Continental-Scale Tectonic Controls of Biogeography and Ecology

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/279241320
Continental-Scale Tectonic Controls of Biogeography and Ecology
Chapter · August 2011

DOI: 10.1525/california/9780520268685.003.0009
CITATIONS READS
55 218
2 authors:
Flávio C. T. Lima Alexande C. Ribeiro

University of Campinas Universidade Federal de Mato Grosso (UFMT)
86 PUBLICATIONS 1,124 CITATIONS 26 PUBLICATIONS 753 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Fishes from the Vale do Juruá, AC: survey of the ichthyofauna of the Juruá, Moa and Liberdade Rivers View project
Systematics and speciation processes of the genus Hypostomus View project
All content following this page was uploaded by Alexande C. Ribeiro on 16 July 2015.
The user has requested enhancement of the downloaded file.

NINE
Continental-Scale Tectonic Controls

of Biogeography and Ecology
FLÁVIO C. T. LI MA and ALEXAN DR E C. R I B E I RO
Fish biogeography in the Neotropical region has been a sub- the South American freshwater fish fauna were facing when
ject of increasing interest in the last few years. Less than thirty interpreting historical vicariant events, Lundberg and col-
years ago, Weitzman and Weitzman (1982) could still claim leagues (1998) provided a synthesis of the geological evolu-
that “ichthyologists have not as yet contributed substan- tion of South American river basins, with an emphasis on the
tive results to the combined studies of biogeography, species hydrogeographic changes in the river basins draining the sedi-
diversification, and evolution of higher fish taxa within South mentary basins adjacent to the Andean cordillera during the
America.” Since then, however, a growing body of information Cenozoic.
on fish taxonomy, distribution, and phylogenetic relation- The synthesis by Lundberg and colleagues (1998) was influ-
ships has opened a path for a substantial improvement in our ential in changing the way South American fish biogeography
understanding of the subject. The first major move toward an is interpreted and inspired new approaches to its comprehen-
adequate assessment of biogeographical patterns presented sion. Examples of recent papers examining broad patterns of
by South American freshwater fishes was the paper by Vari fish biogeography in South America that incorporated views
(1988), which, based on an extensive revision of curimatid on the evolution of hydrogeographic systems expressed by
systematics, discussed in detail freshwater fish biogeographic Lundberg and colleagues (1998) were Lovejoy and colleagues
patterns across South America north of Patagonia. Both Vari (2006), Albert and colleagues (2006), and Hubert and Renno
(1988) and Vari and Weitzman (1990) pointed out the need for (2006). Lovejoy and colleagues (2006) addressed primary
increasing our knowledge of species-level systematics, distri- marine fish groups with representatives in freshwaters in
bution information, and phylogenetic relationships of South South America and concluded that their establishment in the
American freshwater fishes as a necessary step toward the region is probably ascribable to marine incursions that have
formulation of adequate hypotheses about their historical bio- taken place during the Miocene. Albert, Lovejoy, and col-
geography. Subsequent authors have followed those precepts leagues (2006) reviewed the fish clades present in trans-
and provided discussions on possible vicariant events that have Andean river basins and their relationships with cis-Andean
affected some other fish clades—for example, Schaefer (1990) clades, thus expanding the views earlier advanced by Lundberg
on Scoloplax (Scoloplacidae), Schaefer (1997) on Otocinclus and colleagues (1986, 1988), Lundberg and Chernoff (1992),
(Loricariidae), Reis (1998b) on Lepthoplosternum (Callichthy- and Vari (1988) on a substantial vicariant event elicited by the
idae), and Costa (1996) on Simpsonichthys (Rivulidae). formation of a new drainage limit by the uplift of the East-
A widely held assumption in freshwater fish biogeography is ern Andes and the Caribbean coastal cordillera during the
that a good knowledge of the geological evolution of the river Miocene. Hubert and Renno (2006) undertook a parsimony
basins is fundamental to understand the history of the vicari- analysis of endemicity (PAE) using a large database of South
ant events responsible for the current distribution patterns of American characiforms, aiming to identify the relationships
freshwater taxa. With that view in mind, it was evident that among hypothesized areas of fish endemism in the continent,
the meager geological information used to interpret biogeo- in which these authors formally proposed four major hypo-
graphical patterns identified in the aforementioned studies thesis for diversification of the South American biota, two of
was not sufficiently detailed to fully appreciate the impact of which, the “palaeogeography hypothesis” and the “hydrogeo-
the dynamics of geological processes that were thought to be logy hypothesis,” involve changes in river-drainage configura-
responsible for configuring distribution patterns. Prompted tion as the causal factors explaining current fish distributions.
by the perceived difficulties that ichthyologists working with The thorough and detailed review by Lundberg and col-
leagues (1998), however, did not address the hydrogeographic
evolution of the South American river basins that drain
Historical Biogeography of Neotropical Freshwater Fishes, edited by
James S. Albert and Roberto E. Reis. Copyright © 2011 by The Regents shields, and several large basins were excluded from their his-
of the University of California. All rights of reproduction in any form torical narrative. Even though much more stable than the sedi-
reserved. mentary basins adjacent to the Andean orogenic belt, these
145
c09.indd 145 4/7/2011 7:21:11 PM

areas cover an extensive portion of eastern South America and the Cenozoic. Finally, we present evidence for possible events
harbor considerable portions of the Amazon, Orinoco, Guiana, of faunal exchanges driven by river captures among basins of
La Plata, São Francisco, and coastal southeastern river systems. both the Guiana and Brazilian shields as indicated by ich-
That gap was partially filled by A. Ribeiro (2006), who discussed thyofaunal data. We confine our discussion to cis-Andean
extensively the geological events that modified drainage sys- river basins of the Amazon-Orinoco and Guianas (i.e., from
tems during the Mesozoic and Cenozoic and were related to the Cuyuni River on the Venezuela/Guiana border to the rio
several putative vicariant events that have taken place between Araguari in Amapá state in Brazil) and Paraguay river systems,
the coastal river basins of eastern Brazil and the drainages from though the discussion on the foreland-basin relationships has
the adjacent upland crystalline shield area. a wider implication for the biogeography of the La Plata Basin.
Recent popularization of high-resolution images of topogra- The degree of endemism among these isolated river basins is
phy of the world obtained by radar interferometry (the Shuttle variable. Endemic taxa are usually used as evidence for a unique
Radar Topography Mission, or SRTM, available at http://www2. biogeographic history and contribute to the generally accepted
jpl.nasa.gov/srtm/) provides an increasing interest in the geog- view that major basins correspond to major areas of ende-
raphy of South America and its putative interrelationships with mism. This view is so deeply rooted in ichthyologists’ minds
faunal distribution patterns. In South America, a brief examina- that even almost undistinguishable morphotypes are usually
tion of these wonderful maps provides a rapid view of the topo- arbitrarily taken as distinct species under the prerogative of
graphic relief at a continental scale. It also provides outstand- being isolated in different basins instead of being recognized as
ing examples of tectonically imposed landscapes, offering a widespread taxa. The presence of shared species between two
friendly and didactic overview of geological processes involved or more hydrogeographic systems is traditionally attributed to
in molding landscapes of South America. The first conclusion two main causes: (1) species that were present in a paleo-area
taken from analyzing such images is that the South American encompassing both basins before the geological/historical pro-
continent is nothing but a huge plateau fragment of Gond- cess that configured the present observed basin architecture
wanaland around which a younger orogenic belt (the Andean (a simple vicariance model), or (2) species that arose in one of
chain) and lowland areas evolved mostly as a result of Mesozoic the basins and occur subsequently in other basins by dispersal
and Cenozoic tectonic processes responsible for modeling (a simple dispersal model). None of these simplistic assump-
most of the present-day landscape configuration (Figure 9.1). tions, however, explain precisely the processes by which the
Recently, extensive fish collecting undertaken in the upper same species can occur in two or more isolated basins.
rio Xingu and upper rio Tapajós systems by ichthyologists The present architecture of the Brazilian Shield basins is
from several Brazilian institutions (Museu de Zoologia da Uni- strongly associated with the mega tectonic processes that cul-
versidade de São Paulo; Museu de Ciências e Tecnologia da minated in the breakup of Gondwanaland. The present-day
Pontifícia Universidade Católica do Rio Grande do Sul; Museu divides between major basins of the Brazilian Shield were cer-
Nacional, Universidade Federal do Rio de Janeiro) allowed a tainly well developed at the end of the Cretaceous (Cox 1989;
glimpse of fish diversity in the previously virtually unknown Potter 1997; A. Ribeiro 2006). The present-day shared species
upper portions of those river drainages. It confirmed earlier between those basins cannot be explained by a simple cladoge-
impressions (e.g., Jégu et al. 1991; Araújo-Lima and Goulding netic event of such antiquity. If an “ancestral stock” occurring
1997; G. Santos and Ferreira 1999; Jégu and Keith 1999) that in these basins can be still recognized, it does not comprise the
the ichthyofauna of shield rivers draining to the Amazon is present species-level similarity among basins, but deeper levels
more similar to that of other shield rivers than to that of the of relationships among higher inclusive taxa.
western and central portions of the Amazon Basin. In this chapter, we propose a general model for explain-
Thanks to the advances described here, of our knowledge ing the process by which “species-level” similarity can occur
both of freshwater fish systematics and distributions, and of among isolated basins of the upland shields and lowland fore-
the geomorphological history of South America, we are now land basin areas. These patterns are strongly associated with
in a position to formulate adequate hypotheses that address the main tectonic processes affecting the South American con-
the broad biogeographical patterns found in this fauna. In this tinent. As we shall demonstrate, this model is underpinned by
chapter we aim to provide evidence for the recognition of two the recent advances in the knowledge of the tectonic behavior
major “biogeographical provinces” in central and northern of the South American Platform. Although we do not pretend
cis-Andean South America, based on major historical and our model to be an explanation for the distribution patterns
ecological constraints: lowland areas versus upland shield of every fish taxon in northern cis-Andean South America,
areas. Our definition of lowland areas is mostly geographically we expect that, as more taxonomic revisions with good geo-
restricted to, but not strictly confined to, the limits of foreland graphical coverage and phylogenetic hypotheses become avail-
basins, and includes intracratonic/sedimentary basins of the able, the major distribution patterns for fishes in northern cis-
lower Amazon valley and shield deeps such as the Araguaia Andean South America described in the present chapter will
plain and the Takutu rift, situated below 250 meters above sea prove to be considerably more widespread among several Neo-
level (asl). Upland shield areas are defined as those underlain tropical fish clades.
by the ancient Brazilian and Guiana shield areas, generally
lying above 250 meters asl, though some areas below this Materials and Methods
altitude, such as the lower rio Tocantins and lower rio Xingu,
possess outcropping basement (Figure 9.2). We analyzed the distribution patterns of several monophyletic
For reasons explained at the end of the discussion, we taxa, mostly species but also species groups and genera, when
departed from traditional approaches in biogeographical analy- well-corroborated hypotheses of relationships were available.
sis that employ “areas of endemism” as one of their basic tenets. Taxa listed in this chapter as examples for either “lowland” or
We also discuss the relationships of the foreland basins of the “shield” patterns belong to groups recently subjected to taxo-
Orinoco, Amazon, and Paraguay river drainages and point out nomical scrutiny, particularly those where detailed distribu-
their importance as areas of exchange of aquatic biota during tion information, based on extensive survey of specimens in
146 C O N T I N E N TA L A N A LYS I S
c09.indd 146 4/7/2011 7:21:11 PM

