently well-planned translocations or other reintro-
ductions. Rarely documented in detail, lack of awareness of suitable resource needs
can reasonably be implicated in many such failures. Further complications may
arise when those needs include poorly understood interspecific associations, not
least in expanding any initially well-defined conservation perspective or plan into
more nebulous areas such as the balance of food webs influenced by a focal insect
or a competing alien invader. As Dennis elegantly documented, the desirable out-
come of an insect conservation programme is the sustainability (or restoration) of
the arena in which the focal species or wider target entity can thrive, together with
the resources on which it depends or with which it otherwise interacts.
The continuation of mutualisms is a significant functional strand in many such
enterprises, and managers cannot overlook the formidable intricacies these may
pose. In some cases, acknowledgement and understanding that mutualisms occur
and may be difficult to incorporate properly into conservation planning is the major
theme that must be addressed. Perspectives flowing from routine investigations into
whether such associations occur can illuminate potential conservation measures
and, importantly, reduce futility in such programmes. The salutary ‘hindsight’ from
cases such as for Maculinea arion in Britain (Thomas 1980) has had considerable
benefits in providing the ‘foresight’ aiding parallel measures for other myrmecophi-
lous Lycaenidae. That foresight has benefitted individual species and the wider
environments of communities and landscapes in which they occur.
The maxim that conservation managers ‘must think and act like biologists with
the goal of conserving interactions and interacting systems’ needs frequent and
emphatic reiteration three decades after its formulation by Janzen (1987), with –
again in his words – recognition that insects are ‘the building blocks and glue’ of
habitats such as forests. Intricate core interactions such as mutualisms are amongst
the most necessary and specific adhesives that help to structure insect communities
and functional ecological modules in which insects are involved. The examples out-
lined in this book illustrate the complexities and subtleties of mutualistic interac-
tions involving insects, which impinge on their survival and, hence, conservation
need. Species conservation programmes that ignore, fail to appreciate or diminish
the functional relevance of such interactions may also diminish both their credibility
and their chances of success.
References
Amarasekare P (2004) Spatial dynamics of mutualistic interactions. J Anim Ecol 73:128–142

