bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Unexpected diversity in socially synchronized rhythms of shorebirds
1
1
2
3
4,5
1,6
7
Martin Bulla , Mihai Valcu , Adriaan M. Dokter , Alexei G. Dondua , András Kosztolányi , Anne Rutten , Barbara Helm , Brett K.
8
9
10
11
12
13
14
15,16
Sandercock , Bruce Casler , Bruno J. Ens , Caleb S. Spiegel , Chris J. Hassell , Clemens Küpper , Clive Minton , Daniel Burgas
,
17
18
19
20
21
22
23
24
David B. Lank , David C. Payer , Egor Y. Loktionov , Erica Nol , Eunbi Kwon , Fletcher Smith , H. River Gates , Hana Vitnerová ,
25
26
27,28
29
30
31
Hanna Prüter , James A. Johnson , James J. H. St Clair
, Jean-François Lamarre , Jennie Rausch , Jeroen Reneerkens , Jesse R.
31
32
33
29
34
35
31
Conklin , Joanna Burger , Joe Liebezeit , Joël Bêty , Jonathan T. Coleman , Jordi Figuerola , Jos C. E. W. Hooijmeijer , José A.
36,37
38
39
40
41
42
43
44
Alves
, Joseph A. M. Smith , Karel Weidinger , Kari Koivula , Ken Gosbell , Larry Niles , Laura Koloski , Laura McKinnon , Libor
39
45
46,47
48
49
50
51
Praus , Marcel Klaassen , Marie-Andrée Giroux
, Martin Sládeček , Megan L. Boldenow , Michael Exo , Michael I. Goldstein ,
48
31,52
40
47
53,54
55,56
57
Miroslav Šálek , Nathan Senner
, Nelli Rönkä , Nicolas Lecomte , Olivier Gilg
, Orsolya Vincze
, Oscar W. Johnson , Paul A.
58
30
59
60
61
26
62
Smith , Paul F. Woodard , Pavel S. Tomkovich , Phil Battley , Rebecca Bentzen , Richard B. Lanctot , Ron Porter , Sarah T.
26
63
64
65
66
67
31,68
Saalfeld , Scott Freeman , Stephen C. Brown , Stephen Yezerinac , Tamás Székely , Tomás Montalvo , Theunis Piersma
,
69
40
70
1
Vanessa Loverti , Veli-Matti Pakanen , Wim Tijsen , Bart Kempenaers
1
2
Department of Behavioural Ecology and Evolutionary Genetics, Max Planck Institute for Ornithology, Eberhard Gwinner Str, 82319 Seewiesen, Germany. Computational
3
Geo-Ecology, Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Science Park 904, 1098 XH, Amsterdam, The Netherlands. Independent
4
researcher, Gatchinskaya, ap. 27, 197198, St-Petersburg, Russia. Department of Ecology, University of Veterinary Medicine Budapest, Rottenbiller u. 50., H-1077, Budapest,
5
Hungary. MTA-DE "Lendület" Behavioural Ecology Research Group, Department of Evolutionary Zoology, University of Debrecen, Egyetem tér 1., H-4032, Debrecen, Hungary.
6
7
Apiloa GmbH, 82319, Starnberg, Germany. Institute of Biodiversity Animal Health and Comparative Medicine, University of Glasgow, Graham Kerr Building, G12 8QQ,
8
9
10
Glasgow, UK. Division of Biology, Kansas State University, 116 Ackert Hall, 66506-4901, Manhattan, KS, USA. PO Box 1094, 89407, Fallon, NV, USA. Coastal Ecology Team,
11
Sovon Dutch Centre for Field Ornithology, PO box 59, 1790 AB, Den Burg, Texel, The Netherlands. Division of Migratory Birds, Northeast Region, US Fish & Wildlife Service,
12
13
300 Westgate Center Dr, 01035, Hadley, MA, USA. Global Flyway Network, PO Box 3089, WA 6725, Broome, Australia. Institute of Zoology, University of Graz,
14
15
Universitätsplatz 2, 8010, Graz, Austria. Victorian Wader Study group, 165 Dalgetty Road, 3193, Beaumaris, Australia. Department of Forest Sciences, University of Helsinki,
16
17
PO Box 27, FI-00014, Helsinki, Finland. Department of Biological and Environmental Sciences, University of Jyväskylä, PO Box 35, FI-40014, Jyväskylä, Finland. Department
18
of Biological Sciences, Simon Fraser University, 8888 University Drive, V5A 1S6, Burnaby, BC, Canada. Alaska Region, U.S. National Park Service, 240 W. 5th Ave, 99501,
Anchorage, AK, USA. 19State Lab for Photon Energetics, Bauman Moscow State Technical University, 2nd Baumanskaya St., 5-1, 105005, Moscow, Russia. 20Biology
21
Department, Trent University, 2140 East Bank Drive, K9L 0G2, Peterborough, ON, Canada. Department of Fish and Wildlife Conservation, Virginia Polytechnic Institute and
22
State University, 310 West Campus Drive, 24061, Blacksburg, VA, USA. Center for Conservation Biology, College of William & Mary and Virginia Commonwealth University,
23
24
PO Box 8795, 23187, Williamsburg, Virginia, USA. ABR, Inc. Environmental Research and Services, PO Box 240268, 99524, Anchorage, AK, USA. Faculty of Science, Charles
25
University in Prague, Albertov 6, 128 43, Praha 2, Czech Republic. Department of Wildlife Diseases, Leibniz Institute for Zoo- and Wildlife Research, Alfred-Kowalke-Straße
17, 10315, Berlin, Germany. 26Migratory Bird Management, US Fish and Wildlife Service, 1011 East Tudor Road, 99503, Anchorage, USA. 27Biodiversity Lab, Department of
28
Biology and Biochemistry, University of Bath, Claverton Down, BA1 7AY, Bath, UK. Centre for Evolutionary Biology, University of Western Australia, Stirling Highway, WA
29
30
6009, Crawley, Australia. Biologie, Chimie et Géographie, Université du Québec à Rimouski , 300, allée des Ursulines, G5L 3A8, Rimouski. QC, Canada. Canadian Wildlife
31
Service, Environment and Climate Change Canada, P.O. Box 2310, 5019 – 52nd Street, 4th Floor, X1A2P7, Yellowknife NT, Canada. Conservation Ecology Group, Groningen
32
Institute for Evolutionary Life Sciences, University of Groningen, Nijenborgh 7, 9747 AG, Groningen, The Netherlands. Division of Life Sciences, Rutgers University, 604‑
33
34
Allison Road, NJ 08854‑8082, Piscataway, USA. Audubon Society of Portland, 5151 NW Cornell Road, 97210, Portland, OR, USA. Queensland Wader Study Group, 22 Parker
35
36
Street, 4128, Queensland, Australia. Department of Wetland Ecology, Doñana Biological Station (CSIC), Av. Americo Vespucio, s/n., 41092, Seville, Spain. Centre for
37
Environmental and Marine Studies (CESAM), Dep. Biology, University of Aveiro, Campus de Santiago, 3810-193, Aveiro, Portugal. South Iceland Research Centre, University
38
39
of Iceland, Fjolheimar, 800, Selfoss, Iceland. Conservation Services International, PO Box 784, 8204, Cape May, NJ, USA. Department of Zoology and Laboratory of
40
Ornithology, Palacký University, 17. listopadu 50, 771 46, Olomouc, Czech Republic. Department of Ecology, University of Oulu, PO Box 3000, 90014, Oulu, Finland.