Physical map of South America based on radar interferomety (SRTM-NASA) showing the main topographic features and major
F I G U R E 9.1
tectonic structures discussed in this chapter. Limits of the Amazon Foreland basin following Baby et al. (2005).
fish collections, was available. The selection of taxa for which collection, specimens of which were checked to confirm their
an adequate knowledge of both the taxonomy and the geo- identifications, plus reliable literature records—for Abramites
graphic range is available is an obviously essential prerequisite hypselonotus, Vari and Williams (1987) and Taphorn (1992);
before its incorporation into any biogeographical hypothesis for Roeboexodon guyanensis (note: we do not follow the nomen-
(Brown et al. 1996). Some genera poorly known taxonomically clatural suggestion for the change of the name of this species
(e.g., Mylossoma, Hypoptopoma) are also mentioned because predicated by Lucena and Lucinda, in F. Lima et al. 2003),
they constitute monophyletic taxa, with well-known distri- Lucena and Lucinda (2004), and Planquette et al. (1996); for
bution ranges, factors that allow a biogeographical interpre- Anostomus ternetzi, Winterbottom (1980), G. Santos and Jégu
tation of their distribution patterns. In the maps presented, (1989), and Sidlauskas and Santos (2005); for Leporinus brun-
localities plotted were based on material deposited at MZUSP neus, Chernoff et al. (1991) and Santos and Jégu (1996); and
C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 147
c09.indd 147 4/7/2011 7:21:11 PM

F I G U R E 9.2 Map of tropical and subtropical South America showing extent of lowland areas.
for Curimatella meyeri, Vari (1992b). Localities for Thayeria boe- continental portion of the South American plate not affected
hlkei were based on an extensive survey of material conducted by the Caribbean and Andean orogenic zones and is consti-
by the first author, along with C. R. Moreira, as part of revi- tuted by the Brazilian Platform and the Patagonian Platform
sionary studies on the genus Thayeria. (Almeida et al. 2000).The geological structure of the South
Distribution patterns are summarized in Tables 9.1 and 9.2, American Platform can be synthetically described as a Gond-
which include the list of taxa used for the interpretation of wanaland fragment that includes a set of five Arquean cratons
the distribution patterns, with a summary of their distribution (Amazonian, São Francisco, Rio de La Plata, São Luiz, and Luiz
ranges. These data were coded into a data matrix for absence Alves) (Cordani et al. 2000) surrounded by ancient Precam-
(0) and presence (1) of taxa, and a parsimony analysis was brian orogenic belts (both consisting of the crystalline shields)
carried out using the heuristic option of the software PAST, and associated sedimentary cover. The South American Plat-
version 1.90 (Hammer et al. 2001) for hypothesizing hierar- form interacts with the Nazca Plate to the west, creating the
chical interrelationships among areas, following the methods Andean orogenic belt.
proposed by B. Rosen and Smith (1988). Most of the South American platform rocky basement
resulted from a set of paleocontinental amalgamations devel-
oped in response to the convergence of the São Francisco,
Geological Background Congo, and Rio de La Plata cratons during the Neoprotero-
zoic to Early Paleozoic (between 0.9 and 0.5 Ga) originating
STRUCTURAL GEOLOGY AND TECTONIC SETTINGS
the Eastern Gondwanaland supercontinent in the so-called
Distribution and biogeographical patterns discussed in this Brazilian/Pan-African orogenic cycle (Trouw et al. 2000;
chapter take place in the central area of the ancient upland Almeida et al. 2000).
Brazilian crystalline shield and adjacencies, developed as a Within the Brazilian Platform, therefore, shields are consti-
set of lowland areas, in which large river systems evolved as tuted by rocks of the cratons and neighboring ancient oro-
a result of major landscape rearrangements driven by global genic belts that resulted mostly from the Brazilian/Pan-African
tectonic processes acting along most of the South American cycle. These sets of Precambrian rocks present a structural
continent since the Cretaceous period (Potter 1997). Under- inheritance of their collisional origin. Among one of the most
standing this complex history cannot be achieved without conspicuous is the presence of a complex system of Precam-
considering some major aspects of the geological structure of brian rift and shear zones. This complex system of ancient rifts
the South American continent and the way this ancient struc- behaves as weakness zones, more susceptible to undergoing
ture responds to more recent global tectonic forces. Interaction deformations due to tectonic reactivation events (Saadi 1993;
between two connected elements—ancient geological struc- Saadi et al. 2002; Riccomini and Assumpção 1999). Since the
ture and its behavior under present-day tectonics settings—are Gondwanaland breakup (culminating approximately 90 MY)
key factors in elaborating a scenario on the biogeographic his- reactivations along this set of Precambrian fracture zones have
tory of South American freshwater fish fauna. been driving the tectonic behavior of the entire platform. This
Most of the South American Continent consists of the South analysis is based on the concept of resurgent tectonics (Suguio
American Platform (Figure 9.3), which is defined as the stable 2001), in which ancient structures (faults and shear zones)
c09.indd 148 4/7/2011 7:21:11 PM

TABLE 9.1
Examples of Fish Taxa Presenting Shield Distribution Patterns
Taxon OS GU SU FG AN TO XU TA MS Source
Characiformes
Anostomidae
Winterbottom 1980; Santos and

Anostomus ternetzi X X X X X X X
Jégu 1996; MZUSP
Pseudanos gracilis X X X
Pseudanos irinae X X
Winterbottom 1980; G. Santos and
Gnathodolus bidens X X
Jégu 1987
Synaptolaemus Winterbottom 1980; G. Santos and
X X X X X
cingulatus Jégu 1987
Sartor spp. X X X G. Santos and Jégu 1987
Laemolyta fernandezi X X X Mautari and Menezes 2006
Chernoff et al. 1991; Santos and
Leporinus brunneus X X X X
Jégu 1996
Leporinus pachycheilus X X X G. Santos et al. 1996
Hemiodontidae
Argonectes robertsi X X X Langeani 1998
Erythrinidae
Mattox et al. 2006; Planquette
Hoplias aimara X X X X X X X X
et al. 1996
Cynodontidae
Hydrolycus tatauaia X X X X X X X Toledo-Piza et al. 1999
Hydrolycus armatus X X X X X X X Toledo-Piza et al. 1999
Alestidae
Chalceus epakros X X X X X X Zanata and Toledo-Piza 2004
Chalceus Zanata and Toledo-Piza 2004;
X X X X X
macrolepidotus Planquette et al. 1996
Characidae
Serrasalminae
Acnodon spp. X X X X X X X Jégu and Santos 1990; MZUSP
Jégu, Keith, and Belmont-Jégu 2002;
Tometes spp. X X X X X X X Jégu, Santos and Belmont-Jégu
2002; MZUSP
Jégu and Santos 1988; Jégu et al.
Mylesinus spp. X X
1992
Myleus spp. X X X X X X X X X Jégu and Santos 2002
Bryconinae
Brycon falcatus X X X X X X X X X Lima 2001
Characidae incertae sedis
Roeboexodon Lucena and Lucinda 2004;
X X X X X X
guyanensis Planquette et al. 1996; MZUSP
Bryconexodon spp. X X Jégu et al. 1991
Hyphessobrycon
X X X MZUSP
moniliger
Jupiaba acanthogaster X X X X MZUSP
Moenkhausia
X X MZUSP
phaeonota
C. R. Moreira and F. C. T. Lima,
Thayeria boehlkei X X X
unpublished data
Siluriformes
Loricariidae
Hypoptopomatinae
Parotocinclus spp. X X X X X X X Schaefer and Provenzano 1993
Loricariinae
Harttia spp. Le Bail et al. 2000; Boeseman 1971,
(including X X X X X X X X 1976; Provenzano et al. 2005;
Cteniloricaria spp.) Rapp Py-Daniel and Oliveira 2001
Metaloricaria Isbrücker and Nijssen 1982;
X X X
paucidens E. Ferreira 1993
Hypostominae
Fisch-Muller 2003; Werneke, Sabaj,
Baryancistrus spp. X X X X X
et al. 2005; Lujan et al. 2009
c09.indd 149 4/7/2011 7:21:11 PM

TABLE 9.1 (continued)
Taxon OS GU SU FG AN TO XU TA MS Source
Armbruster 2002; Armbruster et al.