Arribas P, Abellan P, Velasco J, Bilton DT, Millan A, Sanchez-Fernandez D (2012) Evaluating
drivers of vulnerability to climate change: a guide for insect conservation strategies. Glob
Chang Biol 18:2135–2146
Ashworth L, Aguilar R, Galetto L, Aizen MA (2004) Why do pollination generalist and specialist
plant species show similar reproductive susceptibility to habitat fragmentation? J Ecol
92:717–719
Aslan CE, Zavaleta ES, Tershy B, Croll D (2013) Mutualism disruption threatens global plant
biodiversity: a systematic review. PLoS One 8(6):e 66993. doi:10.1371/journal.pone.0066993
Banaszak J (1992) Strategy for conservation of wild bees in an agricultural landscape. Agric
Ecosyst Environ 40:179–192
Biesmeijer JC, Roberts SPM, Reemer M, Ohlemuller R, Edwards M (and 7 other authors) (2006)
Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands.
Science 313:351–354
Bluthgen N, Fiedler K (2004) Preferences for sugars and amino acids and their conditionality in a
diverse nectar-feeding ant community. J Anim Ecol 73:155–166
Bond WJ (1994) Do mutualisms matter? Assessing the impact of pollinator and disperser disrup-
tion on plant extinction. Philos Trans R Soc Lond B 344:83–90
Bronstein JL (1989) A mutualism at the edge of its range. Experientia 45:622–637
Bronstein JL (1994) Our current understanding of mutualism. Q Rev Biol 69:31–51
Bronstein JL, Wilson WG, Morris WF (2003) Ecological dynamics of mutualist/antagonist com-
munities. Am Nat 162:S24–S39
Bronstein JL, Dieckmann U, Ferriere R (2004) Coevolutionary dynamics and the conservation of
mutualisms. Interim Report IR-04-061, International Institute for Applied Systems Analysis,
Laxenburg
Bruna EM, Vasconcelos HL, Heredia S (2005) The effect of habitat fragmentation on communities
of mutualists: Amazonian ants and their host plants. Biol Conserv 124:209–216
BUTT (Butterflies Under Threat Team) (1986) The management of chalk grassland for butterflies.
Focus on Nature Conservation, no. 17. Nature Conservation Council, Peterborough
References 221
Conrad KF, Woiwod IP, Parsons M, Fox R, Warren MS (2004) Long-term population trends in
widespread British moths. J Insect Conserv 8:119–136
Dennis RLH (2010) A resource-based habitat view for conservation. Butterflies in the British land-
scape. Wiley-Blackwell, Oxford
Descombes P, Pradervand J-N, Golay J, Guisan A, Pellissier L (2016) Simulated shifts in trophic
niche breadth modulate range loss of alpine butterflies under climate change. Ecography
39:796–804
Drieu R, Rusch A (2016) Conserving species-rich predator assemblages strengthens natural pest
control in a climate warming context. Agric For Entomol. doi:10.1111/afe.12180
Evans DM, Turley NE, Tewksbury JJ (2013) Habitat edge effects alter ant-guard protection against
herbivory. Landsc Ecol 28:1743–1754
Fayle TM, Edwards DP, Foster WA, Yusah KM, Turner REC (2015) An ant-plant by-product mutu-
alism is robust to selective logging of rain forest and conversion to oil palm plantation.
Filz KJ, Schmitt T (2015) Niche overlap and host specificity in parasitic Maculinea butterflies
(Lepidoptera: Lycaenidae) as a measure for potential extinction risks under climate change.
Org Divers Evol 15:555–565
Fortuna MA, Bascompte J (2006) Habitat loss and the structure of plant-animal mutualistic net-
works. Ecol Lett 9:281–286
Forup ML, Henson KSE, Craze PG, Memmott J (2008) The restoration of ecological interactions:
plant-pollinator networks on ancient and restored heathlands. J Appl Ecol 45:743–752
Fox R (2013) The decline of moths in Great Britain: a review of possible causes. Insect Conserv
DIvers 6:5–19
Garibaldi LA, Steffan-Dewenter I, Winfree R, Aizen MA, Bommarco R (and 45 other authors)
(2013) Wild pollinators enhance fruit set of crops regardless of honey bee presence. Science
339: 1608–1611
Ghazoul J (2005) Buzziness as usual? Questioning the global pollination crisis. Trends Ecol Evol
20:367–373
Gilman RT, Fabina NA, Abbott KC, Rafferty NE (2011) Evolution of plant-pollinator mutualisms
in response to climate change. Evol Appl. doi:10.1111/j.1752-4571.2011.00202.x
Harris LF, Johnson SD (2004) The consequences of habitat fragmentation for plant-pollinator
mutualisms. Int J Trop Insect Sci 24:29–43
Hegland SJ, Nielsen A, Lazaro A, Bjerknes A-L, Totland O (2009) How does climate warming
affect plant-pollinator interactions? Ecol Lett 12:184–195
Hochberg ME, Thomas JA, Elmes GW (1992) A modelling study of the population dynamics of a
Large Blue butterfly, Maculinea rebeli, a parasite of red ant nests. J Anim Ecol 61:397–409
Huang J, An J, Wu J, Williams PH (2015) Extreme food-plant specialisation in Megabombus bum-
blebees as a product of long tongues combined with short nesting seasons. PLoS One.
doi:10.1371/lournal.poe.0132358
Ikouchi Y, Suetsugu K, Sumikawa H (2015) Diurnal skipper Pelopidas mathias (Lepidoptera:
Hesperiidae) pollinates Habenaria radiata (Orchidaceae). Entomol News 125:7–11
Janzen DH (1987) Insect diversity of a Costa Rican dry forest: why keep it, and how? Biol J Linn
Soc 30:343–356
Jevanandam N, Goh AGR, Corlett RT (2013) Climate warming and the potential extinction of fig
wasps, the obligate pollinators of figs. Biol Lett 9:20130041. doi.org/10.1098/rsbl.2013.0041
Johnson SD, Neal PR, Peter CI, Edwards TL (2004) Fruiting failure and limited recruitment in
remnant populations of the hawkmoth-pollinated tree Oxyanthus pyriformis subsp. pyriformis
(Rubiaceae). Biol Conserv 120:31–39
Jordano P (1987) Patterns of mutualistic interactions in pollination and seed dispersal: con-
nectance, dependence asymmetries, and coevolution. Am Nat 129:657–677
mutualisms on oceanic islands. Perspect Plant Ecol Evol Syst 12:131–143. doi:10.1016/j.
ppees.2009.10.002
Kersch MF, Fonseca CR (2005) Abiotic factors and the conditional outcome of an ant-plant mutu-
alism. Ecology 86:2117–2126
Klockmann M, Karajoli F, Kuczyk J, Reimer S, Fischer K (2016) Fitness implications of simulated
climate change in three species of copper butterflies (Lepidoptera: Lycaenidae). Biol J Linn
Soc doi. 10/1111/bij.12846
Larsen KJ, Heady SE, Nault LR (1992) Influence of ants (Hymenoptera: Formicidae) on honey-
dew excretion and escape behaviors in a myrmecophile, Dalbulus quinquenotatus (Homoptera:
Cicadellidae), and its congeners. J Insect Behav 5:109–122
Leimar O, Connor EC (2003) By-product benefits, reciprocity and pseudoreciprocity in mutual-
ism. In: Hammerstein P (ed) Genetic and cultural evolution of cooperation. MIT Press,
Cambridge, MA, pp 203–222
Macgregor CJ, Pocock MJO, Fox R, Evans DM (2015) Pollination by nocturnal Lepidoptera, and
the effects of light pollution: a review. Ecol Entomol 40:187–198
Marquis M, Del Toro I, Pelini SL (2014) Insect mutualisms buffer warming effects on multiple
trophic levels. Ecology 93:9–13
Memmott J (1999) The structure of a plant-pollinator food web. Ecol Lett 2:276–280
Menz MHM, Phillips RD, Winfree R, Kremen C, Aizen MA, Johnson SD, Dixon KW (2011)
Reconnecting plants and pollinators: challenges in the restoration of pollination mutualisms.