41
42
43
Australasian Wader Studies Group, 1/19 Baldwin Road, 3130, Blackburn, Australia. LJ Niles Associates, 109 market lane, 8323, Greenwich, USA. Environmental and Life
44
Sciences, Trent University, 1600 West Bank Dr., K0L 0G2, Peterborough, ON, Canada. Bilingual Biology Program, York University Glendon Campus, 2275 Bayview Avenue,
45
46
M4N 3M6, Toronto, ON, Canada. Centre for Integrative Ecology, Deakin University, 75 Pigdons Road, VIC 3216, Waurn Ponds, Australia. Canada Research in Northern
47
Biodiversity and Centre d'Études Nordiques, Université du Québec à Rimouski, 300, allée des Ursulines, G5L 3A8 , Rimouski, QC, Canada. Canada Research in Polar and
48
Boreal Ecology and Centre d'Études Nordiques, Université de Moncton, 18 avenue Antonine-Maillet, E4K 1A6, Moncton, NB, Canada. Faculty of Environmental Sciences,
49
Czech University of Life Sciences Prague, Kamýcká 1176, Suchdol, 16521, Prague, Czech Republic. Depatment of Biology and Wildlife, University of Alaska Fairbanks, PO Box
50
51
756100, 99775-6100, Fairbanks, AK, USA. Vogelwarte Helgoland, Institut für Vogelforschung, An der Vogelwarte 21, D-26386, Wilhelmshaven, Germany. Alaska Coastal
52
Rainforest Center, University of Alaska Southeast, 11120 Glacier Hwy, 99801, Juneau, AK, USA. Cornell Lab of Ornithology, 159 Sapsucker Woods Road, 14850, Ithaca, USA.
53
54
Equipe Ecologie-Evolution, UMR 6282 Biogéosciences, Université de Bourgogne, 6 Bd Gabriel, 21000, Dijon, France. Groupe de Recherche en Ecologie Arctique, 16 rue de
55
Vernot, 21440, Francheville, France. Department of Evolutionary Zoology and Human Biology, University of Debrecen, Egyetem tér 1., H-4032, Debrecen, Hungary.
56
57
Hungarian Department of Biology and Ecology, Babeş-Bolyai University, Clinicilor 5-7, RO-400006, Cluj-Napoca, Romania. Department of Ecology, Montana State
58
59
University, MT 59717, Bozeman, USA. Wildlife Research Division, Environment and Climate Change Canada, 1125 Colonel By Dr., K1A 0H3, Ottawa, ON, Canada. Zoological
Museum, Lomonosov Moscow State University, Bolshaya Nikitskaya St., 6, 125009, Moscow, Russia. 60Institute of Agriculture & Environment, Massey University, Private Bag
61
62
11 222, 4442, Palmerston north, New Zealand. Arctic Beringia Program, Wildlife Conservation Society, 925 Schloesser Dr., 99709, Fairbanks, AK, USA. Delaware Bay
63
64
Shorebird Project, 19002, Ambler, PA, USA. Arctic National Wildlife Refuge, US Fish & Wildlife Service, 101 12th Ave, 99701, Fairbanks, AK, USA. Shorebird Recovery
65
66
Program, Manomet, PO Box 545, 05154, Saxtons River, VT, USA. Fieldday Consulting, V4N 6M5, Surrey, BC, Canada. Milner Centre for Evolution, Department of Biology
67
and Biochemistry, University of Bath, Claverton Down, BA2 7AY, Bath, UK. Servei de Vigilància i Control de Plagues Urbanes, Agència de Salut Pública de Barcelona, Av.
68
Príncep d’Astúries 63, 8012, Barcelona, Spain. Department of Marine Ecology, NIOZ Royal Netherlands Institute for Sea Research, PO Box 59, 1790AB, Den Burg, Texel, The
69
70
Netherlands. Migratory Bird and Habitat Program, U.S. Fish and Wildlife Service, 911 NE 11th Avenue, 97232, Portland, OR, USA. Poelweg 12, 1778 KB, Westerland, The
Netherlands.
1
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
The behavioural rhythms of organisms are thought to be under strong selection, influenced by the
rhythmicity of the environment1-4. Behavioural rhythms are well studied in isolated individuals under
laboratory conditions1,5, but in free-living populations, individuals have to temporally synchronize their
activities with those of others, including potential mates, competitors, prey and predators6-10. Individuals
can temporally segregate their daily activities (e.g. prey avoiding predators, subordinates avoiding
dominants) or synchronize their activities (e.g. group foraging, communal defence, pairs reproducing or
caring for offspring)6-9,11. The behavioural rhythms that emerge from such social synchronization and the
underlying evolutionary and ecological drivers that shape them remain poorly understood5-7,9. Here, we
address this in the context of biparental care, a particularly sensitive phase of social synchronization12 where
pair members potentially compromise their individual rhythms. Using data from 729 nests of 91 populations
of 32 biparentally-incubating shorebird species, where parents synchronize to achieve continuous coverage
of developing eggs, we report remarkable within- and between-species diversity in incubation rhythms.
Between species, the median length of one parent’s incubation bout varied from one to 19 hours, while
period length – the cycle of female and male probability to incubate – varied from six to 43 hours. The
length of incubation bouts was unrelated to variables reflecting energetic demands, but species relying on
crypsis had longer incubation bouts than those that are readily visible or actively protect their nest against
predators. Rhythms entrainable to the 24-h light-dark cycle were less likely at high latitudes and absent in
18 species. Our results indicate that even under similar environmental conditions and despite 24-h
environmental cues, social synchronization can generate far more diverse behavioural rhythms than
expected from studies of individuals in captivity5-7,9. The risk of predation, not the risk of starvation, may be
a key factor underlying the diversity in these rhythms.
Incubation by both parents prevails in almost 80% of non-passerine families13 and is the most common form of
care in shorebirds14. Biparental shorebirds are typically monogamous15, mostly lay three or four eggs in an
open nest on the ground15, and cover their eggs almost continuously13. Pairs achieve this through
synchronization of their activities such that one of them is responsible for the nest at a given time (i.e. an
incubation bout). Alternating female and male bouts generate an incubation rhythm with a specific period
length (cycle of high and low probability for a parent to incubate).
We used diverse monitoring systems (Methods & Extended Data Table 1) to collect data on incubation
rhythms from 91 populations of 32 shorebird species belonging to 10 genera (Fig. 1a), extracted the length of
34,225 incubation bouts from 729 nests, and determined the period length for pairs in 584 nests (see
Methods, Extended Data Fig. 1 & 2).
We found vast between- and within- species variation in incubation bout length and in period length (Fig. 1-3
& Extended Data Fig. 3). Different species, but also different pairs of the same species, adopted strikingly
different incubation rhythms, even when breeding in the same area (see, e.g. incubation rhythms in Barrow,
Alaska, represented by
in Fig. 1b-c; incubation rhythms for each nest are in Supplementary Actograms16).
Whereas in some pairs parents exchanged incubation duties about 20 times a day (Fig. 2a; e.g. Charadrius
semipalmatus, Fig. 1b), in others a single parent regularly incubated for 24 hours (Fig. 2a; e.g. Limnodromus
scolopaceus, Fig. 1b), with exceptional bouts of up to 50 hours (Supplementary Actograms16). Similarly,
whereas incubation rhythms of pairs in 22% of nests followed a strict 24-h period (Fig. 2b; e.g. Tringa flavipes,
Fig. 1b), the rhythms of others dramatically deviated from a 24-h period (Fig. 2b) resulting in ultradian (<20-h
in 12% of nests; e.g. Numenius phaeopus; Fig. 1b), free-running like (e.g. Calidris alpina; Fig. 1b) and infradian
rhythms (>28-h in 8% of nests), some with period lengths up to 48-h (e.g. Limnodromus scolopaceus; Fig. 1b).