Hypancistrus spp. X X X X 2007; Isbrücker and Nijssen 1991;
Lechner et al. 2005a, 2005b
Isbrücker and Nijssen 1989;
Leporacanthicus spp. X X X X
Isbrücker et al. 1993
Boeseman 1982; Le Bail et al. 2000;
Lithoxus spp. X X X X X
E. Ferreira 1993; Lujan 2008
Rapp Py-Daniel 1989; Rapp
Parancistrus spp. X X
Py-Daniel and Zuanon 2005
Burgess 1994; Isbrücker and Nijssen
Scobinancistrus spp. X X X
1989; Fisch-Muller 2003
Pimelodidae
Sorubim
X X X Littman 2007; MZUSP
trigonocephalus
Auchenipteridae
Tocantisia piresi X X X X Mees 1984; MZUSP
Doradidae
Doras spp. (D.
X X X X X Sabaj-Pérez and Birindelli 2008
carinatus group)
Gymnotiformes
Sternopygidae
Archolaemus blax X X X X X Schwassman and Carvalho 1985
Apteronotidae
Megadontognathus
X X X Campos-da-Paz 1999
spp.
Perciformes
Cichlidae
Le Bail et al. 2000; Kullander and
Guianacara spp. X X X X X Nijssen 1989; López-Fernández
et al. 2006
Kullander 1988; Zuanon and
Teleocichla spp. X X X X
Sazima 2002; MZUSP
Gosse 1971; Le Bail et al. 2000;
Retroculus spp. X X X X
MZUSP
Sciaenidae
Petilipinnis grunniens X X X X X Casatti 2002
Pachyurus junki X X X X X Casatti 2001
NOTE : OS, Orinoco shield tributaries; GU, Guyana; SU, Suriname; FG, French Guiana; AN, Amazon northern tributaries; TO, Tocantins; XU, Xingu;
TA, Tapajós, MS, Madeira shield tributaries.
become reactivated subsequently by more recent tectonic when the streams suffer an abrupt deviation as a consequence
events. The evolution of the continental paleodrainage and of the relative movement between rifted blocks. Alternatively,
relief is strongly controlled by resurgent tectonics (Saadi it occurs by differential erosion, because deformation in the
et al. 2005). landscape promotes the adjustment of the drainage to a new
Among large hydrogeographic basins of South America, base level, causing streams on lowered blocks, with a steeper
the Upper Paraná, Uruguay, São Francisco, and large Brazilian gradient and, consequently, more energy, to extend their val-
coastal rivers such as the rio Doce and rio Jequitinhonha show leys headward as a result of erosion, eventually breaking down
strong evidence of acquiring most of their present courses as a the divide and capturing part or all of the drainage of adjacent
result of tectonics associated with the Gondwanaland breakup, slower streams (Tarbuck and Lutgens 2002).
which configured the major shape of their drainage basins If, on one hand, the tectonic activity associated with the
(K. Cox 1989; Potter 1997; A. Ribeiro 2006). Within these evolution of the eastern divergent rifted margin of the South
basins, recent tectonic reactivation events have constantly American platform affects distribution and biogeography of
promoted drainage rearrangements and are the main cause of the central-eastern Brazilian shield fish fauna (A. Ribeiro 2006),
headwater captures between adjacent basins from the Paleo- the tectonic evolution of the Andean cordilleras, on the other,
gene to the present. Examples of large and small drainage devi- is the main geological process affecting drainage dynamics and
ations driven by Paleogene and Neogene tectonic reactivations consequently, fish fauna biogeography of western cis-Andean
on basin divides have been more extensively reported recently South America.
(Ab’Sáber 1957, 1998; Cobbold et al. 2001; Brito-Neves et al. Within the context of Andean tectonics, evolution of fore-
2004; Modenesi-Guattieri et al. 2002; A. Ribeiro 2006; A. land basins is a main point in understanding cis-Andean low-
Ribeiro et al. 2006; Menezes et al. 2008). Stream capture or land fish faunal distribution patterns. Foreland basin systems
piracy can operate basically in two different ways in tectoni- develop as a result of flexural warping of the lithosphere in
cally active areas. It can be a direct effect of tectonic stress, response to supralithospheric and sublithospheric orogenic
c09.indd 150 4/7/2011 7:21:11 PM

TABLE 9.2
Fish Species with Lowland Distributional Patterns
Actinopterygii Class
Taxon OL WA ES TL PP NC Source
Osteoglossiformes
Osteoglossidae
Kanazawa 1966; Jégu and Keith 1999;
Osteoglossum bicirrhosum X X X X X
Maldonado-Ocampo et al. 2008
Arapaimatidae
Arapaima gigas X X X
Clupeiformes
Engraulididae
Jurengraulis juruensis X Whitehead et al. 1988
Anchoviella jamesi X X Whitehead et al. 1988
Pristigasteridae
Whitehead 1985; G. Santos et al. 2004;
Pellona castelnaeana X X
Melo et al. 2005
Whitehead 1985; Menezes and Pinna
Pristigaster cayana X X
2000; G. Santos et al. 2004
Pristigaster whiteheadi X Menezes and de Pinna 2000
Characiformes
Anostomidae
Abramites hypselonotus X X X X Vari and Williams 1987
Leporinus striatus X X X Britski and Garavello 1980; Taphorn 1992
Leporinus trifasciatus X X
Rhytiodus spp. X
Schizodon fasciatus X
Curimatidae
Cyphocharax spiluropsis X Vari 1992a
Curimata aspera and C. cerasina X X Vari 1989a
Curimatella dorsalis X X X Vari 1992b
Curimatella meyeri X Vari 1992b
Potamorhina altamazonica X X Vari 1984
Potamorhina spp. X X X Vari 1984
Psectrogaster curviventris X X Vari 1989b
Steindachnerina bimaculata X X Vari 1991
Steindachnerina guentheri X X Vari 1991
Steindachnerina spp.
X X X Vari 1991
(S. conspersa group)
Gasteropelecidae
Gasteropelecus sternicla X X X X Weitzman 1960; Britski et al. 2007
Thoracocharax spp. X X X X Weitzman 1960
Alestidae
Chalceus erythrurus X Zanata and Toledo-Piza 2004
Cynodontidae
Hydrolycus scomberoides X Toledo-Piza et al. 1999
Characidae
Bryconinae
Brycon amazonicus X X F. Lima 2001
Brycon hilarii X X F. Lima 2001
Clupeacharacinae
Clupeacharax anchovioides X X X F. Lima 2003
Cheirodontinae
Odontostilbe fugitiva X Bührnheim and Malabarba 2006
Serrasalminae
Colossoma macropomum X X Araújo-Lima and Goulding 1997
Jégu and Keith 1999; Machado-Allison
Mylossoma spp. X X X X
and Fink 1995; Britski et al. 2007
Jégu and Keith 1999; Machado-Allison
Piaractus spp. X X X X
and Fink 1995; Britski et al. 2007
Lowe-McConnnell 1964; W. Fink 1993;
Pygocentrus nattereri X X X X X
Jégu and Keith 1999
Serrasalmus elongatus X Jégu and Keith 1999
Stethaprioninae
Stethaprion spp. X Reis 1989
Brachychalcinus spp. X X Reis 1989, 1998b
c09.indd 151 4/7/2011 7:21:11 PM

TABLE 9 . 2 (continued)
Stevardiinae
Gephyrocharax spp. X X X Weitzman 2003; MZUSP
Incertae sedis
Astyanacinus moorii X X MZUSP
Ctenobrycon spp. X X X X X X
Creagrutus barrigai X Vari and Harold 2001
Creagrutus cochui X Vari and Harold 2001
Engraulisoma taeniatum X X X F. Lima et al. 2003; Taphorn 1992
Gymnocorymbus spp. X X X X F. Lima et al. 2003
Hemigrammus barrigonae and
X X X
H. ulreyi
Leptagoniates pi X
Leptagoniates steindachneri X
Markiana spp. X X
Paracheirodon innesi X Weitzman and Fink 1983; Kullander 1986
Paragoniates alburnus X
Parecbasis cyclolepis X MZUSP
Prionobrama spp. X X X
Siluriformes
Cetopsidae
Cetopsis candiru X Vari et al. 2005
Cetopsis coecutiens X X X Vari et al. 2005
Trichomycteridae
Megalocentor echthrus X X Pinna and Britski 1991
Callichthyidae
Leptoplosternum spp. X X Reis 1998b; Reis and Kaefer 2005
Dianema spp. X
Corydoras spp. (C. reynoldsi
X Britto and Lima 2003
group)
Loricariidae
Hypoptopomatinae
Hypoptopoma spp. X X X X X
Otocinclus macrospilus X Schaefer 1997
Otocinclus huaorani X X Schaefer 1997
Otocinclus vestitus X X Schaefer 1997
Otocinclus vittatus X X X X Schaefer 1997
Loricariinae
Apistoloricaria spp. X X
Crossoloricaria spp.* X
Isbrücker and Nijssen 1978a; Taphorn and
Lamontichthys spp.* X X
Lilyestrom 1984b
Planiloricaria cryptodon X Isbrücker and Nijssen 1986
Pseudohemiodon spp. X X X
Hypostominae
Aphanotolurus spp. X X Armbruster 1998b
Hypostomus pyrineusi X Armbruster 2003
Panaque spp. (Panaque Schaefer and Stewart 1993; Chockley and
X X
dentex group) Armbruster 2002
Peckoltia bachi X X Armbruster 2008
Peckoltia brevis X Armbruster 2008
Pseudorinelepis genibarbis X Armbruster and Hardman 1999
Pterygoplichthys pardalis X Weber 1992
Pterygoplichthys punctatus X Weber 1992; Armbruster and Page 2006
Pimelodidae
Brachyplatystoma spp.
X X X X X
(B. filamentosum excepted)
Callophysus macropterus X X
Cheirocerus spp.* X Stewart and Pavlik 1985
Pimelodina flavipinnis X X Stewart 1986
Platynematichthys notatus X X X
Lundberg and Parisi 2002; Parisi et al.
Propimelodus spp. X X
2006; Rocha et al. 2007
c09.indd 152 4/7/2011 7:21:12 PM

TABLE 9 . 2 (continued)
Sorubim elongatus X X Littmann 2007

Sorubim lima X X X Littmann 2007
Sorubim maniradii X Littmann 2007
Auchenipteridae
Auchenipterichthys
X X X Ferraris et al. 2005
coracoideus
Auchenipterus ambyacus X X Ferraris and Vari 1999
Auchenipterus brachyurus X Ferraris and Vari 1999
Entomocorus spp. X X X Vari and Ferraris 1998
Epapterus spp. X X X Reis and Borges 2006
Doradidae
Doras phlyzakion and
X X Sabaj Pérez and Birindelli 2008
D. zuanoni
Leptodoras acipenserinus X Sabaj 2005
Lithodoras dorsalis X
Megalodoras spp. X X
Oxydoras spp. X X X X
Pterodoras spp. X X X X
Rhynchodoras woodsi X Birindelli et al. 2007
Aspredinidae
Hoplomyzon spp.* X X
Xyliphius spp.* X X X
Gymnotiformes
Apteronotidae
Adontosternarchus balaenops X Mago-Leccia et al. 1985
Sternarchorhamphus muelleri X Campos-da-Paz 1995
Batrachoidiformes
Batrachoididae
Thalassophryne nattereri X Collette 1966
Beloniformes
Belonidae
Pseudotylosurus angusticeps X X Collette 1974
Perciformes
Cichlidae
Bujurquina spp. X X X Kullander 1986
Cichla monoculus + C.
X X Kullander and Ferreira 2006
pleiozona + C. kelberi
Cichlasoma amazonarum X Kullander 1983
Laetacara flavilabris X Kullander 1986
Sciaenidae
Plagioscion montei X Casatti 2005
Pleuronectiformes
Achiridae
Apionichthtys nattereri X Ramos 2003b
Tetraodontiformes
Tetraodontidae
Colomesus asellus X X X X Tyler 1964
Sarcopterygii
Ceratodontiformes
Lepidosirenidae
Planquette et al. 1996; Arratia 2003;
Lepidosiren paradoxa X X X X
Maldonado-Ocampo et al. 2008
NOTE : OL, Orinoco lowlands; WA, Western Amazon; ES, Essequibo; TL, Tocantins lowlands; PP, Paraná/Paraguay lowlands; NC, Northern coastal plains.
wedging. Lithospheric flexure under static loads generates basins. These processes were the result of the migration of the
down-bending flexure proximal to the orogen, which migrates tectonic deformations eastward and other mechanisms such
as the load advances (Uba et al. 2006). Foreland basins are thus as megafans dynamics (Horton and DeCelles 2001; Wilkin-
elongated, tectonically imposed lowlands, located between son et al. 2006). Foreland basins can be described as sets of
upland areas of the Andean chain in the west and the Brazil- “expanding lowlands” into which adjacent uplands become
ian Shield in the east. This system of interconnected lowland incorporated as the tectonic load advances eastward. An
areas suffered constant drainage rearrangements, translated by example of such dynamics is exemplified by the origin of the
ephemeral contact and isolation among neighboring drainage Pantanal Wetland, a tectonic depression developed thanks to
c09.indd 153 4/7/2011 7:21:12 PM

