ArticlePDF Available

Qualitative bias in offspring investment in a superorganism is linked to dispersal and nest inheritance

Authors:

Abstract and Figures

How parents invest resources in offspring is a central aspect of life history. While investment strategies have been well studied in solitary organisms, comparatively little has been done on social species, including the many that reproduce by fission. Under colony fission, a parent colony divides resources (individuals) to form two or more offspring colonies. Because individuals differ in characteristics (e.g. size), there is opportunity for both quantitative and qualitative bias in their allocation. In this study we investigated the qualitative aspect of offspring investment during colony fission. Colonies of the ant Cataglyphis cursor fission into multiple offspring colonies as part of their lifecycle, and the distribution of workers is quantitatively biased.We found that investment is also qualitatively biased in terms of worker size and worker genetic characteristics (patrilines). This bias was mainly between the offspring colony that inherited the original nest and offspring colonies that dispersed to new nesting sites. In 74% of cases, dispersing colonies contained larger workers, and the distribution of genetic patrilines was biased in two of six cases in a manner that cannot be explained by the observed variation in worker size between patrilines. Fission also led to a reduction in diversity in offspring colonies compared to the parent colony, in terms of both worker size (70% of cases) and genetic diversity (40% of cases). These patterns are probably the result of differing dispersal probability between workers of different patrilines and of different size. This differential allocation may be adaptive because larger workers may be of disproportionate value to dispersing colonies, and their loss an acceptable cost to the colony inheriting the nest.
Content may be subject to copyright.
Qualitative bias in offspring investment in a superorganism is linked
to dispersal and nest inheritance
Adam L. Cronin
a
,
*
, Thibaud Monnin
b
, David Sillam-Duss
es
b
,
c
,
d
, Fabien Aubrun
b
,
Pierre F
ed
erici
b
, Claudie Doums
e
,
f
a
United Graduate School of Agricultural Sciences, Iwate University, Morioka, Japan
b
Sorbonne Universit
es, UPMC Univ Paris 06, UMR 7618 Institute of Ecology and Environmental Sciences of Paris, Paris, France
c
Sorbonne Universit
es, Institut de Recherche pour le D
eveloppement, U242 Institute of Ecology and Environmental Sciences of Paris, Bondy, France
d
Laboratoire d'
Ethologie Exp
erimentale et Compar
ee, Universit
e Paris 13eSorbonne Paris Cit
e, Villetaneuse, France
e
Institut de Syst
ematique,
Evolution, Biodiversit
e (ISYEB), EPHE, CNRS UPMC Univ Paris 06, MNHN, Sorbonne Universit
es, Paris, France
f
EPHE, PSL Research University, Paris, France
article info
Article history:
Received 31 March 2016
Initial acceptance 2 May 2016
Final acceptance 31 May 2016
MS number 16-00286R
Keywords:
dependent colony foundation
Formicidae
nepotism
polyandry
reproductive strategy
How parents invest resources in offspring is a central aspect of life history. While investment strategies
have been well studied in solitary organisms, comparatively little has been done on social species,
including the many that reproduce by ssion. Under colony ssion, a parent colony divides resources
(individuals) to form two or more offspring colonies. Because individuals differ in characteristics (e.g.
size), there is opportunity for both quantitative and qualitative bias in their allocation. In this study we
investigated the qualitative aspect of offspring investment during colony ssion. Colonies of the ant
Cataglyphis cursor ssion into multiple offspring colonies as part of their lifecycle, and the distribution of
workers is quantitatively biased. We found that investment is also qualitatively biased in terms of worker
size and worker genetic characteristics (patrilines). This bias was mainly between the offspring colony
that inherited the original nest and offspring colonies that dispersed to new nesting sites. In 74% of cases,
dispersing colonies contained larger workers, and the distribution of genetic patrilines was biased in two
of six cases in a manner that cannot be explained by the observed variation in worker size between
patrilines. Fission also led to a reduction in diversity in offspring colonies compared to the parent colony,
in terms of both worker size (70% of cases) and genetic diversity (40% of cases). These patterns are
probably the result of differing dispersal probability between workers of different patrilines and of
different size. This differential allocation may be adaptive because larger workers may be of dispro-
portionate value to dispersing colonies, and their loss an acceptable cost to the colony inheriting the nest.