This variation in period length partly related to the variation in bout length (Fig. 3): in the suborder Scolopaci
period length correlated positively with median bout length, but in the suborder Charadrii species with 24-h
periods had various bout lengths, and species with similar bout lengths had different period lengths.
2
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Figure 1 | Map of studied
breeding sites and the
diversity
of
shorebird
incubation rhythms. a, Map of
breeding sites with data on
incubation
rhythms.
The
colour of the dots indicates the
genus (data from multiple
species per genus may be
available), the size of the dots
refers to data quality ( ,
large: exact breeding site
known; , small: breeding site
estimated, see Methods). For
nearby
or
overlapping
locations, the dots are
scattered to increase visibility.
Contours of the map made
with
Natural
Earth,
http://www.naturalearthdata.
com.
b-c, Illustrations of
between-species diversity (b)
and within-species diversity (c;
note that the three rhythms
for Western sandpiper and
Ringed plover come from the
same breeding location.). Each
actogram depicts the bouts of
female (yellow; ) and male
(blue-grey; ) incubation at a
single nest over a 24-h period,
plotted twice, such that each
row
represents
two
consecutive days. If present,
twilight is indicated by light
grey
bars
(
)
and
corresponds to the time when
the sun is between 6° and 0°
below the horizon, night is
indicated by dark grey bars (
) and corresponds to the
time when the sun is < 6°
below the horizon. Twilight
and night are omitted in the
centre of the actogram (24:00)
to make the incubation rhythm
visible. a, Between-species
diversity. a-c, The circled
numbers (
) indicate
the breeding site of each pair
and correspond to the
numbers on the map.
3
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Figure 2 | Between- and within-species variation in incubation rhythms and their estimated evolution. a-b, Box plots are ordered by species (within
suborder) from the shortest to the longest median bout length, and depict the genus (colour as in Fig. 1a), median (vertical line inside the box), the 25th
and 75th percentiles (box), 25th percentiles minus 1.5 times interquartile range and 75th percentile plus 1.5 times interquartile range or
minimum/maximum value, whichever is smaller (bars), and the outliers (circles). Nmedian bout length = 729 and Nperiod = 584 nests. b, The red vertical line
indicates a 24-h period. c, Observed and reconstructed incubation bout and period length visualised (by colour) on the phylogenetic tree29 using species’
medians (based on population medians) and one of 100 sampled trees (see Methods). The grey circles represent phylogenetically independent
contrasts30 and hence emphasize the differences at each tree node.
Figure 3 | Relationship between bout and period length. Each dot represents a single nest (N = 584 nests), colours depict the genus. In the suborder
Scolopaci the median bout length and period length correlate positively (rSpearman = 0.56, N = 424 nests); in the suborder Charadrii periods longer than
~24h are absent, and there is no simple relationship between bout and period length (N = 160 nests). For species-specific relationships see Extended
Data Fig. 3.
Despite substantial within-species variation, we found a strong evolutionary signal for both bout and period
length with a coefficient of phylogenetic signal λ close to 1 (Extended Data Table 2). This is consistent with the
notion that biological rhythms are largely genetically determined and conserved among related species 8-10.
However, the phylogenetic effect seems unevenly distributed over taxonomic level. Suborder explained 33% of
the phenotypic variance in both bout and period length, with the Scolopaci having longer incubation bouts and
periods than the Charadrii (Extended Data Table 3; Fig. 2 & 3). Species explained 41% of the phenotypic
variation in bout length and 46% in period length, but genus explained little (<1% in both; Extended Data Table
3), suggesting that despite a strong phylogenetic signal, these traits can rapidly diverge (Fig. 2c).
4
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Figure 4 | Predictors of variation in incubation rhythms. a-c, Relationships between bout length and body size, measured as female wing length (a),
breeding latitude (b) and anti-predation strategy, quantified as escape distance (c) for N = 729 nest from 91 populations belonging to 32 species. d, The
relationship between the proportion of nests with a period length that is entrainable by the 24-h light-dark cycle (i.e. period lengths: 3, 6, 12, 24, or 48h)
and breeding latitude (N = 584 nests from 88 populations belonging to 30 species). e, The distribution of period length over latitude. The period was
standardized to 24h so that all 24-h harmonics are depicted as 24h (red line) and respective deviations from each harmonic as deviations from 24h (e.g.
period of 12.5h is depicted as 25h). a-e, Each circle represents the population median; circle size indicates the number of nests. a-c, The solid lines
depict the model-predicted relationships, the dotted lines the 95% credible intervals based on the joint posterior distribution of 100 separate
MCMCglmm runs each with one of the 100 phylogenetic trees with ~1,100 independent samples per tree. The grey areas depict the predicted
relationships for each of 100 runs (i.e. the full range of regression line estimates across 100 models) and illustrate the uncertainty due to the
phylogenetic tree. The predicted relationships stem from a Gaussian phylogenetic mixed-effect models, where the effects of other predictors were kept
constant (a-c, Extended Data Table 4), or from a binomial phylogenetic mixed-effect model (d, Extended Data Table 4).
Two ecological factors may explain the observed variation in bout length. First, the ‘energetic demands
hypothesis’ stipulates that the length of an incubation bout depends on a bird’s energetic state13,17. This
predicts that (1) large species will have longer incubation bouts than smaller species, because they radiate less
body heat per unit of mass and (2) incubation bouts will shorten with increasing breeding latitude, because –
everything else being equal – energy stores will deplete faster in colder environments (Extended Data Fig. 4a-b
shows latitudinal cline in summer temperatures). However, bout length was unrelated to body size (Fig. 4a)
and correlated positively (instead of negatively) with latitude (Fig. 4b). These correlational results across
populations and species support recent experimental findings within species18 and suggest that in biparentallyincubating shorebirds energetic demands are not an important ecological driver underlying variation in bout
length.
An alternative explanation for variation in the length of incubation bouts relates to anti-predation strategies.
Those species that rely primarily on parental crypsis (Extended Data Fig. 5a) benefit from reduced activity near
the nest, because such activity can reveal the nest’s location to potential predators19,20. Thus, in these species,
selection will favour fewer change-overs at the nest, and hence longer incubation bouts. In contrast, species
that are clearly visible when sitting on the nest or that rely on active anti-predation behaviour (Extended Data
Fig. 5b), including having a partner on the watch for predators, leaving the nest long before the predator is
nearby and mobbing the predator15, obtain no advantage from minimizing activity. For these species, bout
length can shorten, which may be advantageous for other reasons (e.g. reduced need to store fat). We
quantified anti-predation strategy as the distance at which the incubating parent left the nest when
approached by a human (escape distance), because cryptic species stay on the nest longer (often until nearly
stepped upon)15. Despite the large geographical distribution of the studied species, with related variability in
the suite of predators and predation pressure21, and even when controlling for phylogeny (which captures
much of the variation in anti-predation strategy, Extended data Fig. 6), escape distance negatively correlated
with the length of incubation bouts (Fig. 4c). This result suggests that bout length co-evolved with the antipredation strategy.
5
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Under natural conditions, most organisms exhibit 24-h rhythmicity, but during summer, when most shorebirds
breed, the 24-h variation in light decreases with latitude leading to continuous polar daylight22 (Extended Data
Fig. 5c-d). Such reduced variation in 24-h light intensity may cause a loss of 24-h rhythmicity23-25. As a
consequence, circadian behavioural rhythms should exhibit a latitudinal cline22. As predicted, incubation
rhythms with periods that do not follow the 24-h light-dark cycle, such as ‘free running-like patterns’ (left
column in Fig. 1b), occurred more often in shorebirds breeding at higher latitudes (Fig. 4d).The absolute
deviations of periods from 24-h and 24-h harmonics also increased with latitude (Fig. 4e; Extended Data Table
4). Although this supports the existence of a latitudinal cline in socially emerged behavioural rhythms22, we
found a substantial number of rhythms that defy the 24-h day even at low and mid latitudes (Fig 4d-e).