Trends Plant Sci 16:4–11
M’Gonigle LK, Williams NM, Lonsdorf E, Kremen C (2016) A tool for selecting plants when
restoring habitat for pollinators. Conserv Lett. doi:10.1111/conl.12261
Munguira ML, Martin J (1999) Action plan for Maculinea butterflies in Europe. Nature and
Environment Series no. 97. Strasbourg, Council of Europe Publishing
Moya-Raygoza G, Martinez AV (2014) Ants (Hymenoptera: Formicidae) and trophobiont leafhop-
per nymphs (Hemiptera: Cicadellidae) become more abundant in shaded conditions: implica-
tions for mutualism. Fla Entomol 97:1378–1385
New TR (2012) Hymenoptera and conservation. Wiley-Blackwell, Oxford
Pellissier L, Litsios G, Fiedler K, Pottier J, Dubuis A, Pradervand J-N, Salamin N, Guisan A
(2012a) Loss of interactions with ants under cold climate in a regional myrmecophilous but-
terfly fauna. J Bigeogr 39:1782–1780
Pellissier L, Rasmanns S, Litsiong G, Fiedler K, Dubuis A, Pottier J, Guisan A (2012b) High host-
plant nitrogen content: a prerequisite for the evolution of ant-caterpillar mutualisms? J Evol
Biol. doi:10.1111/j.1420-9101.2012.02555x
Pellissier L, Alvarez N, Espindola A, Pottier J, Dubuis A, Pradervand J-N, Guisan A (2013)
Phylogenetic alpha and beta diversities of butterfly communities correlate with climate in the
western Swiss Alps. Ecography 36:541–550
Pemberton RW, Wheeler GS (2006) Orchid bees don’t need orchids, evidence from the naturaliza-
tion of an orchid bee in Florida. Ecology 87:1995–2001
Plein M, Bode M, Moir ML, Vesk PA (2016) Translocation strategies for mutualistic species
depend on interspecific interaction type. Ecol Appl 26:1186–1197
Prinzing A, Dauber J, Hammer EC, Hammouti N, Bohning-Gaese K (2007) Perturbed partners:
opposite responses of plant and animal mutualist guilds to inundation disturbances. Oikos
116:1299–1310
Reymond A, Purcell J, Cherix D, Guisan A, Pellissier L (2013) Functional diversity decreases with
temperature in high elevation ant fauna. Ecol Entomol 38:364–373
Sagata K, Gibb H (2016) The effect of temperature increases on an ant-Hemiptera-plant interac-
tion. PLoS One 11(7):e0155131. doi:10.1371/lournal.pone.0155131
Schultz CB, Russell C, Wynn L (2008) Restoration, reintroduction, and captive propagation for
at-risk butterflies: a review of British and American conservation efforts. Israel J Ecol Evol
54:41–61
References 223
Senapathi D, Biesmeijer JC, Breeze TD, Kleijn D, Potts SG, Carvalheiro LG (2015) Pollinator
conservation – the difference between managing for pollination services and preserving polli-
nator diversity. Curr Opin Insect Sci 12:93–101
Stanton ML (2003) Interacting guilds: moving beyond the pairwise perspective on mutualisms.
Am Nat 162:S10–S23
Stuble KL, Patterson CM, Rodriguez-Cabal MA, Ribbons RR, Dunn RR, Sanders NJ (2014) Ant-
mediated seed dispersal in a warmed world. PeerJ:e286. doi:10.7717/peerj.286
Taki H, Kevan PG (2007) Does habitat loss affect the communities of plants and insects equally in
plant-pollinator interactions? Preliminary findings. Biodivers Conserv 16:3147–3161
Tanaka HO, Yamane S, Nakashizuka T, Momose K, Itioka T (2007) Effects of deforestation on
mutualistic interactions of ants with plants and hemipterans in tropical rainforest of Borneo.
Asian Myrmecol 1:31–50
Thomas JA (1980) Why did the large blue become extinct in Britain? Oryx 15:243–247
Thomas JA, Simcox DJ, Clarke RT (2009) Successful conservation of a threatened Maculinea but-
terfly. Science 325:80–83
Thompson JN, Fernandez CC (2006) Temporal dynamics of antagonism and mutualism in a geo-
graphically variable plant-insect interaction. Ecology 87:103–112
Thomson FJ, Auld TD, Ramp D, Kingsford RD (2016) A switch in keystone seed-dispersing ant
genera between two elevations for a myrmecochorous plant, Acacia terminalis. PLoS One
11(6):e 0157632. doi:10.1371/journal.pone.0157632
Tylianakis JM, Didham RK, Bascompte J, Wardle DA (2008) Global change and species interac-
tions in terrestrial ecosystems. Ecol Lett 11:1351–1363
Ugelvig LV, Andersen A, Boomsma JJ, Nash DR (2012) Dispersal and gene flow in the rare, para-
sitic Large Blue butterfly Maculinea arion. Mol Ecol 21:3224–3236
Valiente-Banuet A, Aizen MA, Alcantara JM, Arroyo J, Cocucci A (and 14 other authors) (2014)
Beyond species loss: the extinction of ecological interactions in a changing world. Funct Ecol.
doi:10.1111/1365-2435.12356
Vazquez DP, Bluthgen N, Cagnolo L, Chacoff NP (2009) Uniting pattern and process in plant-
animal mutualistic networks: a review. Ann Bot 103:1445–1457
Wadygmar AM, Cumming MN, Weis AE (2015) The success of assisted colonization and assisted
gene flow depends on phenology. Glob Chang Biol. doi:10.1111/gcb.1988
Wang R-W, Sun B-F (2009) Seasonal change in the structure of fig-wasp community and its impli-
cation for conservation. Symbiosis 47:77–83
Warren M, Robertson MP, Greeff JM (2010) A comparative approach to understanding factors
limiting abundance patterns and distributions in a fig tree-fig wasp mutualism. Ecography
33:148–158
Warren RJ II, Bradford MA (2014) Mutualism fails when climate response differs between inter-
acting species. Glob Chang Biol 20:466–474
Wilcock C, Neyland R (2002) Pollination failure in plants: why it happens and when it matters.
Trends Plant Sci 7:270–277
Wynhoff I (2001) At home on foreign meadows. The reintroduction of two Maculinea butterfly
species. Wageningen Agricultural University, Wageningen
Yao I (2014) Costs and constraints in aphid-ant mutualism. Ecol Res 29:383–391
Zhou A, Kuang B, Gao Y (2015) Does the host plant affect the benefits from mutualisms? The
invasive mealybug and ghost ant association. Ecol Entomol 40:782–786
Index
A Ant-plants, 10, 23
Acacia, 18, 23, 122, 186 Ants, 22, 24
A. collinsii, 71 Aphaenogaster, 209, 212
A. cornigera, 64, 68 A. picea, 212
A. drepanolobium, 93 A. rudis, 90, 154, 162, 212
A. hindsii, 94 Aphids, 9, 28
A. ligulata, 91 Aphnaeinae, 133
A. terminalis, 209 Aphomomyrmex afer, 66
Aconitum, 206 Aphytophagy, 122
Acoustic signals, 131 Apidae, 46
Acrodipsas, 124, 125 Apis, 175
Acrodipsas myrmecophila, 125 Apis mellifera, 58, 158, 160, 161, 193, 205
Adomerus rotundus, 92 Appeasement hypothesis, 134
Aeropetes tulbaghia, 40 Aprovechados, 173
Agaonidae, 46 Araliaceae, 137
Agavaceae, 46 Arctiidae, 106
Alfalfa, 123 Argentine ant, 162
Allomerus octoarticulatus, 24, 93 Arhopala, 17, 120, 128
Allosmaitia strophius, 128 A. amphimuta, 129
Allotinus apries, 105 A. dajagaka, 129
Amblyopini, 10 A. zylda, 129, 130
Amblyopinus tiptoni, 10 Aricia agestis, 132
Ambrosia beetles, 145 Arion subfuscus, 162
Anagraecum sesquipedale, 15, 22 Aristolochiaceae, 162
Aneuretinae, 85 Asarum canadense, 89, 162
Angelonia, 45 Asclepiadaceae, 84
Anolis carolinensis, 161 Assisted colonisation, 216
Anonychomyrma, 125 Asteraceae, 111
Anoplolepis longipes, 157 Asymmetry, 13
Anostostomatidae, 6 Atta
Antagonisms, 6 A. cephalotes, 95, 96
Ant blues, 125 A. colombica, 95
Anthene emolus, 135 A. sexdens, 96
Anthomyiidae, 39, 53, 176, 185 Attini, 94