©2016 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
How individuals allocate resources during reproduction is a
central focus of life history theory. While there can be advantages to
producing larger offspring (Krist, 2011; Marshall &Keough, 2008),
parents with nite resources face a trade-off between offspring size
and number (Smith &Fretwell, 1974), and the vast literature
exploring the inuence of varied quantitative investment on
offspring characteristics indicates that the optimal investment
strategy is context dependent (Bernardo, 1996; Burgess, Bode, &
Marshall, 2013). Less attention has been paid to the importance of
qualitative variation in investment, although some authors have
suggested this may be of at least equal importance (Krist, 2011;
Mousseau &Fox, 1998; Nager, Monaghan, &Houston, 2000). For
example, differential deposition of hormones in avian eggs can
inuence offspring phenotype and behaviour (reviewed in Gil,
2008; Groothuis, Müller, von Engelhardt, Carere, &Eising, 2005),
and indeed, the available evidence suggests that variation in the
composition of invested resources may have diverse effects in a
wide range of species (De Fraipont, Clobert, John, &Meylan, 2000;
Gil, 2008; McGraw, Adkins-Regan, &Parker, 2005).
Theoretical and empirical studies of offspring investment have
focused on solitary organisms (Stearns &Hoekstra, 2005), while
the vast and often ecologically dominant social fauna has suffered a
dearth of attention. Social insect colonies are superorganisms, and
new colonies are created through one of two different modes that
differ strongly in the manner in which resources are allocated.
Under independent colony foundation (ICF) young queens depart
*Correspondence: A. L. Cronin, United Graduate School of Agricultural Sciences,
Iwate University, 3-18-8 Ueda, Morioka 020-8550, Japan.
E-mail address: adamcronin@gmail.com (A. L. Cronin).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
http://dx.doi.org/10.1016/j.anbehav.2016.06.018
0003-3472/©2016 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 119 (2016) 1e9
alone from the parent colony, disperse and mate on the wing, and
start new colonies on their own. Under this strategy, the quantity
and quality (e.g. size, energetic resources) of individual sexual
offspring (dispersing queens and males) produced by the colony
will be the main determinant of colony reproductive success. In
contrast, under dependent colony foundation (DCF, also known as
swarming, budding or colony ssion), young queens start new
colonies with the help of nestmate workers (reviewed in: Cronin,
Molet, Doums, Monnin, &Peeters, 2013; Peeters &Molet, 2010).
These queens leave the parent colony accompanied by numerous
workers and are continuously helped and protected. Colony
reproductive success in this case depends largely on the number
and characteristics of workers (e.g. age, size and genetic lineage)
and other resources (such as brood, food and building material)
allocated to the new colonies. However, despite the abundance and
ecological importance of species employing DCF (Cronin et al.,
2013), and the growing body of theoretical (Bulmer, 1983; Cronin,
Loeuille, &Monnin, 2016; Crozier &Pamilo, 1996; Rangel, Reeve,
&Seeley, 2013) and empirical (Amor et al., 2011; Briese, 1983;
Ch
eron, Cronin et al., 2011; Cronin, F
ed
erici, Doums, &Monnin,
2012; Fern
andez-Escudero, Sepp
a, &Pamilo, 2001; Gotwald,
1995; Lenoir, Qu
erard, Pondicq, &Berton, 1988; Seeley, 1996; van
Veen &Sommeijer, 2000) literature on the quantitative aspects of
reproductive allocation during DCF, studies on qualitative variation
during DCF in social insects have been largely restricted to assess-
ing the potential for nepotism, for which no convincing evidence
has been found (Breed, 2014; Heinze, Elsishans, &H
olldobler, 1997;
Rangel, Mattila, &Seeley, 2009; Solís, Hughes, Klingler, Strassmann,
&Queller, 1998).