Many shorebirds predominantly use tidal habitats, at least away from their breeding ground15. To anticipate
tidal foraging opportunities, these species may have activity patterns with a period length resembling the tidal
period. As changing to a different rhythm is costly26, these tidal activity patterns might carry over to
incubation. Although half of our species are tidal away from their breeding grounds, and some forage in tidal
areas also during breeding (~12% of populations), in only 5% of nests did pairs display a period length that can
be entrained by the tide. Moreover, tidal species had similar (not longer) periods than non-tidal ones
(Extended Data Table 4). Hence, unlike the 24-h light-dark cycle, tidal life-history seems to play at best a
negligible role in determining incubation rhythms.
Three main questions arise from our results. First, is variation in incubation bout length in cryptic species
related to the actual predation pressure? This can be tested by comparing bout length between populations of
a particular species that are exposed to different predator densities, or between years that differ in predation
pressure. Second, it remains unclear how the diverse social rhythms emerge. Are these rhythms a
consequence of behavioural flexibility, or a ‘fixed’ outcome of synchronization between the circadian clocks of
the two individuals involved? An experimental study on ring doves (Streptopelia risoria) suggests that parents
may even use two timers - circadian oscillation and interval timing - to determine when to incubate27. Parents
rapidly adjusted their schedules to phase-shifted photoperiods and their incubation rhythm ‘free-ran’ in
constant dim illumination (implying a circadian mechanism), whereas an experimental delay in the onset of an
incubation bout did not change the length of the bout because the incubating parent refused to leave the nest
until its incubation bout reached the ‘typical’ duration (implying interval timing). Third, what – if any – are the
fitness consequences for the parents of having a certain incubation rhythm? For example, the costs of having a
particular incubation rhythm may be unevenly distributed between the two parents (e.g. because one parent
is on incubation duty when food is more readily available, or because one parent ‘enforces’ its own rhythm at
a cost to the other parent).
In conclusion, our results reveal that under natural conditions social synchronization can generate much more
diverse rhythms than expected from previous work5-7,9,28, and that theses rhythms often defy the assumptions
of entrainment to the 24-h day-night cycle. Not risk of starvation, but risk of predation seems to play a key role
in determining some of the variation in incubation rhythms. We describe this diversity in the context of
biparental incubation, but diverse behavioural rhythms may also arise in many other social settings (e.g. in the
context of mating interactions25, vigilance behaviour during group foraging). Essentially, the reported diversity
suggests that the expectation that individuals within a pair (or group) should optimize their behavioural
rhythms relative to the 24-h day may be too simplistic, opening up a wide field to study the evolutionary
ecology of plasticity in circadian clocks.
6
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
REFERENCES
1, Dunlap, J. C., Loros, J. J. & DeCoursey, P. J. Chronobiology: Biological Timekeeping (Sinauer Associates, 2004).
2, Young, M. W. & Kay, S. A. Time zones: a comparative genetics of circadian clocks. Nat Rev Genet 2, 702-715 (2001).
3, Helm, B. & Visser, M. E. Heritable circadian period length in a wild bird population. Proc R Soc B (2010).
4, Koskenvuo, M., Hublin, C., Partinen, M., Heikkilä, K. & Kaprio, J. Heritability of diurnal type: a nationwide study of 8753
adult twin pairs. J Sleep Res 16, 156-162 (2007).
5, Kronfeld-Schor, N., Bloch, G. & Schwartz, W. J. Animal clocks: when science meets nature. Proc R Soc B 280 (2013).
6, Bloch, G., Herzog, E. D., Levine, J. D. & Schwartz, W. J. Socially synchronized circadian oscillators. Proc R Soc B 280
(2013).
7, Castillo-Ruiz, A., Paul, M. J. & Schwartz, W. J. In search of a temporal niche: social interactions. Prog Brain Res 199, 267280 (2012).
8, Mistlberger, R. E. & Skene, D. J. Social influences on mammalian circadian rhythms: animal and human studies. Biol Rev
79, 533-556 (2004).
9, Davidson, A. J. & Menaker, M. Birds of a feather clock together–sometimes: social synchronization of circadian
rhythms. Curr Opin Neurobiol 13, 765-769 (2003).
10, Mrosovsky, N. Locomotor activity and non‐photic influences on circadian clocks. Biol Rev 71, 343-372 (1996).
11, Regal, P. J. & Connolly, M. S. Social influences on biological rhythms. Behaviour 72, 171-198 (1980).
12, Emlen, S. T. & Oring, L. W. Ecology, sexual selection, and the evolution of mating systems. Science 197, 215-223
(1977).
13, Deeming, D. C. Avian Incubation: Behaviour, Environment and Evolution (Oxford University Press, 2002).
14, Szekely, T. & Reynolds, J. D. Evolutionary transitions in parental care in shorebirds. Proc R Soc B 262, 57-64 (1995).
15, del Hoyo, J., Elliott, A. & Sargatal, J. Handbook of the Birds of the World. Vol. 3. Hoatzing to Auks. (Lynx Edicions,
1996).
16, Bulla, M. et al. Supporting Information for 'Unexpected diversity in socially synchronized rhythms of shorebirds'.
Version 1. Open Science Framework, https://osf.io/wxufm/ (2016).
17, Williams, J. B. in Avian Energetics and Nutritional Ecology (ed C. Carey) Ch. 5, 375-416 (Chapman & Hall, 1996).
18, Bulla, M., Cresswell, W., Rutten, A. L., Valcu, M. & Kempenaers, B. Biparental incubation-scheduling: no experimental
evidence for major energetic constraints. Behav Ecol 26, 30-37 (2015).
19, Martin, T. E., Scott, J. & Menge, C. Nest predation increases with parental activity: separating nest site and parental
activity effects. Proc R Soc B 267, 2287-2293 (2000).
20, Smith, P. A., Tulp, I., Schekkerman, H., Gilchrist, H. G. & Forbes, M. R. Shorebird incubation behaviour and its influence
on the risk of nest predation. Anim Behav 84, 835-842 (2012).
21, McKinnon, L. et al. Lower predation risk for migratory birds at high latitudes. Science 327, 326-327 (2010).
22, Hut, R. A., Paolucci, S., Dor, R., Kyriacou, C. P. & Daan, S. Latitudinal clines: an evolutionary view on biological rhythms.
Proc R Soc B 280, 20130433 (2013).
23, van Oort, B. E. et al. Circadian organization in reindeer. Nature 438, 1095-1096 (2005).
24, Steiger, S. S. et al. When the sun never sets: diverse activity rhythms under continuous daylight in free-living arcticbreeding birds. Proc R Soc B 280 (2013).
25, Lesku, J. A. et al. Adaptive sleep loss in polygynous pectoral sandpipers. Science 337, 1654-1658 (2012).
26, Foster, R. G. & Wulff, K. The rhythm of rest and excess. Nat Rev Neurosci 6, 407-414 (2005).
27, Silver, R. & Bittman, E. L. Reproductive Mechanisms: Interaction of Circadian and Interval Timing. Ann N Y Acad Sci
423, 488-514 (1984).
28, Paul, M. J., Indic, P. & Schwartz, W. J. Social synchronization of circadian rhythmicity in female mice depends on the
number of cohabiting animals. Biol Lett 11, 20150204 (2015).
29, Revell, L. J. & Freckleton, R. Two new graphical methods for mapping trait evolution on phylogenies. Methods. Ecol.
Evol. 4, 754-759, http://dx.doi.org/10.1111/2041-210x.12066 (2013).
30, Felsenstein, J. Phylogenies and the comparative method. Am Nat 125, 1-15 (1985).