DOI 10.1007/978-3-319-58292-4
226 Index
Auchenorrhyncha, 105, 107, 109 C

Azteca, 24, 93, 157 Cactus bees, 163
A. instabilis, 186 Caesalpiniaceae, 69, 134
A. pittieri, 109 Caladenia, 43
A. trigona, 168 Calathea
Azyga orbigera, 186 C. crotalifera, 16
C. ovandensis, 91
Calliphoridae, 40
B Camponotus, 137
Bacillus, 144 C. brutus, 106
Bamboo, 116, 137 C. cruentatus, 89
Banistereopsis malifolia, 131 C. fastigatus, 196
Bark beetles, 145 C. japonicus, 131
Barteria C. pennsylvanicus, 210
B. dewevrei, 93 C. rufipes, 28
B. fistulosa, 93 Carnation, 41
Basidiomycetes, 95 Carpenter bee, 183
Basistemon, 45 Carvacrol, 128
Batesian mimicry, 180 Caryophyllaceae, 41, 56
Batesian mimics, 173 Cataulacus mckeyi, 93, 187
Bat pollination, 52 Caternaultiella rugosa, 106
Bats, 155 Cecidomyiidae, 28
Bees, 8, 12, 55, 58, 158, 160, 205 Cecropia, 157
hawk moths, 40 C. obtusifolia, 157
hotels, 196 C. sciadophylla, 157
Beetles, 10, 16, 75 Centaurea, 41
Biological markets, 8 Centridini, 46
Biological trading, 9 Centris, 45
Birds, 155, 158 Cephalotus pusillus, 28
Bogus yucca moths, 179 Ceratocystis, 145
Bombus, 182, 193, 205 Ceratosolen, 178, 211
B. appositus, 184 C. arabicus, 178, 211
B. impatiens, 68 C. fusciceps, 178
B. (Megabombus) gerstaeckeri, 206 C. galili, 178, 211
B. occidentalis, 184, 185 C. gravelyi, 20
B. pascuorum, 185 Chaetothyriales, 93
B. terrestris, 160 Chaitophorus, 113
Botanophila, 176 C. populicola, 113, 163, 210
Brachymyrmex obscurior, 215 Chamaecrista
Brachystele unilateralis, 160 C. debilis, 69
Bradysia, 54 C. nictitans, 73
Braunsapis puangensis, 55 Cheaters, 17
Bright copper, 125 Chelidonium majus, 92
Bromeliaceae, 84 Cheritra freya, 134
Bucculatrix thurberiella, 65–66 Chiastocheta, 53
Bull-horn acacia, 68 Chilocoris kuwanae, 115
Bumblebees, 67, 158 Chilopsis linearis, 183
Bursaria spinosa, 125 Chloroea virescens, 160
Butterflies, 40 Chrysomya, 40
Butterfly gardening, 40 Chrysothamnus viscidiflorus, 111
Buzz pollination, 158 Cicadellidae, 215
Index 227
Cirsium arvense, 41 Diadasia, 163