In addition to the number of workers, an important consider-
ation under DCF is the characteristics of the workers allocated to
offspring nests. Workers are not identical entities, but can differ in a
variety of ways including genetic makeup, experience, age and
morphology. These factors are not necessarily independent, as for
example some degree of genetic determination of worker
morphology has been demonstrated in several species of ants
(Fraser, Kaufmann, Oldroyd, &Crozier, 2000; Hughes, Sumner, Van
Borm, &Boomsma, 2003; Jaffe, Kronauer, Kraus, Boomsma, &
Moritz, 2007). Trait variation among workers means that they are
not of equivalent tness value for a new colony, and the number of
workers allocated may thus be an incomplete measure of repro-
ductive investment. Furthermore, because social organisms are
subject to selection at multiple levels (Gardner &Grafen, 2009;
H
olldobler &Wilson, 2009), the optimal allocation strategy from
a colony perspective may not be the same as that from an individual
perspective. A range of benets to having a genetically diverse
workforce has been demonstrated in social insects (Baer &Schmid-
Hempel, 1999; Hughes &Boomsma, 2004; Seeley &Tarpy, 2007;
Ugelvig, Kronauer, Schrempf, Heinze, &Cremer, 2010), and we
might thus expect allocation to favour the maintenance of genetic
diversity in newly produced colonies. However, despite the lack of
evidence for nepotism, individual selection could potentially favour
the association of workers with more closely related full-sister
gynes (young queens), leading to reduced genetic diversity and
thus the opposite pattern to that expected by colony level selection.
In this study, we investigated the potential for qualitative bias in
reproductive allocation during DCF in a superorganism, using the
thermophilic ant Cataglyphis cursor. This species reproduces
exclusively via colony ssion (a term we use here synonymously
with DCF, following Cronin et al., 2013), with a parent colony pro-
ducing on average four (range 2e7) offspring colonies containing a
highly variable number of workers (Ch
eron, Cronin et al., 2011;
Cronin et al., 2012; Lenoir et al., 1988). Workers exhibit contin-
uous variation in size (Cagniant, 1983; Cl
emencet &Doums, 2007),
an important trait in thermophilic ants because it is linked with
heat resistance and thus affects foraging behaviour and interspe-
cic scramble competition (Cerd
a, 2001; Cerd
a&Retana, 1997).
Colonies are headed by a single queen, mated with on average
5.6 ±1.3 males (range 4e8; Pearcy, Aron, Doums, &Keller, 2004),
and thus contain multiple worker patrilines (Ch
eron, Monnin,
F
ed
erici, &Doums, 2011; Pearcy et al., 2004). In addition,
although queens produce workers sexually and males by arrheno-
tokous parthenogenesis as in most Hymenoptera, this species is
unusual in that new gynes (i.e. young virgin queens) can be pro-
duced either sexually or via thelytokous parthenogenesis (Doums
et al., 2013; Pearcy et al., 2004).
Our investigation focuses on worker size and genetic patriline.
From both of these perspectives, there are reasons to expect the
pattern of allocation will deviate from our null hypothesis that
workers are randomly distributed. First, worker size in ants is
positively associated with higher longevity (Porter &Tschinkel,
1985), resilience to starvation (Heinze, Foitzik, Fischer, Wanke, &
Kipyatkov, 2003), foraging efciency (e.g. temperature resistance,
Cerd
a&Retana, 1997; Cerd
a&Retana, 2000; Cl
emencet, Cournault,
Odent, &Doums, 2010) and colony defence (H
olldobler &Wilson,
1990). However, as size increases so does production cost, and
only well-established colonies may be able to produce the large
workers they need. Newly founded colonies may thus benet
disproportionately from the presence of larger workers compared
to larger, established colonies. We might therefore expect larger
workers to be preferentially allocated to small colonies and/or
colonies that disperse to new nesting sites (where a new nest must
be excavated), with their corresponding depletion in the parent
colony. Second, from a genetic standpoint, bias could arise if
workers exhibit nepotistic tendencies and preferentially associate
with gynes of their own patriline during ssion for inclusive tness
purposes. However, in C. cursor, this can only occur in colonies with
sexually produced gynes, because parthenogenetically produced
gynes are equally related to all workers regardless of patriline. If
nepotism does occur, we might therefore expect differences in
patriline distribution of workers between sister colonies (i.e.