7
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Supplementary Information is freely available at open science framework https://osf.io/wxufm/
16
Acknowledgements We thank all that made the data collection possible. We are grateful to W. Schwartz, E. Schlicht, W.
Forstmeier, M. Baldwin, H. Fried Petersen, D. Starr-Glass, and B. Bulla for comments on the manuscript and to F. KornerNievergelt, J. D. Hadfield, L. Z. Garamszegi, S. Nakagawa, T. Roth, N. Dochtermann, Y. Araya, E. Miller and H. Schielzeth for
advice on data analysis. Data collection was supported by various institutions and people listed in the Supplementary
16
Data 1: https://osf.io/sq8gk . The study was supported by the Max Planck Society (to B.K.). M.B. is a PhD student in the
International Max Planck Research School for Organismal Biology.
Author Contributions M.B. and B.K. conceived the study. All authors except B.H. collected the primary data. MB
coordinated the study and managed the data. MB and M.V. developed the methods to extract incubation. M.B. extracted
bout lengths and with help from A.R. and M.V. created actograms. M.B. with help from M.V. analysed the data. M.B.
prepared the Supporting Information. M.B. and B.K. wrote the paper with input from the other authors. Except for the
first, second, and last author, the authors are listed alphabetically by their first name.
Author Information All information, primary and extracted data, computer code and software necessary to replicate our
results, as well as the Supplementary Actograms are open access and archived at Open Science Framework
16
https://osf.io/wxufm/ . The authors declare no competing financial interests. Correspondence and requests for
materials should be addressed to M.B. (bulla.mar@gmail.com) and B.K. (b.kempenaers@orn.mpg.de).
8
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
METHODS
Recording incubation. Incubation data were obtained between 1994 and 2015, for as many shorebird species
(N = 32) and populations (N = 91) as possible, using six methods (for specifications of the equipment see
Extended Data Table 1). (1) In 261 nests, a radio frequency identification reader (‘RFID’) registered presence of
tagged parents at the nest. The passive-integrated tag was either embedded in a plastic flag31,32, with which
the parents were banded, or glued to the tail feathers33. In 200 nests the RFID was combined with a
temperature probe placed between the eggs. The temperature recordings allowed us to identify whether a
bird was incubating even in the absence of RFID readings; an abrupt change in temperature demarcated the
start or end of incubation31. (2) For 396 nests, light loggers were mounted to the plastic flag or a band that was
attached to the bird’s leg34,35. The logger recorded maximum light intensity (absolute or relative) for a fixed
sampling interval (2-10 min). An abrupt change in light intensity (as opposed to a gradual change caused, e.g.
by civil twilight) followed by a period of low or high light intensity demarcated the start or end of the
incubation period (Extended Data Fig. 2). (3) For nine nests a GPS tag, mounted on the back of the bird,
recorded the position of the bird36. The precision of the position depends on cloud cover and sampling
interval36. Hence, to account for the imprecision in GPS positions, we assumed incubation whenever the bird
was within 25 m of the nest (Extended Data Fig. 2b). (4) At three nests automated receivers recorded signal
strength of a radio-tag attached to the rump of a bird; whenever a bird incubated, the strength of the signal
remained constant24 (Supplementary Actograms p. 257-916). (5) At 53 nests video cameras and (6) for 8 nests
continuous observations were used to identify the incubating parents; parent identification was based on
plumage, colour rings or radio-tag. In one of the populations, three different methods were used, in seven
populations representing seven species two methods. In one nest, two methods were used simultaneously
(Extended Data Fig. 2b).
Extraction of incubation bouts. An incubation bout was defined as the total time allocated to a single parent
(i.e. the time between the arrival of a parent at and its departure from the nest followed by incubation of its
partner). Bout lengths were only extracted if at least 24h of continuous recording was available for a nest; in
such cases, all bout lengths were extracted. For each nest, we transformed the incubation records to local time
as (UTC time +
nest′s longitude
15
). Incubation bouts from RFIDs, videos and continuous observations were mostly
extracted by an R-script and the results verified by visualizing the extracted and the raw data16,31,37,38;
otherwise, MB extracted the bouts manually from plots of raw data39,40 (plots of raw data and extracted bouts
for all nests are in the Supplementary Actograms16; the actograms were generated by ‘ggplot’ and ‘xyplot’
functions from the ‘ggplot2’ and ‘lattice’ R package41-43). Whenever the start or end of a bout was unclear, we
classified these bouts as uncertain (see next paragraph for treatment of unsure bouts). In case of light logger
data, the light recordings before and after the breeding period, when the birds were definitely not incubating,
helped to distinguish incubation from non-incubation. Whenever an individual tagged with a light logger
nested in an environment where the sun was more than 6° below the horizon for part of a day (i.e. night), we
assumed an incubation bout when the individual started incubating before the night started and ended
incubating after the night ended. When different individuals incubated at the beginning vs. at the end of the
night, we either did not quantify these bouts or we indicated the possible time of exchange (based on trend in
previous exchanges), but classified these bouts as uncertain (see Supplementary Actograms16). In total, we
extracted 34,225 incubation bouts.
The proportion of uncertain bouts within nests had a distribution skewed towards zero (median = 0%, range:
0-100%, N = 729 nests), and so did the median proportion of uncertain bouts within populations (median = 2%,
range: 0-74%, N = 91 populations). Excluding the uncertain bouts did not change our estimates of median bout
length (Pearson’s correlation coefficient for median bout length based on all bouts and without uncertain
9
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
bouts: r = 0.96, N = 335 nests with both certain and uncertain bouts). Hence, in further analyses all bouts were
used to estimate median bout length.
Note that in some species sexes consistently differ in bout length (Figure 1b, e.g. Northern lapwing). As these
differences are small compared to the between-species differences and because in 27 nests (of 8 species) the
sex of the parents was unknown, we here use median bout length independent of sex.
Extraction of period length. The method used for extracting the period length of incubation rhythm for each
nest is described in the Extended Data Fig. 1.
Extraction of entrainable periods. We classified 24-h periods and periods with 24-h harmonics (i.e. 3, 6, 12,
48h) as strictly entrainable by 24-h light fluctuations (N = 142 nests out of 584). Including also nearest adjacent
periods (±0.25h) increased the number of nests with entrainable periods (N = 277), but results of statistical
analyses remained quantitatively similar. We consider periods and harmonics of 12.42h (i.e. 3.1, 6.21, 12.42,
24.84h) as strictly entrainable by tide. However, because the periods in our data were extracted in 0.25-h
intervals (Extended Data Fig. 1), we classified periods of 3, 6.25, 12.5, 24.75h (i.e. those closest to the strict
tide harmonics) as entrainable by tide (N = 32 nests out of 584). Including also the second nearest periods (i.e.
3.25, 6, 12.25, 25) increased the number of nests entrainable by tide to N = 55.
Population or species life-history traits. For 643 nests, the exact breeding location was known (nests or
individuals were monitored at the breeding ground). For the remaining 86 nests (from 27 populations
representing 8 species, where individuals were tagged with light loggers on the wintering ground), the
breeding location was roughly estimated from the recorded 24-h variation in daylight, estimated migration
tracks, and the species’ known breeding range44-51. One exact breeding location was in the Southern
Hemisphere, so we used absolute latitude in analyses. Analyses without populations with estimated breedinglocation or without the Southern Hemisphere population generated quantitatively similar estimates as the
analyses on full data.
For each population, body size was defined as mean female wing length52, either for individuals measured at
the breeding area or at the wintering area. In case no individuals were measured, we used the mean value
from the literature (see open access data for specific values and references53).