F I G U R E 9.3 Major tectonic provinces of South American Platform. I, South American Platform; II, Patagonian massif; III, Andean orogenic
belt; IV, foreland basins; AM, Amazon craton; SL, São Luis craton; SF, São Francisco craton; LA, Luiz Alves craton; RP, Rio de la Plata craton;
B, Borborema province; T, Tocantins province; M, Mantiqueira province; DF, Dom Feliciano belt. (Modified from Cordani et al. 2000 and
Cordani and Sato 1999).
tectonic reactivations of Precambrian faults along the Trans- mation. Marine incursions during the Miocene also are known
brasiliano lineament approximately 2.5 MY (Soares at al. 1998; from several intracontinental basins in South America. In the
Assine 2004). This system of interconnected foreland basins, Amazon Basin, several short marine incursions appeared in
the Chaco and Pantanal, formed during the late Cenozoic in the Miocene Solimões and Pebas formations in Brazil, Peru,
response to Nazca–South American plate convergence and its Ecuador, Colombia, and Venezuela. South of the Chaco fore-
related eastward interaction with the Brazilian shield (Uba land basin, the Miocene Paranense Sea covered a wide area in
et al. 2006). According to Assine (2004), during the Cretaceous northern Argentina and Uruguay (Hulka et al. 2006).
and afterward the western border of the upland upper Paraná Despite uncertainty concerning the extension of seaways
Basin extended westward to the present-day Pantanal Wetland. in South America, it is reasonable to consider that the latest
The area represented the natural extension of the present-day events represent a starting point for biogeographic analysis of
Brazilian crystalline shield to the west-southwest, acting as strictly freshwater fishes inhabiting lowland foreland drainage
divide between drainages of the upper Paraná and Chaco basins. Lundberg and colleague (1998, 38, fig. 18) illustrate the
basins. During the last compressive event along the Andean most recent major marine seaway dating from the Late Ter-
belt (~2.5 MY) flexural subsidence associated with fault reac- tiary (c. 11.8–10 MY), named as the Paranan Sea in the south
tivation on its borders originated the Pantanal Wetland (Ass- and the Pebasian Sea in the north. According to V. Ramos and
ine 2004) and configured the present-day divide between the Aleman (2000), maximum flexural subsidence resulted from
western margins of the upper Paraná and upper Paraguay. The rapid tectonic loading of the Andes between 15 and 13 MY
set of foreland basins present along the Andean slope are thus was responsible for the marine transgression of the Paranense
dynamic landscapes capturing drainages from adjacent upland Sea that invaded most of the Andean foothills between 40º S
shield rivers and promoting hydrological connections to and the Maracaibo area. According to the same authors, this
each other. marine transgression of the Paranense Sea could be correlated
South American foreland basins are also areas of constant and connected with Amazonian transgressions of the middle
marine incursions (Lovejoy et al. 2006). Subsidence of fore- Miocene (12 MY) (Figure 9.4).
land basins combined with eustatic sea-level rises promotes Along the core area of the Brazilian shield there are also
marine incursions along several lowland areas of foreland enclaves of tectonically developed lowland areas or depres-
basins adjacent to the Andean chain (Lundberg et al. 1998). sions that are not directly related to the evolution of fore-
Between the Oligocene and Late Miocene, shallow restricted land basins but to the constant tectonic reactivation events
marine incursions transgressed into southeastern Bolivia and undergone by the complex system of Precambrian faults
are represented there by the Middle-Late Miocene Yecua For- of the crystalline basement. Examples of such areas are the
c09.indd 154 4/7/2011 7:21:12 PM

Guiana
Shield
Tet
Amazon Amazon arm
hys
Sea
rma
Brazilian shield
ea
s eS
nen
ra
Pa
0 1000 km
F I G U R E 9.4 Possible extension of the Miocene (12 MY) transgression of the Paranense Sea (modified from V. Ramos and Aleman 2000).
Araguaia and Tocantins depression (Saadi 1993; Saadi et al. control of drainage patterns along the Takutu rift in the cen-
2005). According to Saadi and colleagues (2005), there is a tral portion of the Guiana Shield is also mentioned by Costa
close relationship between the Araguaia-Tocantins depres- and colleagues (1996). In the central portion of the Brazilian
sion and the Pantanal basin, in terms of their structural shield, neotectonic activity controlling drainage patterns is
control by the Transbrasiliano lineament. The same authors mentioned by Innocencio (1989).
mentioned that in the east of the Araguaia-Tocantins depres-
sion the topography becomes increasingly higher eastward RUNNING WATER DYNAMICS:
as a result of a recent Cenozoic uplift. In this region, sev- UPLANDS VERSUS LOWLANDS
eral hydrological anomalies associated with recent tectonic
adjustments exist, such as unresolved drainages termed There are significant differences between the river dynam-
águas emendadas (coalescence of headwaters of distinct river ics of upland shield areas and lowland foreland basins that
systems). directly affect fish faunal distribution patterns. The first and
According to Riccomini and Assumpção (1999), there is more prominently observed difference between uplands and
evidence of Quaternary, and particularly Holocene, faulting lowlands refers to river gradient. Upland rivers draining the
in almost all geological provinces of Brazil and a close rela- ancient crystalline basement of the Brazilian Shield typi-
tionship of geoid anomalies with uplifted areas of neotec- cally possess stepped gradients, and often are intercalated by
tonic and seismic activity. This vision contradicts previous sequences of rapids and/or waterfalls (Innocencio 1989).
ideas of tectonic stability. Drainage patterns are strongly con- Such rivers are typically ancient, superposed streams (i.e., the
trolled by neotectonic activity, and evidence of this control stream establishes its course without regard to the underly-
has been extensively described in the geological literature. In ing structures) (Tarbuck and Lutgens 2002) or streams with
the Amazon Basin, tectonic control of the drainage pattern courses structurally oriented by the rock basement features,
excluding the obvious and direct effect of the Andean orogeny such as fault lines (L. Soares 1977). Upland rivers are typically
is mentioned by Soares (1977), Costa and colleagues (1996), well fitted into their valleys of exposed crystalline rock and
and Costa and colleagues (2001). According to Costa and col- do not present lateral movements. They lack alluvial plains,
leagues (2001), paleogeographic configurations of the Ama- or else have poorly developed ones. A typical example is
zon River, as well the present observed pattern, are controlled the rio Juruena at the upper rio Tapajós basin, a clear-water
by Meso-Cenozoic tectonics. In the Guiana Shield, tectonic river lacking floodplains and presenting a dendritic drainage
c09.indd 155 4/7/2011 7:21:12 PM

pattern. Exceptions to this pattern are the tributaries of the and species composition are very distinct among those differ-
upper rio Xingu, such as the rio Culuene, rio Sete de Setembro, ent water types (Saint-Paul et al. 2000; Goulding et al. 1988;
rio Batovi, rio Curisevo, rio Suiá-Miçu, and the upper rio Xingu see however Henderson and Crampton 1997 for a slightly dis-
itself, and some tributaries of the upper rio Negro basin, such tinct view on the matter).
as the middle and lower portions of the rio Tiquié, where a flat
relief allowed the formation of extensive alluvial plains where
these rivers meander extensively. Distribution Patterns
In contrast, lowland rivers draining sedimentary basins pres- SHIELDS
ent broad floodplains and highly dynamic lateral movements.
Meanders typically occur and oscillate along the whole exten- The distinct nature of the ichthyofauna of the highlands
sion of the floodplain (Christofoletti 1980; Salo et al. 1986). of both the Brazilian and Guiana shields in northern South
The Amazon River itself, however, is actually an anastomosing America was known, or at least suspected, since Eigenmann
river, possessing a floodplain width of 20–75 km, with a highly (1909b, 1912) and Haseman (1912). Although not areas of
dynamic channel migration, particularly in its upstream endemism in the present sense, Eigenmann’s (1909) ichthy-
reaches (Kalliola et al. 1992; Mertes et al. 1996). ological “provinces” combine faunistic evidence with relief
Other important characteristics of river behavior in lowland data, underlying Eigenmann’s perception that fish distribution
areas refer to megafan dynamics (Horton and De Celles 2001; and diversity were in some way determined by the geological
Wilkinson et al. 2006). Megafans are large, fan-shaped, par- setting. For example, Eigenmann (1909, 319, 328) recognized
tial cones of river-laid sediment with radii arbitrarily defined a “Guiana Province” and a “south-east or East Brazilian pla-
as >100 km, and they typically develop immediately down- teau,” corresponding with the Guiana and Brazilian shields,
stream of a topographic discontinuity, such as the Andean respectively. Eigenmann (1912) stated that “the Guiana
mountain front, with the fan apex located at the point at highland . . . is presumably one of the oldest land-masses of
which the formative river exits the higher country (Wilkin- South America” (p. 94) and noticed that some of the fishes
son et al. 2006). Megafans are testimonies of the degree of found in the upper Potaro River, a tributary of the Essequibo
lateral movement undergone by river channels along low- River in Guiana, might represent “relicts of the original fauna
land foreland basins. Thus, differently from upland rivers, of the Guiana plateau” (p. 104). Haseman (1912, 58–60), who
lowland foreland basins underwent constant and much faster collected extensively in the highland drainages in the Brazil-
hydrogeographic changes. Given these dynamics, associ- ian shield, referred to a impoverished ichthyofauna inhabit-
ated with the fact that foreland basin drainages constantly ing this area, and called attention to Eigenmann’s findings
are connected and disconnected from each other by the tec- of a depauperate fish fauna in the upper Potaro River. Both
tonic evolution of foreland systems, widespread distribution Eigenmann and Haseman clearly realized that the shield areas
patterns are expected for the lowland fish fauna. However, were considerably older than the lowlands and that they har-
geographically limited endemism and local faunistic simi- bored a distinct fish assemblage. However, perhaps because
larities between adjacent drainage basins caused by river cap- of a perception by subsequent authors who dealt with South
tures are the expected patterns of fish distribution in upland American freshwater fish biogeography in the incipient state
shield areas. of knowledge of fish distributions at the time Eigenmann and
Rivers draining shield areas or weathered soils on the low- Haseman formulated their hypothesis, little attention was
lands, generally away from the foreland basins (see Klam- paid to their insights. Subsequent authors preferred, rather, to
mer 1984), are either clear or black water, with low to very identify areas of fish endemism that were delimited based on
low concentrations of dissolved inorganic solids (Sioli 1984; perceived faunal discontinuities across major river drainages
Goulding et al. 2003; Lewis et al. 2006). Possible exceptions are or drainage systems (e.g., Menezes 1972; Ringuelet 1975; Vari
some tributaries of the rio Negro, such as the rio Branco, the 1988). Géry (1964, 1969) was probably the earliest author who
rio Padauari, and the rio Demini, which were termed “semi- envisaged a relationship between the ichthyofaunas from the
muddy” by Goulding and colleagues (2003, 42, 216). Geo- Guiana and Brazilian shields, though hypothesizing that the
chemical analysis of the water of the rio Branco indicated that shield ichthyofauna would have “circumvented” the lowlands
this river is in fact chemically and sedimentologically interme- through the upper Orinoco and upper Amazon systems
diate between black- and muddy-water rivers (E. Ferreira et al. during a period of sea incursion in the Tertiary. The existence of
2007). The Guyanese rivers, the tributaries of the Río Orinoco a distinct ichthyofauna occurring in the Guiana and Brazilian
draining the Guiana Shield, such as the Río Caura and Río shield portions of the Amazon Basin was finally identified and
Caroní, and some tributaries draining the lowlands, such as remarked upon by G. Santos and Jégu (1987) and discussed
the Río Capanaparo and the Río Cinaruco, are also either clear subsequently by Jégu and colleagues (1991), G. Santos and
or black water (Taphorn 1992; Kullander and Nijssen 1989; Ferreira (1999), and Jégu and Santos (2002). As for the Río
Lewis et al. 2006). In contrast, the Amazon river and its tribu- Orinoco system, Mago-Leccia (1978) was the first to notice the
taries possessing their headwaters in the Andes, such as the rio ichthyofaunistic distinction between its eastern (lower) and
Madeira and the rio Japurá (or Río Caquetá), or those drain- upper portions, which drain the Guiana Shield, from the west-
ing the western lowlands, such as the rio Purus and the rio ern, lowland portion of that river basin. Table 9.1 lists exam-
Juruá, constitute the so-called white- or muddy-water rivers, ples of fish taxa that seem to be restricted to the shield areas in
with a high load of sediments and dissolved inorganic solids northern cis-Andean South American river systems.
(Sioli 1984; Goulding, Barthem, et al. 2003; Lewis et al. 2006).
Tributaries of the Río Orinoco possessing their headwaters in LOWLANDS
the Andes, such as Río Apure and Río Meta, can also be consid-
ered white-water rivers (Taphorn 1992; Lewis et al. 2006). The Eigenmann (1909b, 317–19) was the first to recognize the
distinction between muddy, clear, and black waters is impor- cis-Andean foreland basins and associated lowlands in South
tant because it is a well-established fact that both fish biomass America as possessing an ichthyofauna distinct from the river
c09.indd 156 4/7/2011 7:21:12 PM