Cistaceae, 67 Diaziella, 167
Citrus mealybug, 115 Diffuse mutualism, 3
Clusia, 157 Dinochloa trichogona, 116
Coadaptations, 37 Dioscorea prehensilis, 75
Coccidae, 109, 186 Dioscoreaceae, 75
Coccinellidae, 28, 107, 186 Disa nivea, 180
Coccoidea, 112 Dischidia, 84
Coccus, 17 Dolichoderinae, 162
C. viridis, 186 Dolichoderus, 116
Co-diversification, 49 D. cuspidatus, 116
Coevolution, 66 Domatia, 24, 55, 75–84, 93
Coextinctions, 15, 16 Dorsal nectary organs (DNOs), 121, 130
Coffee green scale, 186 Drepanididae, 158
Colias, 41 Drosicha corpulenta, 106
Commensalisms, 6 Drosophilid fly, 169
Cordia Drupadia
C. alliodora, 109 D. ravindra, 134
C. nodosa, 24, 93 D. theda, 134
Coryciinae, 44 Duroia hirsuta, 15, 24, 86
Corydalis caseana, 184
Cotesia cyaniridis, 121
Crematogaster, 17, 93, 129, E
134, 172 Ecosystem engineers, 95
C. californica, 163 Ecosystem processes, 193
C. nigriceps, 93 Ectatomminae, 69
C. opuntiae, 73 Edge effects, 200
C. quadriformis, 196 Elaiosomes, 87, 90, 162, 185
C. scutellaris, 179 Elephant, 19
Croton billbergianus, 170 Elizabethiella baijanthi, 49
Cucurbitaceae, 41 Ephydridae, 54
Cycadothrips chadwicki, 38 Epicephala, 52
Cycads, 38 Epicephala-Breynia vitis-idaea, 20
Cydnidae, 92 Epicephala-Glochidion, 20, 53
Cynopterus sphinx, 155 Epichloe elymi, 176
Cyphomyrmex, 95 Eriococcus coriaceus, 212
Cytinaceae, 67 Eris, 72
Cytinus hypocistis, 67 Eucalyptus camaldulensis, 212
Euglossa viridissima, 157
Euglossine bees, 181
D Euglossini, 195
Dalbulus, 108 Euphorbiaceae, 157, 170
D. gelbus, 215 Eurybia lycisca, 15
D. maidis, 215 Eurymelidae, 112
D. quinquenotata, 215 Exploitation barriers, 174
Dalechampsia scandens, 157 Exploiters, 21
Decarema, 17 Exploitionists, 71
Defences, 59, 65 Extinction cascades, 18
Deforestation, 198
Dendroctonus, 145
Dendrolasius, 106 F
Desert willow, 183 Fabaceae, 55, 65, 66, 90, 93
Devil’s gardens, 15, 86 Facilitations, 6
228 Index
Ferocactus Globeflower flies, 53