offspring colonies derived from the ssion of the same parent
colony) containing sexually produced gynes from different patri-
lines. Alternatively, workers of different patrilines in bees and ants
may exhibit different dispersal propensities (Kryger &Moritz, 1997;
Sepp
a, Fern
andez-Escudero, Gyllenstrand, &Pamilo, 2008), and
this could lead to bias in the distribution of patrilines between the
offspring colony that inherited the nest and those that dispersed to
new nesting sites. Finally, size-based allocation of workers may be
confounded by genetic effects if these factors are related. Although
genetic variation has not been convincingly linked to differences in
worker size in two previous studies in C. cursor (Eyer, Freyer, &
Aron, 2013; Fournier, Battaille, Timmermans, &Aron, 2008) this
pattern has been demonstrated in other ant species (Evison &
Hughes, 2011; Fraser et al., 2000; Hughes et al., 2003; Jaffe et al.,
2007; Rheindt, Strehl, &Gadau, 2005) and we reassessed this
possibility here.
METHODS
We used 54 offspring colonies of C. cursor from a previous study
(Ch
eron, Cronin et al., 2011). These colonies were collected as
products of the ssion of 14 parent colonies in the eld at Argel
es-
sur-Mer in southern France. Parentand offspringcolonies refer to
colonies before and after ssion, respectively. That is, each parent
colony divided into several offspring colonies, and offspring col-
onies derived from the same parent colonyare thus sistercolonies.
In C. cursor the queen of the monogynous parent colony (i.e. the
mother queen) is retained in approximately half of all reproducing
colonies (Ch
eron, Cronin et al., 2011) so that one of the resulting
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e92
offspring colonies may contain the original mother queen (who
may also relocate from the original nest). Also note that because
parent colonies cannot be sampled nondestructively we recon-
structed them a posteriori based on their own offspring colonies
(see below). We focused our analyses on two comparisons: (1)
between offspring colonies and reconstructed parent colonies and
(2) between offspring colonies, with particular focus on differences
between dispersing colonies and those that continued to occupy
the original nest.
Morphological Analysis
The morphometric data set comprised 1952 individuals from 54
offspring colonies, with 30e65 randomly sampled workers in each
(excluding colony 544E, which contained only three workers;
Table 1). The characteristics of each colony are given in Table 1 (see
also Ch
eron, Cronin et al., 2011). Note that the genetic analysis (see
below) indicated that the queen of colony 473B was not the mother
queen as reported in Ch
eron, Cronin et al. (2011), but a gyne
belonging to one of the patrilines found in workers. Colony 473B
was therefore queenless.
As a measure of worker size, we used the length of the tibia of
the right hindleg, as previous biometric studies on C. cursor indi-
cated that all classical measures of worker size are highly inter-
correlated (r¼0.76e0.92; Cagniant, 1983; Cl
emencet &Doums,
2007), and tibia length is correlated with worker resistance to
temperature (Cl
emencet et al., 2010). Measurements were per-
formed with NI Vision Assistant 7.0 (National Instruments Corp.,
Austin, TX, U.S.A.) using photographs taken with a Canon Power-
shot S80 camera mounted on a Leica MZ6 stereomicroscope.
Genetic Analysis
For the genetic analysis, we used a subset of 20 offspring col-
onies produced by the ssion of six parent colonies (Table 1) and
comprising a total of 745 workers and 20 gynes/queens. Four of the
parent colonies produced only parthenogenetic gynes (426, 471,
511 and 634) and two produced at least one sexual gyne (present in
offspring colonies 447C and 473B-E). For each of the 20 offspring
colonies, workers used in morphological analyses were genotyped
at six microsatellite loci (Ccur 11, Ccur 46, Ccur 58, Ccur 65, Ccur 99
and Ccur 100) to identify worker patrilines, following Ch
eron,
Monnin et al. (2011). The genotypes of queens and gynes were
obtained from the data set of Ch
eron, Monnin et al. (2011). How-
ever, whereas Ch
eron, Cronin et al. (2011) listed the gynes in
offspring colonies 511A and 511D as sexually produced, reanalysis
of these data indicates that these gynes were in fact parthenoge-
netically produced. In half of the cases (471, 511, 463), one offspring
colony continued to occupy the original nest, whereas in the other
cases (426, 447, 473) all colonies dispersed (Table 1).