systems draining shield areas. His “Amazon province” is the and colleagues (1998), was used by several subsequent authors
combination of the lowlands of the Orinoco, Amazon, and to interpret vicariant events in cis-Andean South American
La Plata basins, in his words “the most extensive and intri- freshwater fishes. Armbruster (1998b) proposed that the vicari-
cate fresh water system in the world . . . a network of rivers ant event leading to the allopatric speciation of both Apha-
practically uninterrupted, extending from the mouth of the notolurus species should be ascribed to the establishment of
Orinoco through the Cassiquiare, Rio Branco, Rio Negro, Rio a water divide between the Orinoco and Amazon systems.
Madeira, Rio Guaporé, Rio Paraguay, Parana and La Plata to Montoya-Burgos (2003) associated a major cladogenetic event
Buenos Aires.” He also recognized the relative youth of the within the genus Hypostomus with the establishment of the
lowlands when compared with the highlands, as well as the Michicola Arch as a divide between the Rio Paraguay/Paraná
ichthyofaunistic similarities between the Amazon and La Plata and Amazon river systems, although some taxa included in
basins. However, as discussed later, most subsequent authors his “Paraguay/Paraná clade” actually occur in river drainages
failed to appreciate the distinction between the ichthyofaunas from the southeastern portion of the Brazilian Shield. Hubert
from the lowlands and shield areas. Weitzman and Weitzman and Renno (2006) suggested that “the Vaupes and Michicola
(1982), in their analysis of the distribution patterns of Nan- arches enhanced allopatric differentiation in western South
nostomus and Carnegiella, discussed the representatives of each America” (p. 1429), and Albert, Lovejoy, and colleagues (2006,
genus occurring in the Guiana Shield and in the lowlands of 22–23) noticed that the “[gymnotiform] fauna of the Orinoco
the Amazon and Orinoco basins but centered on the perspec- Basin is much more similar to that of western Amazon, from
tive of their failed attempt to correlate fish distributions with which it is currently isolated hydrologically, than it is to the
the purported Quaternary forest refugia. Based on distribution drainages of the Guianas Shield or eastern Amazon, with
data of cichlids, Kullander (1986, 28–41) was the first to discuss which it is now connected.” They also remarked, contra Vari
extensively fish biogeography in the western Amazon region (1988), that the connection between both systems provided
and to suggest it as constituting an area of fish endemism (p. by the Rio Negro/Río Casiquiare cannot explain the similar-
37, fig. 9). Although without specifying the geological basis ity between them, since they are “poor routes for dispersal in
underlying the pattern, he was also the first to notice the area electric fishes, possibly because of the physical barriers (i.e.,
relationship between the western Amazon and the Orinoco rapids) at Pto. Ayacucho and São Gabriel da Cachoeira and
(pp. 33–35, fig. 5). Vari (1988, 360) noticed that some curi- the chemical barriers (e.g., differences in pH, temperature,
matids, such as Curimata aspera, were restricted to the western conductivity) between the black water Rio Negro and Casiqui-
Amazon, but remarked that such distribution patterns “corre- are Canal and the white water Orinoco and Amazon rivers.”
late primarily with white water conditions of that area rather Finally, Wilkinson and colleagues (2006, 164–65) gave some
than being representative of historical events.” He also attri- examples of fishes potentially dispersed via megafan dynamics
buted the occurrence of several curimatids in both the Orinoco across the water divides of the Paraguay, Amazon, and Orinoco
and Amazon systems to the connection between those systems river systems. In Table 9.2 we list fish taxa that exemplify low-
provided by the Río Casiquiare (p. 355). land/foreland basin distribution patterns.
Araújo-Lima and Goulding (1997, 27) mentioned that
Colossoma macropomum occurs in the western-central Ama- Shields and Lowlands
zon Basin in muddy- or black-water rivers but is limited in
clear-water, shield-draining rivers to their lower reaches, below We believe that the proposed distribution patterns delineated
large waterfalls or cataracts. G. Santos and Ferreira (1999, in this chapter resulted from ecological constraints in asso-
353–54) and especially Jégu and Keith (1999, 1136, 1138–39, ciation with major historical events affecting freshwater fish
fig. 3) discussed the presence of a typical fish assemblage in faunas in South America. We provide a pattern of relation-
the western and central portion of the Amazon Basin, in the ship among these areas based on a parsimony analysis in
lowland, white-water rivers. Jégu and Keith (1999, 1134) state: Figure 9.5. For some of the proposed patterns, such as the fish
“From the foothills of the Andes to the area downstream of fauna associated with foreland basin evolution, the scenario
Santarém, along 3000–5000 km of waterways (in the Amazon is underpinned by a relatively clear causal explanation. How-
and its Andean tributaries), the Serrasalminae guild found in ever, other patterns present a more obscure origin, which is
the white waters is strikingly constant. These species are gener- probably a mix of both historical and ecological causes. Fish
ally restricted to an area within 20–30 km upstream from the faunas restricted to shields and the disjunct distribution of
confluence with the Amazonas River, but they are occasionally several taxa in both the Guiana and Brazilian shields are
found as far as 80–100 km upstream.” Jégu and Keith (1999) examples of such complex association between ancient his-
also noticed a decrease on the number of species of the “várzea tory and ecological constraints. It is important to stress that
Serrasalminae guild” in the lower Amazon Basin and the occur- some ecological factors that clearly influence fish distribu-
rence of some representatives of that assemblage in a single tion patterns in northern cis-Andean South America, such as
river system draining shield areas, the rio Tocantins. Lundberg water typology, are, as mentioned previously, a consequence
and colleagues (1998), after presenting his scenario for the evo- of geomorphological processes and, as such, possess a histori-
lution of the lowland river systems in South America, which cal component.
included a suggestion that the establishment of the modern An even more obvious ecological factor correlated with
divides between the Amazon, Orinoco, and La Plata systems geomorphological processes is the contrast between the high-
happened during the late Miocene, predicted that “establish- energy rivers crossing the steep slope of the Andes and the
ment of sympatry among relatively close species within clades more gently sloped shields with the sluggish water flow and
of lowland fishes is an expectation and such patterns have enormous expanses of floodplains found in the lowlands. It
been noticed. . . . [C]oincident, or nearly so, with the forego- has been noticed that most fish taxa displaying shield dis-
ing was the Orinoco-Amazonas vicariance event” (p. 43). The tribution patterns are highly rheophilic fishes, in contrast
fragmentation of the foreland basins as a consequence of arch with the typical floodplain/river-channel specialists found in
uplifting during the late Miocene, hypothesized by Lundberg the lowlands (e.g., G. Santos and Jégu, 1987; G. Santos and
c09.indd 157 4/7/2011 7:21:12 PM