Anti-predation strategy was assessed by estimating escape distance of the incubating bird when a human
approached the nest, because species that are cryptic typically stay on the nest much longer than non-cryptic
species, sometimes until nearly stepped upon48,54. Escape distance was obtained for all species. Forty-four
authors of this paper estimated the distance (in m) for one or more species based on their own data or
experience. For 10 species, we also obtained estimates from the literature48. We then used the median
‘estimated escape distance’ for each species. In addition, for 13 species we obtained ‘true escape distance’.
Here, the researcher approached a nest (of known position) and either estimated his distance to the nest or
marked his position with GPS when the incubating individual left the nest. For each GPS position, we
calculated the Euclidian distance from the nest. In this way we obtained multiple observations per nest and
species, and we used the median value per species (weighted by the number of estimates per nest) as the
‘true escape distance’. The species’ median ‘estimated escape distance’ was a good predictor of the ‘true
escape distance’ (Pearson’s correlation coefficient: r = 0.89, N = 13 species). For analysis, we defined the
escape distance of a species as the median of all available estimates.
For each species, we determined whether it predominantly uses a tidal environment outside its breeding
ground, i.e. has tidal vs. non-tidal life history (based on48,50,51). For each population with exact breeding
location, we scored whether tidal foraging habitats were used by breeding birds for foraging (for three
10
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
populations this information was unknown)53. For all populations with estimated breeding location we
assumed, based on the estimated location and known behaviour at the breeding grounds, no use of tidal
habitat.
Statistical analyses. Unless specified otherwise, all analyses were performed on the nest level using median
bout length and extracted period length.
We used phylogenetically informed comparative analyses to assess how evolutionary history constrains the
incubation rhythms (estimated by Pagel’s λ coefficient of phylogenetic signal55,56) and to control for potential
non-independence among species due to common ancestry. This method explicitly models how the covariance
between species declines as they become more distantly related55,57,58. We used the Hackett59 backbone
phylogenetic trees available at http://birdtree.org60, which included all but one species (Charadrius nivosus)
from our dataset. Following a subsequent taxonomic split61, we added Charadrius nivosus to these trees as a
sister taxon of Charadrius alexandrinus. Phylogenetic uncertainty was accounted for by fitting each model with
100 phylogenetic trees randomly sampled from 10,000 phylogenies at http://birdtree.org60.
The analyses were performed with Bayesian phylogenetic mixed-effect models (Fig. 4 and Extended Data Table
2 and 4) and the models were run with the ‘MCMCglmm’ function from the R package ‘MCMCglmm’ 62. In all
models, we also accounted for multiple sampling within species and breeding site (included as random
effects). In models with a Gaussian response variable, an inverse-gamma prior with shape and scale equal to
0.001 was used for the residual variance (i.e. variance set to one and the degree of belief parameter to 0.002).
In models with binary response variables, the residual variance was fixed to one. For all other variance
components the parameter-expanded priors were used to give scaled F-distributions with numerator and
denominator degrees of freedom set to one and a scale parameter of 1,000. Model outcomes were insensitive
to prior parameterization. The MCMC chains ran for 2,753,000 iterations with a burn-in of 3,000 and a thinning
interval of 2,500. Each model generated ~1,100 independent samples of model parameters (Extended Data
Table 2 and 4). Independence of samples in the Markov chain was assessed by tests for autocorrelation
between samples and by using graphic diagnostics.
First, we used MCMCglmm to estimate Pagel’s λ (phylogenetic signal) for bout and period length (Gaussian),
and to show that our estimates of these two incubation variables were independent of how often the
incubation behaviour was sampled (‘sampling’ in min, ln-transformed; Extended Data Table 2). Hence, in
subsequent models, sampling was not included.
Then, we used MCMCglmm to model variation in bout length and period length. Bout length was modelled as
a continuous response variable and latitude (°, absolute), female wing length (mm, ln-transformed) and
approach distance (m, ln-transformed) as continuous predictors. Predictors had low collinearity (at nest,
population and species level; all Pearson or Spearman correlation coefficients |r| < 0.28). To test for potential
entrainment to 24-h, period length was modelled as a binary response variable (1 = rhythms with period of 3,
6, 12, 24, or 48 h; 0 = rhythms with other periods) and latitude as a continuous predictor. To test how circadian
period varies with latitude or life history, period was transformed to deviations from 24-h and 24-h harmonics
and scaled by the time span between the closest harmonic and the closest midpoint between two harmonics.
For example, a 42h period deviates by -6h from 48h (the closest 24-h harmonic) and hence -6h was divided by
12h (the time between 36h – the midpoint of two harmonics - and 48h -the closest harmonic). This way the
deviations spanned from -1 to 1 with 0 representing 24-h and its harmonics. The absolute deviations were
then modelled as a continuous response variable and latitude as continuous predictor. The deviations were
also modelled as a continuous response and species life history (tidal or not) as categorical predictor.
11
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
In all models the continuous predictor variables were centred and standardized to a mean of zero and a
standard deviation of one.
We report model estimates for fixed and random effects, as well as for Pagel’s λ, by the modes and the
uncertainty of the estimates by the highest posterior density intervals (referred to as 95% CI) from the joint
posterior distributions of all samples from the 100 separate runs, each with one of the 100 separate
phylogenetic trees from http://birdtree.org.
To help interpret the investigated relationships we assessed whether incubation rhythms evolved within
diverged groups of species by plotting the evolutionary tree of the incubation rhythm variables (Fig. 2c), as
well as of the predictors (Extended Data Fig. 6).
The source of phylogenetic constraint in bout and period length was investigated by estimating the proportion
of phenotypic variance explained by suborder, genus and species (Extended Data Table 3). The respective
mixed models were also specified with ‘MCMCglmm’62 using the same specifications as in the phylogenetic
models. Because suborder contained only two levels, we first fitted an intercept mixed model with genus,
species, and breeding site as random factors, and used it to estimate the overall phenotypic variance. We then
entered suborder as a fixed factor and estimated the variance explained by suborder as the difference
between the total variance from the first and the second model. To evaluate the proportion of the variance
explained by species, genus and breeding site, we used the estimates from the model that included suborder.
R version 3.1.163 was used for all statistical analyses.
Code availability. All statistical analyses, figures, and the Supplementary Actograms are replicable with the
open access information, software and r-code available from Open Science Framework, https://osf.io/
wxufm/16.
Data availability. Primary and extracted data that support the findings of this study are freely available from
Open Science Framework, https://osf.io/wxufm/16.
12
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
REFERENCES
31, Bulla, M., Valcu, M., Rutten, A. L. & Kempenaers, B. Biparental incubation patterns in a high-Arctic breeding shorebird:
how do pairs divide their duties? Behav Ecol 25, 152-164 (2014).
32, Reneerkens, J., Grond, K., Schekkerman, H., Tulp, I. & Piersma, T. Do uniparental Sanderlings Calidris alba increase egg
heat
input
to
compensate
for
low
nest
attentiveness?
PLoS
ONE
6,
e16834,
http://dx.doi.org/10.1371/journal.pone.0016834 (2011).
33, Kosztolányi, A. & Székely, T. Using a transponder system to monitor incubation routines of Snowy Plovers. J Field
Ornithol 73, 199-205 (2002).
34, Conklin, J. R. & Battley, P. F. Attachment of geolocators to bar-tailed godwits: a tibia-mounted method with no
survival effects or loss of units. Wader Study Group Bull 117, 56-58 (2010).
35, Burger, J., Niles, L. J., Porter, R. R. & Dey, A. D. Using geolocator data to reveal incubation periods and breeding
biology in Red Knots Calidris canutus rufa. Wader Study Group Bull 119, 26-36 (2012).