F. viridescens, 163 Glochidion acuminatum, 52
F. wislezeni, 12, 73 Gossypium, 59
Ficus, 46, 47, 167, 171, 211 G. thurberi, 65
F. aurea, 155 Gracillariidae, 52
F. burtt-davyi, 49 Grasshopper, 200
F. carica, 167, 179 Greya, 5
F. condensa, 167 G. politella, 197
F. hispida, 175 Guayaquila xiphias, 139
F. obtusifolia, 168 Gyrostigma, 19
F. paracamptophylla, 167
F. racemosa, 167, 175, 178, 200
F. semicordata, 20 H
F. sycomorus, 178, 211 Habenaria radiata, 203
Figs, 46 Hadena
Fig wasps, 46, 155 H. bicruris, 56
Five-spotted gama leafhopper, 215 H. consparcatoides, 56
Flies, 39, 58 Halictid bees, 51
Floral scents, 56 Harmonia axyridis, 107
Flower flies, 40 Hawk moths, 40, 201
Food bodies, 8, 75–84 Heliconius, 41, 146, 147
Forelius, 121 H. erato, 147
F. pruinosus, 65 H. melpomene, 147
Formica, 121, 123 H. numata, 147
F. altipetens, 136 Helleborus foetidus, 89
F. francoeuri, 187 Hemaris, 40
F. fusca, 107 Hemiargus isola, 121
F. neogates, 136 Hemideina, 6
F. obscuripes, 5, 110, 136 Hemiptera, 9
F. obscuriventris, 110 Hesperiidae, 16
F. podzolica, 5 Heteroptera, 106, 107
F. polyctena, 185 Heteropterys pteropetala, 69
F. propinqua, 113 Homalotylus shuvakhinae, 186
F. rufibarbis gelida, 69 Houardia abdita, 114
F. subsericea, 210 Hover flies, 40, 125
Fouquieraceae, 183 Humboldtia brunonis, 55
Fouquieria splendens, 183 Hummingbird pollination, 52
Fuchsia excorticata, 6 Hydnophytum, 84
Fungus comb, 143 Hydrellia tarsata, 54
Fungus gardens, 171 Hydrocharitaceae, 54
Fungus gnat, 54 Hylaeus, 158
Hylemya, 185
Hymenoptera, 63
G Hypolycaena erylus, 134
Gall-formers, 177
Gall midge, 28
Geometridae, 56 I
Giant scale insect, 106 Idarnes, 177
Gigantochloa scortechini, 137 Imperial hairstreak, 122
Glaucopsyche lygdamus, 5, 136 Inga, 65
Global pollination crisis, 195 I. vera, 196
Globeflower, 53 Ipomopsis aggregata, 185
Index 229
Iridaceae, 58 M
Iridomyrmex, 112, 122 Macaranga, 17, 120, 128
I. anceps, 122 M. rufescens, 129
I. gracilis, 67 Macropis
I. mayri, 169 M. europaea, 38
I. nitidiceps, 125 M. fulvipes, 38
I. rufoniger, 122, 169, 212 Macrotermes, 144
I. viridiaeneus, 91 Macrotermitinae, 143
Isola blue butterfly, 121 Macrozamia, 38
Ithomiinae, 146 Maculinea, 38, 120, 124, 125, 202, 210
M. alcon, 125, 127, 130
M. arion, 124, 125, 127, 128, 202, 217
J M. nausithous, 126, 217
Jalmenus evagoras, 122, 169 M. rebeli, 125, 127, 131
Jumping spiders, 72 M. teleius, 217
Malaicoccus formicarii, 116
Malpighiaceae, 46, 69, 131
L Malvaceae, 59, 65
Lamiaceae, 92 Marantaceae, 16, 91
Lamium amplexicaule, 92 Margarodidae, 106
Lapeirousia, 58 Medicago sativa, 123
Large blues, 120, 124, 202, 210 Megalomyrmex, 186
Lasius, 91 M. adamsae, 187
L. alienus, 124 Melanargia galathea, 42
L. flavus, 131 Melesitose, 114
L. niger, 115, 124, 131 Melinea, 147
L. umbratus, 210 Melittidae, 38, 44
Lathyrus montanus, 41 Membracidae, 110
Leaf-cutting ants, 5, 7, 94 Meranops, 67
Leafhoppers, 108 Metaphidippus, 72
Leonardoxa africana, 66, 93, Metapleural glands, 67
114, 187 Metrosideros polymorpha, 158
Leporella fimbriata, 67 Microdon mutabilis, 125
Leptidea sinapis, 41 Microtermes, 144
Leptothorax, 91 Microtis parviflora, 67
Light pollution, 204 Miletinae, 133
Lilioceris latipennis, 75 Mimetes cucullatus, 162
Linepithema humile, 162, 163 Mimicry ring, 146
Lithophragma parviflorum, 5, 197 Mimosaceae, 94, 196
Lithophragma-Prodoxidae, 46 Mirid bug, 65
Lizard, 161 Mistletoes, 134
Lophocereus schottii, 51 Moegistorhynchus longirostris, 58
Lophocereus-Upiga, 46 Monkshoods, 206
Loranthaceae, 28 Moraceae, 46, 47
Lucilia, 40 Moths, 20, 40, 56, 65
Lucilia papuensis, 40 Mountain pride butterfly, 40
Lycaeides melissa, 123 Mouse, 89
Lycaena (Gaeides) xanthoides, 187 Mullerian mimicry, 6, 146
Lycaena Mycangia, 145
L. phlaeas, 131 Myrmecia urens, 67
L. tityrus, 131 Myrmecodia, 84
Lycaenidae, 119 Myrmecophiles, 11
Lyonetiidae, 66 Myrmecophilous organs, 120
Lysimachia vulgaris, 38 Myrmecophily, 119, 133
230 Index
Myrmecophytes, 10, 21 Papaveraceae, 92, 184

Myrmecoxeny, 133 Paralucia aurifera, 125
Myrmelachista schumanni, 15, 24, 86 Paraputa anomala, 114
Myrmica, 126, 202, 210, 217 Parasites, 173
M. incompleta, 136 Parasitoids, 71, 177
M. rubra, 68, 126, 131, 154 Parategeticula, 46, 53
M. ruginodis, 89, 127 Parrhasius polibeta, 137
M. sabuleti, 202 Partner fidelity, 22, 23
M. scabrinodis, 127, 202 Passiflora, 41, 69, 147
M. schenki, 131 Passiflora oerstedii, 71
Myrmicaria Passifloraceae, 93
M. arachnoides, 105 Pathogens, 10
M. opaciventris, 106 Pegoscapus hoffmeyeri, 168
Myrsinaceae, 38 Pelopidas mathias, 203
Perizoma, 56
Perizoma affinitatum, 57
N Peromyscus
Narathura japonica, 130 P. leucopus, 89
Nectar robbers, 67 P. nudipes, 10
Nectar thief, 16 Petalomyrmex phylax, 66, 93, 187
Nemestrinidae, 39 Pheidole, 84
Nierembergia linariifolia, 45 P. bicornis, 38, 84
Niphanda fusca, 131 P. pallidula, 179
Noctuidae, 56 Phengaris, 120
Nothomyrmeciinae, 85 Philidris nagasau, 22
Nudina artaxidia, 106 Philoliche gulosa, 58
Nursery pollinators, 55 Phlox, 41
Nutrient sink, 123 Phyllanthaceae, 52
Nymphalidae, 40–42, 146 Phyllanthus, 53
Phyllobaenus, 38
Phytotelmata, 15
O Pieridae, 41
Oak-blue, 130 Piper, 84
Ocotillo, 183 P. obliquum, 84
Odontotermes formosanus, 144 P. urystachyum, 65, 84
Oecophylla smaragdina, 134, 172, 175, 178, 196 Piperaceae, 65
Oestrid flies, 19 Pitcher plants, 15
Oil-collecting bee, 38, 44 Pittosporaceae, 125
Onagraceae, 6 Plantaginaceae, 45
Ophrys, 180 Plant castration, 85
Opuntia imbricata, 23 Plant pathogens, 94
Orchidaceae, 180 Platyneura mayri, 178
Orchid bees, 157, 195 Platypodidae, 145
Orchids, 15, 22, 43, 67, 160, 195, 203 Plebejus argus, 124
Oxyanthus, 201 Poaceae, 116, 215
Oxyanthus pyriformis, 41, 201 Polemoniaceae, 69, 186
Polemonium viscosum, 69
Polygala vayredae, 185
P Polygalaceae, 185
Pachycereae, 51 Polygonaceae, 54
Pachycereus marginatus, 52 Polyommatinae, 213
Pachycondyla, 91 Polyommatus
P. apicalis, 91 P. bellargus, 132
P. harpax, 91 P. coridon, 132
P. luteola, 23 P. icarus, 128, 132
Index 231
Poplar aphid, 163 Rheum nobile, 54