F I G U R E 9.5 Area cladogram depicting hierarchical relationships among river drainages discussed in this chapter obtained by a parsimony analy-
sis base on examples of shared taxa between areas.
Ferreira, 1999; Jégu and Keith, 1999; Jégu and Santos, 2002). pattern of the Amazon River was quite distinct from the pres-
The piedmont area of the headwaters of the Amazon and Ori- ent-day configuration. Potter (1997, 338, fig. 5) pointed out
noco basins draining the Andean slope (i.e., above 200 meters that the pre-Miocene Amazon watershed possessed headwaters
asl) also possesses a distinctive fish assemblage, as previously much further eastward than the present day. In this ancient
noticed by Ibarra and Stewart (1989) and Galacatos and col- Amazon River drainage, both the Guiana Shield, in the north,
legues (1996) in the upper Río Napo in Ecuador and Taphorn and the Brazilian Shield, in the south, represent headwaters
(1992, 490, 505) in the Rio Meta system in Venezuela. of the same interconnected drainage basin. Also, according to
Several taxa presenting foreland distribution patterns J. Costa and colleagues (2001), the Purus Arch worked as an
appear to be restricted to the westernmost, higher portions uplifted area between the Amazonas and Solimões basins with
of the Amazon and Orinoco basins because of their ecologi- the axial paleodrainage diverging from it from the Late Paleo-
cal requirements. Examples are some cichlids in the Peruvian zoic until the Early Tertiary. From the Early to mid-Tertiary,
Amazon (Kullander 1986, 29); the genus Attonitus (Vari and due to the uplift of the Andes, the paleodrainage of the west
Ortega 2000, fig. 3); and Brycon hilarii, B. whitei, and the Sal- side of the Purus Arch was reversed and formed the Amazon
minus species inhabiting the Orinoco and western Amazon River flowing eastward. From that time to the present, the
systems (Taphorn 1992, 490; F. Lima, unpublished data). Also drainage system was reorganized and gave rise to the formation
noteworthy as an indication of preference for swift-flowing of the modern landscape patterns, which was driven mostly
waters are the distribution of the genus Creagrutus (Vari and by neotectonic processes, not necessarily associated with the
Harold 2001, 44, fig. 18) and of the characiform family Par- Andean tectonics (J. Costa et al. 2001). Some of the examples
odontidae as a whole, both well represented on the western of disjunct distribution patterns presented by us (e.g., Figure
portion of the Amazon and Orinoco basins and across the river 9.6) fit perfectly to the limits of the Pre-Miocene Amazon Basin
systems draining the shields but lacking in the lowlands. From as discussed by Potter (1997) and reiterated by J. Costa and
the examples given in Table 9.2, however, it becomes plain colleagues (2001). Distributional disjunctions are probably
that most lineages with rheophilic representatives typical of the result of the ecological limits imposed by the present-day
shield areas are lacking in the much more geologically recent Amazon River, which acts as a barrier for reophilic, clear- and/
and unstable Andean piedmont. or black-water dwelling fishes, which probably were once dis-
For practical reasons we group fish faunal distribution pat- tributed throughout the whole pre-Miocene Amazon Basin.
terns in two main sets: fish fauna from shields—groups of These faunal components are today limited to the upper por-
fishes that appear to be restricted to upland shield areas by tions of this ancient basin, being thus relictual to the Guiana
both historical and ecological constraints—and a fish fauna and Brazilian shield tributaries.
associated with foreland basins—groups that clearly under-
went distribution range expansions associated with foreland CENTRAL BRAZILIAN SHIELD
basin evolution.
Several fish species distributed throughout the shields do not
DISJUNCT SHIELDS
have their distribution ranges delimited by basins divides.
Rather, these species occur in more than one river basin.
We interpret the present-day disjunction of the Brazilian and Jupiaba acanthogaster, for example, is widespread across riv-
Guiana shield areas based on a major historical event associ- ers draining the Brazilian Shield from the rio Tocantins basin
ated with the tectonic evolution of the Amazon Basin. The westward to the upper rio Madeira and the upper rio Paraguay
Amazon River is installed within a megashear system, dated basins (Figure 9.7). This pattern contradicts the idea that river
from the late Jurassic, the Amazon graben, which separated basins correspond to major areas of endemism. It provides evi-
the Guyanese and Brazilian cratons (Grabert 1983; Caputo dence of constant faunal exchanges by neighboring drainage
1991). As discussed by Potter (1997), the pre-Miocene drainage basins thanks to continuous headwater captures associated
c09.indd 158 4/7/2011 7:21:12 PM

Roeboexodon guyanensis
Leporinus brunneus
Anostomus ternetzi
Political boundary
Rivers
F I G U R E 9.6 Distribution of Anostomus ternetzi, Leporinus brunneus (Anostomidae), and Roeboexodon guyanensis (Characidae).
Tocantisia piresi
Jupiaba achantogaster
Hyphessobrycon moniliger
Moenkhausia phaeonota
Thayeria boehlkei
Political boundary
Rivers
Distribution of Tocantisia piresi (Auchenipteridae), Jupiaba achantogaster, Hyphessobrycon moniliger, Moenkhausia phaeonota, and
F I G U R E 9.7
Thayeria boehlkei (Characidae).
with neotectonic activity as previously proposed by A. Ribeiro activity is probably the main cause of the evolution of huge
(2006) and demonstrated by A. Ribeiro and collegues (2006). tectonically developed depressions, such as the Pantanal
In the case of the central Brazilian shield, as illustrated by the and Araguaia-Tocantins depressions. Reactivations along the
distribution pattern presented by Jupiaba acanthogaster, Transbrasiliano Lineament have surely provided opportunities
the major tectonic feature responsible for the accelerated for constant faunal mixing among neighboring upland
fluvial dynamism of constant headwater captures is the drainages river systems. In addition to the several fish taxa
Transbrasiliano Lineament. As previously mentioned, this listed in Table 9.2, the ichthyofaunistic interchange between
megashear zone is tectonically active, and its neotectonic the upper rio Paraguay and the upper rio Tapajós system,
c09.indd 159 4/7/2011 7:21:12 PM

Curimatella meyeri
Rhytiodus spp
Parecbasis cyclolepis
Political boundary
Rivers
F I G U R E 9.8 Distribution of Curimatella meyeri (Curimatidae), Rhytiodus spp. (Anostomidae), and Parecbasis cyclolepis (Characidae).
discussed by F. Lima and collegues (2007), is another example by the Amazonian floodplain várzea forests (Wittman et al.
of the faunal mixing that has occurred along this lineament. 2006). That distribution pattern is the one described for
In some cases, upland shield groups clearly represent ancient several Serrasalminae taxa (the so-called várzea guild) by Jégu
distributions, since lowland foreland basin systems are new top- and Keith (1999, fig. 3). Except for some taxa also found in
ographic landscapes relative to shield areas, but, presumably, in the rio Tocantins basin, which are discussed later, fish taxa
other cases, initial diversification might have occurred on low- presenting that distribution pattern generally only occur in
lands, with such groups secondarily occurring in upland areas. the clear-water rivers in their lower section, below the first
However, determination of such a sequence of events cannot be rapids that mark the beginning of the shield-draining portion
described without phylogenetic hypotheses. Recently, Menezes of the river.
and colleagues (2008) demonstrated how the tectonic evolution The lower and middle portions of the rio Negro basin, down-
of the Rio Paraná basin throughout tectonic reactivations along stream from the São Gabriel waterfalls, drain the lowlands,
the borders of the basin configured the present-day distribution and several taxa typically displaying western-central Amazon
patterns of the Glandulocaudinae, in which basal groups are distribution patterns also occur in that area—for example,
restricted to upland rivers draining the Brazilian Shield and Osteoglossum bicirrhosum (Osteoglossidae), Arapaima gigas
more derived groups underwent diversification along lowland (Arapaimatidae), all the Amazonian Pristigasteridae, Hydrolycus
areas of the Paraná-Paraguay and coastal rivers of southern scomberoides (Cynodontidae), Cetopsis coecutiens (Cetopsidae),
to northeastern Brazil. A. Ribeiro and colleagues (2005) also Pseudorinelepis genibarbis (Loricariidae), Cichla monoculus (Cich-
discussed the distribution patterns presented by the characid lidae), and Plagioscion montei (Sciaenidae). However, in spite of
genera Creagrutus and Piabina, concluding that upland shield the lack of physical barriers such as rapids and waterfalls in
areas of Central Brazil are the locale of initial diversification of that river stretch, several other fish taxa displaying western-
the group, and subsequently phylogenetic radiation occurred central Amazon Basin distribution patterns are absent from the
in adjacent areas, including trans-/cis-Andean basins. At the middle and lower rio Negro basin. As first noticed by Goulding
moment, we are unable to provide examples of groups that ini- and colleagues (1988, 98–101), the scarcity or absence of sev-
tially diversified in the lowland areas but secondarily invaded eral common white-water species that also occur in the lower
the shield areas. We suggest that groups that are well diversified courses of clear-water rivers in the rio Negro basin can only be
in the lowlands but with some typical shield representatives, ascribed to their inability to cope with the highly acidic black
such as Anostomidae, Cichlidae, and the order Gymnotiformes, waters or else with the low biological productivity of that sys-
are the ones where those examples will be found. tem. Examples given by Goulding and colleagues (1988) were
Rhytiodus spp., Schizodon fasciatus (Anostomidae), Mylossoma
WESTERN-CENTRAL AMAZON
spp. (Characidae), Psectrogaster spp. (Curimatidae), Thoracho-
charax stellatus (Gasteropelecidae), Cichlasoma spp. (Cichlidae;
What is here considered as a western-central Amazon pat- see also Kullander 1983, 286), Pseudotylosurus microps (Beloni-
tern of distribution includes fishes that are widespread across dae), and Colomesus asellus (Tetraodontidae). Other examples
the lowlands of the Amazon Basin—that is, across the white- of common white-water fish species that are notoriously
water tributaries and along the main channel of the Amazon absent from black waters are Potamorhina altamazonica (Vari
river (Figure 9.8)—and that coincide with the area covered 1984), Prochilodus nigricans (Castro and Vari 2004), Pygocentrus
c09.indd 160 4/7/2011 7:21:13 PM