36, Bouten, W., Baaij, E. W., Shamoun-Baranes, J. & Camphuysen, K. C. J. A flexible GPS tracking system for studying bird
behaviour at multiple scales. J Ornithol 154, 571-580 (2012).
37, Bulla, M. R-SCRIPT and EXAMPLE DATA to extract incubation from temperature measurements. Version 1. figshare,
https://dx.doi.org/10.6084/m9.figshare.1037545.v1 (2014).
38, Bulla, M. R-SCRIPT and EXAMPLE DATA to extract incubation bouts from continuous RFID and video data. Version 1.
figshare, http://dx.doi.org/10.6084/m9.figshare.1533278.v1 (2015).
39, Bulla, M. Example of how to manually extract incubation bouts from interactive plots of raw data - R-CODE and DATA.
Version 1. figshare, https://dx.doi.org/10.6084/m9.figshare.2066784.v1 (2016).
40, Bulla, M. Procedure for manual extraction of incubation bouts from plots of raw data.pdf. Version 1. figshare,
https://dx.doi.org/10.6084/m9.figshare.2066709.v1 (2016).
41, Wickham, H. ggplot2: Elegant Graphics for Data Analysis (Springer Publishing Company, Incorporated, 2009).
42, Sarkar, D. & Andrews, F. latticeExtra: Extra Graphical Utilities Based on Lattice. R package version 0.6-24.
http://CRAN.R-project.org/package=latticeExtra (2012).
43, Sarkar, D. Lattice: Multivariate Data Visualization with R. (Springer, 2008).
44, Lisovski, S. Geolocator-ArcticWader-BreedingSiteEstimation. Version 2015-08-05. GitHub repository,
https://github.com/slisovski/Geolocator-ArcticWader-BreedingSiteEstimation (2015).
45, Lisovski, S. & Hahn, S. GeoLight – processing and analysing light-based geolocator data in R. Methods. Ecol. Evol. 3,
1055-1059, https://dx.doi.org/10.1111/j.2041-210X.2012.00248.x (2012).
46, Lisovski, S. et al. Geolocation by light: accuracy and precision affected by environmental factors. Methods. Ecol. Evol.,
1-10, https://dx.doi.org/10.1111/j.2041-210X.2012.00185.x (2012).
47, Conklin, J. R., Battley, P. F., Potter, M. A. & Fox, J. W. Breeding latitude drives individual schedules in a transhemispheric migrant bird. Nat Commun 1, 67 (2010).
48, Poole, A. The Birds of North America Online (Cornell Laboratory of Ornithology, 2005).
49, Lappo, E., Tomkovich, P. & Syroechkovskiy, E. Atlas of Breeding Waders in the Russian Arctic (UF Ofsetnaya Pecha,
2012).
50, Chandler, R. J. Shorebirds of the Northern Hemisphere (Christopher Helm, 2009).
51, Brazil, M. Birds of East Asia: Eastern China, Taiwan, Korea, Japan, and Eastern Russia (Christohper Helm, 2009).
52, Dale, J. et al. Sexual selection explains Rensch's rule of allometry for sexual size dimorphism. Proc R Soc B 274, 29712979 (2007).
53, Bulla, M. et al. Supplementary Data 3 - Study sites: location, population wing length, monitoring method, tide. Version
11. figshare, https://dx.doi.org/10.6084/m9.figshare.1536260.v11 (2016).
54, Cramp, S. Handbook of the Birds of Europe, the Middle East, and North Africa: The Birds of the Western Palearctic
Volume III: Waders to Gulls (Oxford University Press, 1990).
55, Freckleton, R. P., Harvey, P. H. & Pagel, M. Phylogenetic analysis and comparative data: a test and review of evidence.
Am Nat 160, 712-726 (2002).
56, Pagel, M. Inferring evolutionary processes from phylogenies. Zool Scr 26, 331-348 (1997).
57, Martins, E. P. & Hansen, T. F. Phylogenies and the comparative method: a general approach to incorporating
phylogenetic information into the analysis of interspecific data. Am Nat, 646-667 (1997).
58, Pagel, M. Inferring the historical patterns of biological evolution. Nature 401, 877-884 (1999).
13
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
59, Hackett, S. J. et al. A phylogenomic study of birds reveals their evolutionary history. Science 320, 1763-1768 (2008).
60, Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K. & Mooers, A. O. The global diversity of birds in space and time. Nature
491, 444-448 (2012).
61, Küpper, C. et al. Kentish versus Snowy Plover: phenotypic and genetic analyses of Charadrius alexandrinus reveal
divergence of Eurasian and American subspecies. Auk 126, 839-852 (2009).
62, Hadfield, J. D. MCMC methods for multi-response generalized linear mixed models: the MCMCglmm R package. J Stat
Softw 33, 1-22 (2010).
63, R-Core-Team. R: A Language and Environment for Statistical Computing. Version 3.1.1. R Foundation for Statistical
Computing, http://www.R-project.org/ (2014).
64, Anderson, D. R. Model Based Inference in the Life Sciences: A Primer on Evidence (Springer, 2008).
65, Hijmans, R. J. raster: Geographic data analysis and modeling. R package version 2.3-24. http://CRAN.Rproject.org/package=raster (2015).
66, Bivand, R. & Lewin-Koh, N. maptools: Tools for reading and handling spatial objects. R package version 0.8-30.
http://CRAN.R-project.org/package=maptools (2014).
67, Revell, L. J. in Modern Phylogenetic Comparative Methods and Their Application in Evolutionary Biology (ed L. Z.
Garamszegi) Ch. 4, 77-103 (Springer, 2014).
68, Johnson, O. W. et al. Tracking Pacific Golden-Plovers Pluvialis fulva: transoceanic migrations between non-breeding
grounds in Kwajalein, Japan and Hawaii and breeding grounds in Alaska and Chukotka. Wader Study 122, 13-20
(2015).
69, Kosztolanyi, A., Cuthill, I. C. & Szekely, T. Negotiation between parents over care: reversible compensation during
incubation. Behav Ecol 20, 446-452 (2009).
70, St Clair, J. J. H., Herrmann, P., Woods, R. W. & Székely, T. Female-biased incubation and strong diel sex-roles in the
Two-banded Plover Charadrius falklandicus. J Ornithol, 1-6 (2010).
71, Spiegel, C. S., Haig, S. M., Goldstein, M. I. & Huso, M. Factors affecting incubation patterns and sex roles of Black
Oystercatchers in Alaska. Condor 114, 123-134 (2012).
72, Praus, L. & Weidinger, K. Predators and nest success of Sky Larks Alauda arvensis in large arable fields in the Czech
Republic. Bird Study 57, 525-530 (2010).
73, Orme, D. et al. caper: Comparative Analyses of Phylogenetics and Evolution in R. R package version 0.5.2.
http://CRAN.R-project.org/package=caper (2013).