Populus, 114 Rhinoceros, 19
Populus tremuloides, 210 Rhinoleucophenga myrmecophaga, 169
Pore cupola organs (PCOs), 121 Rhytidoponera, 4, 209
Prairie star, 197 R. metallica, 90
Pristomyrmex R. tasmaniensis, 67
P. punctatus, 130 R. victoriae, 163
P. pungens, 115 R. violacea, 90
Prodoxidae, 5, 46, 197 Riodinidae, 119
Prodoxus, 177, 179 Rubiaceae, 24, 41, 84, 86
Prodoxus decipiens, 179
Prosoeca, 58
Prosoeca ganglbaueri, 181 S
Prostomata, 66 Salix, 114
Proteaceae, 162 Salix exigua, 28
Protection hypothesis, 134 Salticidae, 73
Protectionists, 71 Sanguisorba officinalis, 217
Protectors, 59 Santalales, 134
Protozoa, 16 Saproxylic insects, 16
Pseudococcidae, 109 Saraca thaipingensis, 134
Pseudococus citriculus, 115 Satyrus fevula, 42
Pseudocopulation, 67 Saxifragaceae, 5, 197
Pseudomyrmex, 23, 71, 83, 185 Scale insects, 17, 24
P. ferruginea, 64, 68 Scarlet gilia, 185
P. ferrugineus, 94, 171 Schefflera vinosa, 137
P. flavicornis, 71 Sciaridae, 54
P. gracilis, 171 Scolytidae, 145
P. nigrocinctus, 71 Scolytus, 145
P. spinicola, 71 Scrophulariaceae, 44, 180
Psittacanthus robustus, 28 Selenothrips rubrocinctus, 197
Psyllids, 18 Semutophila saccharopa, 137
Pteromalidae, 167 Senita cactus, 51
Pterygodium cruciferum, 44 Senita moth, 51
Publilia Sensory trap, 92
P. concava, 107, 110 Sexual mimicry, 43
P. daphnoides, 90 Silene, 56
P. modesta, 110 S. ciliata, 56
Pyralidae, 51 S. dioica, 57
Silver-studded blue, 124
Slug, 162
Q Social parasite, 186
Qualea grandiflora, 169 Solanaceae, 45, 200
Quercus, 130 Solanum americanum, 200
Solenopsis
S. invicta, 162
R S. xyloni, 12
Rabdophaga salicisbrassicoides, 28 Sphingidae, 40, 59
Rafflesia, 40 Spiders, 72, 169
Rafflesia pricei, 40 Squamellaria, 22
Ranunculaceae, 53, 89 Stability, 20, 26
Rapala pheretima, 135 Staphylinidae, 10
Reciprocal antagonism, 51 Stenamma, 91
Red imported fire ant, 162 Sternorrhyncha, 105, 107, 109
Rediviva peringueyi, 44 Stratiotes aloides, 54
Restoration, 205 Stridulatory mechanisms, 122
232 Index
Supay chakras, 86 U
Sycoecinae, 167 Ulex, 89
Syconia, 47 Upiga virescens, 51
Symbiosis, 10
Syrphidae, 40, 57
V
Vespula pensylvanica, 158
T Viola, 41
Tapinoma sessile, 136 Vochysia
Technomyrmex albipes, 157, 175 V. thyrsoidea, 28
Tegeticula, 46, 174, 176 V. vismiaefolia, 83
T. cassandra, 176 Vochysiaceae, 28, 83, 169
T. intermedia, 176
T. yuccasella, 176
Temnothorax crassispinus, 92 W
Tentacle organs (TOs), 121 Warning signals, 146
Tentacular organ, 122 Wasps, 72, 121
Termites, 7 Water soldier, 54
Termitomyces, 143, 144 Waterstoniella, 168
Termitophiles, 11 Weevils, 38
Tetramorium Western yellowjacket wasp, 158
T. caespitum, 128 Widow species, 218
T. tsushimae, 92 Wild ginger, 162
Tetraponera aethiops, 93 Wood white, 41
Thisbe irenea, 130, Wyethia, 111
132, 169
Thrips, 38, 46, 197
Thymus praecox, 202 X
Thynnine wasp, 43 Xanthopan morganii, 15
Tillandsia, 84 Xyleborini, 145
Tolerance, 6 Xylocopa, 175
Trachymyrmex, 171 X. californica, 183
T. zeteki, 187 X. californica arizonensis, 183
Tramp ants, 162
Tramp species, 38, 157
Tranes, 38 Y
Tree weta, 6 Yellow loosestrife, 38
Treehoppers, 107, 110, 139 Yucca, 46, 49, 52
Tricholomataceae, 143 Y. baccata, 174
Trigona, 175 Y. filamentosa, 176, 179
Tripsacum dactyloides, 215 Yucca moths, 46
Trollius, 39 Yuccas, 46
Trollius-Chaetosticta, 46
Trophallaxis, 131
Trophobionts, 105 Z
Trophobiosis, 105 Zaluzianskaya microsiphon, 180