nattereri (Goulding and Ferreira 1984), Cetopsis candiru (Vari Hubert and colleagues (2007a) that those intracratonic arches
et al. 2005), plus several examples listed in Table 9.2. Pygo- may have played a role in determining vicariant events for
centrus nattereri and Colomesus asellus are known from “semi- the clade comprised by Serrasalmus and Pygocentrus species
muddy” tributaries of the rio Negro basin such as the rio should also be dismissed as a gross misinterpretation of the
Branco and rio Padauiri, but are very scarce or absent in the neotectonic processes in the lower Amazon, which were much
rio Negro itself. Also, several large-sized migratory characi- more complex than a simple model of arch deformation (e.g.,
forms such as Colossoma macropomum, Brycon amazonicus, J. Costa et al. 2001). Though the lower Amazon River is a rela-
Semaprochilodus taeniurus, and S. insignis enter the lower and tively “recent” addition to the western-central Amazon River
middle rio Negro in postspawning migrations but migrate ecosystem, incorporated into that system since the breaching
downstream into the Amazon river to breed (Araújo-Lima and of the Purus arch, dated either as taking place during the late
Goulding 1997; Goulding et al. 1988; Borges 1986; M. Ribeiro Miocene, ~8 MY (Lundberg et al. 1998; Costa et al. 2001), or
and Petrere 1990), consequently using black-water habitats late Pliocene, ~2.5 MY (Campbell et al. 2006), faunistic dif-
only seasonally, and even so in a facultative fashion, since all ferences between this portion of the basin from the upstream
these species possess populations elsewhere that do not mig- reaches of the basin seems more likely to be due to ecological,
rate into black waters. As noticed by Albert, Lovejoy, and col- rather than to historical, factors.
leagues (2006) and supported by Winemiller, López-Fernández,
and colleagues (2008), physical and chemical barriers do not
COMPOSITE SYSTEMS
allow the Negro and Casiquiare rivers to be an efficient bio-
logical connection between the Orinoco and Amazon systems. As previously noticed by Jégu and Keith (1999) and Hrbek,
Crossa, and colleagues (2007), the only river system in the
EASTERN LOWER AMAZON
Amazon Basin draining shield areas that possesses a large
number of fish taxa otherwise only known from lowlands of
A striking trend observed in some fish species possessing the central and western Amazon Basin is the rio Tocantins
western-central Amazon distribution patterns, earlier noticed system. The large tributary of the rio Tocantins, the rio Ara-
by Jégu and Keith (1999), is that some of them are scarce guaia, unlike other shield-draining tributaries of the Amazon
or absent in the lower portions of the Amazon River. Three River, possesses a huge sedimentary basin, the Bananal plain,
Serrasalminae, Colossoma macropomum, Piaractus brachypo- covering 90,000 km2 (Latrubesse et al. 2005). This alluvial
mus, and Serrasalmus elongatus, are absent from the lower valley is very recent, a result of a still-subsiding Quaternary
Amazon river (Jégu and Keith 1999; for the tambaqui, see tectonic deep (Saadi 1993; Saadi et al. 2005). Typical western-
also Araújo-Lima and Goulding 1997, 24–25). We can add central Amazon fishes that occur in the rio Tocantins system
at least a fourth species to that list, Brycon amazonicus (F. are Arapaima gigas (Arapaimatidae), Osteoglossum bicirrhosum
Lima, unpublished data). We suggest that the probable causal (Osteoglossidae), Pellona castelnaeana, Pristigaster cayana (Pris-
factor for the absence of these and presumably several other tigasteridae), Leporinus trifasciatus (Anostomidae), Psectrogaster
western-central Amazon fish species from the area is the occur- amazonica (Curimatidae), Thorachocharax stellatus (Gasterop-
rence of circadian variations in the water level of the lower elecidae), Mylossoma spp., Pygocentrus nattereri (Characidae),
Amazon linked to tidal influence (see Goulding, Barthem, Cetopsis candiru, C. coecutiens (Cetopsidae), Auchenipterichthys
et al. 2003, 38, for a depiction of the extent of tidal influence on coracoideus (Auchenipteridae), the monophyletic clade that
the lower Amazon Basin). The circadian variation of water level includes Cichla monoculus, C. pleiozona, and C. kelberi (Cich-
considerably decreases the effects of the seasonal flood pulses lidae), and Colomesus asellus (Tetraodontidae). The freshwater
and presumably adversely affects migratory, highly fecund dolphin, Inia geoffrensis, and the large podocnemid turtle
total-spawners such as Colossoma macropomum and Brycon Podocnemis expansa, both of which possess a lowland distribu-
amazonicus, which depend on a seasonal and extended flood- tion pattern within the Amazon Basin, are also well-known
ing of the floodplains for an effective recruitment. However, members of the aquatic fauna from the rio Tocantins system.
some middle- to large-sized fishes presenting western-central Interestingly, phylogeographic data on Arapaima gigas (Hrbek
Amazon distribution patterns that possess low fecundity, et al. 2007) and on the turtle Podocnemis expansa (Pearse et al.
large eggs, and multiple spawning, such as Osteoglossum 2006) suggest virtually no differentiation between the popula-
bicirrhosum, Arapaima gigas, and Pygocentrus nattereri, occur in tions inhabiting the rio Tocantins system and the ones from
the lower Amazon, presumably because their life-history traits the western-central Amazon.
are more adjusted to the local circadian variations of water From the preceding discussion, it becomes plain that two
level. The absence of some fishes presenting a western-central of the larger tributaries of the Amazon Basin, the rio Tocan-
Amazon distribution pattern is presumably at least one of the tins and the rio Negro (see discussion on the western-central
reasons that led Hubert and Renno (2006, 1429) to consider Amazon pattern), can be considered as possessing a “compos-
this area as a distinct area of endemism, and it was interpreted ite” nature, the first due to the subsidence of a portion of the
by them as having as its causal factor the presence of the Purus basin that has transformed what was previously a presumably
structural arch acting as a boundary between the western and typical plateau drainage into a lowland river system, and the
eastern Amazon. However, as remarked by Rossetti and col- second because its upper portion lies on the Guiana Shield,
leagues (2005, 86), structural arches situated along the Amazon while its middle and lower portions drain the lowlands. The
river graben such as the Carauari, Purus, and Gurupá arches consequence is that the ichthyofauna and presumably the
are basement structures buried under a mantle of variable-aged remaining aquatic biota inhabiting these river systems are, at
sediments that, other than determining constraints on flood- least to a certain extent, a mixture of lowland/highland taxa.
plain development (Mertes et al. 1996; Dumont et al. 1988, While hydrochemistry has played a major role in determining
fig. 2), are not expressed superficially in eastern Amazon and, the absence in the rio Negro basin of a large number of typi-
consequently, could not plausibly have acted as biogeographical western-central Amazon fishes, it is less clear why many
cal barriers from the late Miocene onward. The suggestion by typical várzea-floodplain/white-water fishes failed to establish
c09.indd 161 4/7/2011 7:21:13 PM

Abramites hypselonotus
Political boundary
Rivers
F I G U R E 9.9 Distribution of Abramites hypselonotus (Anostomidae).
populations in the rio Tocantins system. Several explanations a matter still to be determined. Examples of fishes possess-
may be advanced, such as the distinct type of water of the rio ing shield distribution patterns that only occur within this
Tocantins system, the absence of a typical várzea forest in that river system in the shield, clear-water tributaries, as the rio
river system, or else the presence of several rapids along the Machado and rio Aripuanã, are Leporinus brunneus (Figure
lower courses of the rio Tocantins and rio Araguaia that might 9.6) and Jupiaba acanthogaster (Figure 9.7). In contrast, several
have acted as a barrier to the dispersal of several lowland taxa. fishes possessing western-central Amazon distribution patterns
Of course, there is no way to choose among those alternative occur in the rio Madeira system only in its lowland, white-
hypotheses, and it is possible that actually all of them have water tributaries, such as Parecbasis cyclolepis and Rhytiodus
played a role in determining which lowland taxa would be spp. (Figure 9.8).
able to occupy the relatively recent floodplains of the rio
Tocantins system. We suggest that the rio Tocantins system is UNDER THE ANDEAN SHADOW: THE DYNAMISM
an example of “biotic merging,” mentioned by Lundberg and OF FORELAND BASINS
colleagues (1998, 44) as a possible biogeographical outcome of
geomorphological events. Historical relationships among foreland basins are corrobo-
Other river systems possessing a composite nature in north- rated both by the fact that distribution ranges of several
ern cis-Andean South America are the Essequibo River in Guy- species (e.g., Abramites hypselonotus; Figure 9.9, and the numer-
ana and the rio Madeira system. The upper portions of the rio ous examples in Table 9.2) transpose basin divides, but also
Branco and the Essequibo River drain the Takutu graben (e.g., by phylogenies of several taxa of different taxonomic levels.
Milani and Thomaz Filho 2000; Leite et al. 2007, fig. 2), which The evidence suggests that hydrological changes acting along
established a continuous lowland area between both river sys- the slope of the Andean chain throughout the evolution of
tems. The upper Ireng River, a tributary of the upper rio Branco foreland basins are a constant and ancient process, promot-
system, and the Rupununi River, a tributary of the Essequibo ing both expansions of distribution ranges and cladogenesis
River, are considered to be in contact during wet years through through time. Examples of several levels of faunal relation-
the flooded plains of Lake Amuku (Lowe-McConnell 1964). In ships among foreland basins based on phylogenetic patterns
fact, some fishes presenting western-central Amazon distribu- were provided by several authors. Vari (1984) demonstrates
tion patterns extend their distribution into the rio Branco and that Potamorhina laticeps from the Lake Maracaibo drain-
Essequibo River, such as Arapaima gigas, Osteoglossum bicirrho- age is the sister group of a more inclusive clade representing
sum, Pygocentrus nattereri, and Colomesus asellus. These species a sequence of sister groups including P. pristigaster (Amazon
are not known elsewhere from Guyanese river systems, except Basin), P. squamoralevis (Paraná-Paraguay drainage), P. altama-
from their presence in the coastal lowlands of the Amapá state, zonica (Amazon and Orinoco), and P. latior (Amazon Basin).
Brazil, Brycon amazonicus and Colomesus asellus excepted (Jégu Weitzman and Fink (1985) provide evidence that the genus
and Keith 1999). The rio Madeira system has its western head- Argopleura (from the Magdalena) is the sister group of a more
waters either in the Andes or in the lowlands, and its eastern inclusive clade including the genus Iotabrycon (from Guayas
headwaters at the Brazilian shield (Goulding et al. 2003). area of the western Andean slope) plus the genera Xenurobry-
As a consequence, those distinct areas harbor very distinct con (with three species, X. macropus from the Paraguay basin
fish assemblages, although the extent of that distinctness is and X. pteropus and X. heterodon from the western Amazon
c09.indd 162 4/7/2011 7:21:13 PM