14
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
EXTENDED DATA
Extended Data Figure 1 | Extracting period length of incubation rhythms. a-c, Each column represents an example for a specific nest
with long, intermediate and short incubation bouts. a, From the extracted bout lengths we created a time series that indicated for each
nest, and every 10 min, whether a specific parent (female, if sex was known) incubated or not. Exchange gaps (no parent on the nest)
had to be < 6 h to be included (for treatment of exchange gaps > 6 h see d, e). b, We then estimated the autocorrelation for each 10
63
min time-lag up to 4 days (R ‘acf’ function ). Positive values indicate a high probability that the female was incubating, negative values
indicate that it was more likely that the male was incubating. We used only nests that had enough data to estimate the autocorrelation
pattern (N = 584 nests from 88 populations of 30 species). The visualized autocorrelation time series never resembled white or random
noise indicative of an arrhythmic incubation pattern. To determine the period (i.e. cycle of high and low probability for a parent to
incubate) that dominated the incubation rhythm, we fitted to the autocorrelation estimates a series of periodic logistic regressions. In
each regression, the time lag (in hours) transformed to radians was represented by a sine and cosine function
, where f(t) is the autocorrelation at time-lag t, a0 is the intercept, b is the slope for sine and c the slope
for cosine, T represents the length of the fitted period (in hours), and e is an error term. We allowed the period length to vary from 0.5
64
h to 48 h (in 15 min intervals, giving 191 regressions). c, By comparing the Akaike’s Information Criterion (AIC) of all regressions, we
estimated, for each nest, the length of the dominant period in the actual incubation data (best fit). Regressions with ΔAIC (AICmodelAICmin) close to 0 are considered as having strong empirical support, while models with ΔAIC values ranging from 4-7 have less
64
support . In 73% of all nests, we determined a single best model with ΔAIC <= 3 (c, middle ΔAIC graph), in 20% of nests two best
models emerged and in 6% of nests 3 or 4 models had ΔAIC <= 3 (c, left and right ΔAIC graphs). However, in all but three nests, the
models with the second, third, etc. best ΔAIC were those with period lengths closest to the period length of the best model (c, left and
right ΔAIC graphs). This suggests that multiple periodicities are uncommon. d-e, The extraction of the period length (described in a-c)
requires continuous datasets, but some nests had long (>6 h) gaps between two consecutive incubation bouts, for example because of
equipment failure or because of unusual parental behaviour. In such cases, we excluded the data from the end of the last bout until the
same time the following day, if data were then available again (d), or we excluded the entire day (e).
15
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Figure 2 | Extracting incubation bouts from light logger data. a, An example of a nest with a light intensity signal from
both parents (yellow line
and blue-grey line
). The incubation bouts for a given parent reflect periods dominated by lower
light values compared to those of the partner. Note the sharp drop in the light levels at the beginning of each incubation bout and the
sharp increase in the light levels at the end. Change-overs between partners occur when the light signal lines cross. Such pronounced
changes in light intensity detected by the logger were used to assign incubation even when only a single parent was tagged. Note that
after the chicks hatch and leave the nest (July 9, vertical bar), the light intensity signal from both parents remains similar. b, An example
of a nest where one incubating parent was simultaneously equipped with a light logger and with a GPS tag. The yellow line (
)
indicates light levels, red dots ( ) indicate the distance of the bird to the nest. As expected, low light levels co-occur with close proximity
to the nest, and hence reflect periods of incubation. Although light levels decrease during twilight (light grey horizontal bar;
), the
recordings were still sensitive enough to reflect periods of incubation, i.e. the light signal matches the distance (e.g. May-25: female
incubated during dawn, but was off the nest during dusk). a-b, Rectangles in the background indicate extracted female (light yellow
polygon, ) and male (light blue-grey polygon, ) incubation bouts.
16
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Figure 3 | Relationship between bout and period length for 30 shorebird species. Each dot represents one nest (N =
584 nests), colours indicate the genus.
17
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Figure 4 | Ecological correlates of latitude. a, Variation in minimum temperature across the globe represented by mean
minimum June temperature for the Northern Hemisphere and mean minimum December temperature for the Southern Hemisphere.
b, Correlation between absolute latitude and the mean minimum temperature based on the month represented by mid-day of
incubation data for each nest (N = 729). For maximum temperature the correlation was similar (r = -0.91, N = 729 nests). c, Daily
variation in sun elevation (i.e. in light conditions) represented as the difference between the noon and midnight sun-elevation for the
summer solstice in the Northern Hemisphere and the winter solstice in the Southern Hemisphere. d, Correlation between absolute
latitude and daily variation in sun elevation for mid-day of incubation data for each nest (N = 729 nests). The points are jittered, as else
they form a straight line. a, c, Red points indicate the breeding sites (N = 91). a-b, The minimum and maximum monthly temperature
65
data were obtained from www.worldclim.org using the ‘raster’ R-package .c-d, Sun-elevation was obtained by the ‘solarpos’ function
66
from the ‘maptools’ R-package .
Extended Data Figure 5 | Between-species variation in parental crypsis during incubation. a-b, Shorebirds vary in how visible they are
on the nest while incubating. The nearly invisible Great knot (Calidris tenuirostris; a; central and facing right) sits tight on the nest when
approached by a human until nearly stepped upon. In contrast, the conspicuous Eurasian oystercatcher (Haematopus ostralegus; b) is
visible on the nest from afar and when approached by a human leaves the nest about 100 m in advance (Credits: a, M. Šálek; b, Jan van
de Kam).
18
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Figure 6 | Phylogenetic relationships for predictors. a, Body size, estimated as female wing length. b, Latitude
29,67
using species’ medians (a-b; based on population
(absolute), c, Escape distance. a-c, We visualised the evolution of these traits
medians), species’ estimates of escape distance (c) and one of the 100 sampled trees (see Methods).
19
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Table 1 | Incubation monitoring methods and systems.
53
For details about methods used in each populations, see Supplementary Data .
*At one nest a bird with MK logger was recaptured and the logger exchanged for Intigeo logger. This nest appears in N for both logger
types.
**Simultaneously equipped with light logger (Intigeo). This nest appears in N for both GPS-tracer and Intigeo.
20
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Table 2 | Effects of phylogeny and sampling on bout length and period length.
The posterior estimates (modes) of the effect sizes with the highest posterior density intervals (95% CI) and the median and range of
the effective sample sizes (N (range)) come from the joint posterior distribution of 100 separate runs each with one of 100 separate
phylogenetic trees from http://birdtree.org. Nbout = 729 nests from 91 populations belonging to 32 species. Nperiod = 584 nests from 88
populations belonging to 30 species. Sampling (how often the incubation behaviour was sampled) was ln-transformed and then meancentred and scaled (divided by SD). For procedures and specifications related to phylogenetic Bayesian mixed models see Methods.
Estimating Pagel’s λ on the species level (Nbout = 32 species, Nperiod = 30 species) with phylogenetic generalized least-squares using the
73
function ‘pgls’ from the R-package ‘caper’ gave similar results (median (range) λbout = 0.73 (0.63-1) and λperiod = 0.95 (0.64-1), based on
100 estimates each for one of the 100 trees).
Extended Data Table 3 | Source of phylogenetic signal
The posterior estimates (modes) of the effect sizes with the highest posterior density intervals (95% CI) and the effective sample sizes
62
(N) come from a posterior distribution of 1,100 simulated values generated by ‘MCMCglmm’ in R . Nbout = 729 nests from 91
populations belonging to 32 species. Nperiod= 584 nest from 88 populations belonging to 30 species.
21
bioRxiv preprint first posted online Nov. 1, 2016; doi: http://dx.doi.org/10.1101/084806. The copyright holder for this preprint (which was
not peer-reviewed) is the author/funder. It is made available under a CC-BY-NC-ND 4.0 International license.
Extended Data Table 4 | Effect of latitude, body size, escape distance and life history on biparental incubation rhythms in shorebirds
The posterior estimates (modes) of the effect sizes with the highest posterior density intervals (95% CI) and the median and range of
the effective sample sizes (N (range)) come from the joint posterior distribution of 100 separate runs each with one of the 100 separate
phylogenetic trees from http://birdtree.org. Nbout = 729 nests from 91 populations belonging to 32 species. For models on light
entrainable rhythm, absolute deviations and deviations from 24-h: N = 584 nests from 88 populations belonging to 30 species. Latitude
(in bout model: absolute value), wing length (ln-transformed), and escape distance (ln-transformed) were mean-centred and scaled
(divided by SD). The estimates for the light- entrainable rhythm are on a binomial scale. For procedures and specifications related to
phylogenetic Bayesian mixed models see Methods.
22