10.1007/978-3-319-58292-4 Mutualism and Insect Conservation PDF

Uploaded by

Copyright:

Available Formats

10.1007/978-3-319-58292-4 Mutualism and Insect Conservation PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10.1007/978-3-319-58292-4 Mutualism and Insect Conservation PDF

Uploaded by

Copyright:

Available Formats

Tim R.

Mutualisms and Insect

ISBN 978-3-319-58291-7 ISBN 978-3-319-58292-4 (eBook)

Library of Congress Control Number: 2017940274

© Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

Melbourne, Australia Tim R. New

Boucher DH (ed) (1985) The biology of mutualism. Croom Helm, London/Sydney

The following publishers and organisations have graciously granted permission to

Part I The Meaning and Dynamics of Mutualisms

Part II Lessons from Classic Mutualisms

Part III Environmental Impacts and Insect Conservation:

Mutualisms, essentially comprising some form of cooperation between species, and

Leigh EG (2010) The evolution of mutualism. J Evol Biol 23:2507–2528

Keywords Antagonism • Biological trading • Interactions • Interdependence •

The functional variety of ecological interactions between species continues to chal-

© Springer International Publishing AG 2017 3

The functional relationships between taxa in many interspecific associations that

Table 1.1 Hypothetical Reproduction of focal species

(as an insect group participating in a great variety of mutualisms) are purported to

Table 1.2 Some factors that Passage of symbionts from parent

The relative importance of all these influences on ant-aphid relationships is difficult

In short, many mutualistic environments or participating species are exploited or

One plant species One pollinator species

Several plant species One pollinator species

One plant species Several pollinator species

All plant species All pollinator species

co-adaptations may become increasingly intricate and specialised. However, ‘asym-

Keywords Ants • Coevolution • Coextinction • Mutualism • Partner choice •

2.1 Coextinction and Coevolution: Ecological Specialisation

© Springer International Publishing AG 2017 15

and by which it obtains nectar from flowers of Calathea crotalifera (Marantaceae)

No. extinct butterfly species

200 400 600 800 1000

No. extinct plant species

vulnerability through host obligations is almost impossible to confirm. Nevertheless,

2.2 Stability and Integration of Mutualisms

tive developments. Evolution of mutualism is thus promoted by each partner devel-

Mortality risk of myrmecophiles

2.3 Wider Community Stability

Several of the preceding examples of mutualisms are based on the exchange of a

Mutualism Only ants Total

A major problem in predicting outcomes of mutualisms, emphasised by Ringel

enhance chances of community persistence. Richer communities may indeed be

Anderson B, Johnson SD (2008) The geographical mosaic of coevolution in a plant-pollinator

ciations, other than occasional allusions to pesticide impacts. A parallel system, of

Keywords Coadaptation • Ecological specialization • Evolution • Figs •

3.1 Introduction: Pollination Systems

Pollination involves interactions between two predominant groups of organisms:

© Springer International Publishing AG 2017 37

involve transient visits by the pollinator with no more enduring interdependence

Fig. 3.1 Scheme to

Pollinator sharing amongst related orchids is apparently widespread. Thus, a

3.2 Obligate Pollination Mutualisms

Although many angiosperms are pollinated by associates as vectors, relatively few

decomposing flowers and inflorescences after pollination has occurred, a scenario in

3.2.1 Figs and Fig Wasps

3.2.2 Yuccas and Yucca Moths

Melbourne, Australia Tim R. New

Part III Environmental Impacts and Insect Conservation:

2.1 Coextinction and Coevolution: Ecological Specialisation

2.2 Stability and Integration of Mutualisms

2.3 Wider Community Stability

3.1 Introduction: Pollination Systems

3.2 Obligate Pollination Mutualisms

3.2.1 Figs and Fig Wasps

3.2.2 Yuccas and Yucca Moths

3.2.3 Other Examples

3.3 Nursery Pollination

3.4 Broader Perspective

4.1 Introduction: Ant-Plant Interactions

4.2 Extrafloral Nectaries

Fig. 4.3 Ants as anti-

4.3 Food Bodies and Domatia

4.5 Ant Gardens

4.7 Ants, Plants and Fungi

5.1 Introduction: Trophobiotic Associations

5.2 Ant-Hemiptera Mutualisms

6.1 Introduction: Ants and Lycaenid Butterflies

7.1 Introduction: The Remaining Variety

7.2 Termites and Fungi

7.3 Beetles and Fungi

7.4 Mullerian Mimicry

8.2 Impacts on Mutualisms