Basin) plus the genera Scopaeocharax (western Amazon) and taxa shared between the rio Tiquié (upper rio Negro basin)
Tyttocharax (Amazon Basin). According to Vari (1989a), Curi- and the lowlands of the western Amazon, and presumably
mata mivartii, from the Magdalena, Cauca, San Jorge, and indicates that a faunal interchange between the western low-
Sinú river basins, is the sister group of a more inclusive clade lands and the westernmost portion of the Guiana Shield has
including C. aspera (middle and upper Amazon Basin) and C. occurred.
cerasina (Rio Orinoco), which are sister groups. Vari (1991) also
provided evidence of a monophyletic group within the genus Conclusions
Steindachnerina, including a sequence of sister groups repre-
sented by S. conspersa (Lower Paraná and Paraguay), S. bimacu- One of the corollaries that follow from the preceding discus-
lata (Amazon and Orinoco), S. leucisca (Amazon Basin), and sion is that, in a strict sense, neither the Orinoco nor the Ama-
S. binotata (upper Madeira River basin). Schaefer and Stewart zon Basin should be considered, as a whole, as an area of fish
(1993) also conclude that the armored catfishes of the Panaque endemism. Rather, those river systems are composite areas,
dentex species complex consist of a monophyletic group, dis- geologically and biogeographically speaking. Taxa thought to
tributed through the extreme western headwaters of the Ama- be endemic to either of those river basins actually are endemic
zon Basin, the Rio Purus of western Brazil, and the western Rio to a portion of the river basin, more often either in the low-
Orinoco basin of Venezuela. Finally, Reis (1998b) demonstrates lands or in the shield areas. Taxa occurring in the lowlands
a close phylogenetic relationship among species of the genus commonly possess sister taxa in the lowlands of neighboring
Lepthoplosternum distributed along the foreland basin systems river systems. For example, fish species presenting a western-
of South America adjacent to the Andean chain, including the central Amazon distribution pattern generally have sister taxa
western Amazon region, the upper Rio Madeira, the Paraguay in the La Plata, Orinoco, and/or some trans-Andean river sys-
Basin, and the coastal lowland area of southern Brasil (Reis tem. In contrast, fish species occurring in shield areas in the
1998, 356, fig. 11). Amazon Basin typically have sister taxa in the shield-draining
It is expected that other aquatic animals such as aquatic portions of the Orinoco Basin and/or in Guyanese river sys-
mammals, reptiles, decapod crustaceans, mollusks, and, per- tems. Taxa relatively ubiquitous in the Amazon Basin, occur-
haps, some amphibians, present distribution patterns mirror- ring in river systems draining both lowlands and shield areas,
ing the dichotomy between lowlands and shield areas here such as Boulengerella cuvieri (Vari 1995, fig. 45), Caenotropus
described for fishes, and, in fact, some examples are avail- labyrinthicus (Vari et al. 1995, fig. 20), Serrasalmus rhombeus, or
able. For instance, the Amazon River dolphin, Inia geoffrensis, Brachyplatystoma filamentosum, cannot be used as evidence for
possesses a distribution pattern that encompasses essentially an “Amazon Basin” area of endemism, since they also occur at
the lowlands of the Amazon and Orinoco systems, though it the Guyanese river systems and in the Rio Orinoco Basin, and
occurs in the upper rio Negro and Río Casiquiare (Emmons actually display a distribution pattern that should rather be
and Feer 1999, map 137). As a consequence, contrary to the called “northern cis-Andean South American.”
fish taxa examples listed previously, its distribution across As noted at the beginning of the chapter, the biogeographi-
both river systems cannot be ascribed to a “foreland basin” cal hypotheses presented here avoid the use of the concept of
distribution pattern. The giant Amazon river turtle, Podecnemis “area of endemism.” First of all, there is no consensus as to the
expansa (Pelomedusidae), seems also to be restricted to low- definition of an area of endemism (see Lomolino et al. 2006,
land areas, occurring in the western-central Amazon Basin, rio 435–36), and in fact its very existence should not be presup-
Tocantins, rio Negro/rio Branco, Río Orinoco, and Essequibo posed (Hovenkamp 1997). Traditionally, areas of endemism
River systems (e.g., Pearse et al. 2006). Several crabs possess for freshwater fishes in South America have been defined as
distribution patterns that fit either the shield or the lowlands corresponding to distinct faunistic assemblages found within
distribution patterns described here. the boundaries of the main river drainages. The recognition
For example, the pseudothelphusids Kingsleya latifrons, that river basins might consist actually of composite areas,
K. siolii, and Fredius denticulatus display a Guiana Shield distri- with different taxa presenting phylogenetic relationships
bution pattern, occurring in the Guyanese river systems and pointing to distinct individual patterns, was earlier remarked
northern tributaries of the lower rio Amazonas (Magalhães upon by Vari (1988, 358) when discussing the curimatid fauna
2003, figs. 90, 96, and 98, respectively). Trichodactylid crabs from the rio São Francisco basin. The data presented in this
displaying lowland/foreland distribution patterns are Dilocar- chapter amply support that the same holds true for all major
cinus pagei, Poppiana argentiniana, Valdivia camerani, and Zil- river basins of northern cis-Andean South America. In fact,
chiopsis oronensis, in the Paraná/Paraguay and western-central as pointed out previously by Minckley and colleagues (1986,
Amazon basins; the genus Sylviocarcinus, with one species, S. 610–11) and A. Ribeiro (2006, 242–43), fish assemblages from
australis, in the Paraná/Paraguay system and two species, S. any given river system should be expected to be the result of
devillei and S. maldonadoensis, distributed across the western/ the accumulation of diverse faunistic interchanges between
central Amazon Basin and also into the rio Tocantins system; neighboring river systems through geological time and, as
and Rotundovaldivia latidens and Trichodactylus faxoni, with such, to be composite. This statement, in fact, is true for any
western-central Amazon distribution patterns (Magalhães given “area of endemism” (McDowall 2004).
2003, figs. 106, 120, 154, 162, 128, 130, 132, 126, 142, respec- As stated by Potter (1997, 332), “Rivers continually accom-
tively). The trichodactylid Moreirocarcinus emarginatus occurs modate themselves to tectonic and climatic changes and
in the western portion of the Amazon and Orinoco basins, and hence [it] is rarely meaningful to think of rivers ever having
also in the Río Magdalena in trans-Andean South America, pos- had either a definite beginning or an end. . . . [D]ifferent parts
sessing, as such, what could be interpreted as a foreland basin of a river system may have different ages.” Present drainage
distribution pattern, except for its occurrence in the upper rio configurations are only the last chapter of a continuous and
Negro basin, which lies in the Guiana shield (Magalhães 2003, complex geomorphological evolution. It is thus expected that
fig. 116). Its distribution pattern, however, is similar to the the complexity of the evolution of river landscapes will be mir-
one discussed by Britto and colleagues (2007) for some fish rored by an intricate biogeographical history of their biotas.
c09.indd 163 4/7/2011 7:21:13 PM

Rather than trying to define elusive “areas of endemism,” we basis of such evident accelerated dynamics, we believe that the
suggest that a better approach to understand the historical biogeography of freshwater fishes cannot be interpreted based
biogeography of freshwater fish fauna in South America is to on the naive assumption of using a single “pattern approach”
identify distribution patterns of monophyletic taxa, and then founded exclusively on comparisons between areas of ende-
to seek historical and/or contemporary evidence that might mism. As pointed out by McDowall (2004, 345): “Area rela-
explain the identified pattern. General hypotheses will account tionships [patterns] are . . . an [derivative, secondary] outcome
for general trends in distributions patterns, but they will never of the [primary] processes that generate distribution patterns
paint the whole picture. For example, how could we hope to of individual species, about which we are interested in learn-
reconstitute the historical biogeography for a purported area of ing. . . . [T]he overall observed patterns are the accumulation of
endemism called “rio Tocantins”? There is evidence pointing the individual patterns . . . . If we do not give attention to indi-
to ichthyofaunal exchanges between the headwaters of the vidual histories, we face the prospect that varied causations of
rio Tocantins and all the major river drainages with which it that history will be subsumed within the general history, and
possesses divides: rio Paraguai, rio Paraná, rio São Francisco, [that we may lose] lose a great deal from understanding biotic
and rio Xingu. Also, as discussed earlier, the rio Tocantins basin distributions.” We conclude, thus, that understanding the
is within an active tectonic depression which resulted in the processes of landscape evolution and associating such events
development of extensive alluvial plains in the rio Araguaia with observed phylogenetic and distribution patterns are the
and which was hypothesized previously as an explanation for only useful ways to elaborate scenarios of faunal evolution
the presence of several taxa presenting western-central Amazon through time.
distribution patterns in the area. Apparently, all those distinct
biogeographical events could be simply summed up by stating ACKNOWLEDGMENTS
that the rio Tocantins system is a composite area (Figure 9.5).
But, then, what exactly is the meaning of recognizing the rio We are grateful to several friends who joined us in field expe-
Tocantins system as a historical, biogeographical entity at all? ditions during which we obtained data and exchanged some
It could be argued that, in order to reflect area relationships ideas about fish biogeography in South America: W. G. R.
more accurately, the rio Tocantins basin could be “sliced” into Crampton, A. Cabalzar, M. C. Lopes, H. Ortega, R. E. Reis, H.
smaller areas of endemism, some of them presumably shared A. Britski, F. A. Machado, J. A. S. Zuanon, L. S. Sousa, E. F. G.
with neighboring river systems. However, except for perhaps Ferreira, C. R. Moreira, and J. Alves de Souza. We thank
using geomorphological criteria, there would be no means to M. de Vivo for lively and inspiring discussions on several
determine exact limits for those areas, since subsequent faunal biogeographical topics, and N. A. Menezes, H. A. Britski, J. S.
exchanges certainly obscured faunistic limits and resulted in Albert, J. G. Lundberg, and an anonymous reviewer for con-
the blending of previously distinct fish assemblages. On the structive comments and criticisms on the manuscript.
c09.indd 164 4/7/2011 7:21:13 PM

View publication stats

Continental-Scale Tectonic Controls of Biogeography and Ecology

Uploaded by

Copyright:

Available Formats

Continental-Scale Tectonic Controls of Biogeography and Ecology

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

Continental-Scale Tectonic Controls of Biogeography and Ecology

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Continental-Scale Tectonic Controls of Biogeography and Ecology

Chapter · August 2011

Flávio C. T. Lima Alexande C. Ribeiro

SEE PROFILE SEE PROFILE

Systematics and speciation processes of the genus Hypostomus View project

The user has requested enhancement of the downloaded file.

Continental-Scale Tectonic Controls

c09.indd 145 4/7/2011 7:21:11 PM

c09.indd 146 4/7/2011 7:21:11 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 147

c09.indd 147 4/7/2011 7:21:11 PM

c09.indd 148 4/7/2011 7:21:11 PM

Winterbottom 1980; Santos and

c09.indd 149 4/7/2011 7:21:11 PM

Armbruster 2002; Armbruster et al.

c09.indd 150 4/7/2011 7:21:11 PM

c09.indd 151 4/7/2011 7:21:11 PM

c09.indd 152 4/7/2011 7:21:12 PM

Sorubim elongatus X X Littmann 2007

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 153

c09.indd 153 4/7/2011 7:21:12 PM

c09.indd 154 4/7/2011 7:21:12 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 155

c09.indd 155 4/7/2011 7:21:12 PM

c09.indd 156 4/7/2011 7:21:12 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 157

c09.indd 157 4/7/2011 7:21:12 PM

c09.indd 158 4/7/2011 7:21:12 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 159

c09.indd 159 4/7/2011 7:21:12 PM

c09.indd 160 4/7/2011 7:21:13 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 161

c09.indd 161 4/7/2011 7:21:13 PM

F I G U R E 9.9 Distribution of Abramites hypselonotus (Anostomidae).

c09.indd 162 4/7/2011 7:21:13 PM

C ON TI N EN TAL - S C AL E TEC TON I C CONT R OLS 163

c09.indd 163 4/7/2011 7:21:13 PM

c09.indd 164 4/7/2011 7:21:13 PM

You might also like