To get a more precise estimate of the population allelic fre-
quencies used to estimate relatedness, we added tothe genetic data
set one worker from 17 additional colonies collected the same year
at the same site but which did not ssion (Ch
eron, Cronin et al.,
2011). Data were analysed using Relatedness 5.0.1 (Queller &
Goodnight, 1989) weighting colonies equally.
Worker patrilines, and father genotypes, were determined for
each colony using Matesoft (Moilanen, Sundstr
om, &Pedersen,
2004). The probability of nondetection of a patriline because two
males harboured the same allele combination by chance was low
(0.0003; estimated following Boomsma &Ratnieks, 1996) as our
loci are highly heterozygous (mean He ¼0.75; Cl
emencet &
Doums, 2007). For parent colonies in which the mother queen
was missing (426, 447, 473, 634), her genotype was deduced from
the parthenogenetically produced gynes and/or from worker ge-
notypes using the broadoption in Matesoft. In all cases, a single
mother queen genotype was possible (but see Results regarding a
possible queen turnover in the parent colony 511).
Statistical Analysis
We rst examined how the characteristics of workers in
offspring colonies compared to the parent colony. For this purpose,
it was necessary to reconstruct parent colonies because they could
not be nondestructively sampled prior to ssion and, as we only
measured a subset of all available workers in offspring colonies,
simply compiling data would not be representative of the parent
colony as offspring colonies were of different size. We thus recon-
structed a distribution of 1000 possible parent colonies by sampling
from the measured and genotyped worker subsets as follows: all
workers from the largest offspring colony were assigned to the
reconstructed parent colony: For each remaining offspring colony,
Table 1
Colony metrics and collection data
Parent colony Type of analyses Offspring colonies Total
AB C D E FGH
426 M/G e62 (35) 153 (35-1F) 79 (35) eeee294 (105)
447 M/G e148 (35-1F) 112 (35)
S
30 (30-2F) eeee290 (100)
456 M 1150 (35) 152 (35) 326 (35) ee eee1628 (105)
464 M 72 (35) 180 (35) eee eee252 (70)
471 M/G 492 (30)
Q
137 (30) 160 (30) ee eee789 (90)
473 M/G e78 (35)
S
56 (35-3F)
S
42 (35)
S
102 (35-1F)
S
eee278 (140)
511 M/G 394 (65) 890 (65)
Q
e45 (45-2F) eeee1329 (175)
544 M 243 (30)
Q
305 (48) 276 (48) 213 (48) 3 (0) eee1037 (174)
571 M 81 (35) 152 (35) e112 (35) 90 (35) 60 (35) ee495 (175)
581 M 85 (35) 158 (35)
Q
e48 (18) eeee291 (88)
598 M 234 (35) 177 (35) 92 (35) 56 (35) 69 (35) 90 (35)
Q
e62 (35) 780 (245)
634 M/G 293 (35) 69 (35) 77 (35) 138 (30) eeee577 (135)
641 M 257 (35) 58 (35) e252 (35) 121 (35) e103 (35) e791 (175)
667 M 49 (35) 1284 (35)
Q
160 (35) 71 (35) e244 (35) ee1808 (175)
Number of workers collected and analysed in 54 offspring colonies studied by Ch
eron, Cronin et al. (2011). Offspring colonies are named with the label of their parent colony
followed by a letter from A to H, with colonies designated A(e.g. 456A) having inherited the nest of the parent colony and others (BeH) having dispersed to new nests. There
was thus no colony A when all offspring colonies dispersed. The number of workers analysed from each offspring colony is given between parentheses. M and G indicate,
respectively, whether morphological and genetic analyses were carried out, with the same workers used for both analyses. Foreign workers are indicated by values subtracted
from the total sampled worker population and denoted Fand were not included in analyses. Each offspring colony contained either the mother queen (indicated by the
superscript Q) or a gyne. Cases in which offspring colonies contained sexually produced gynes are marked with a superscript S whereas unmarked nests contained
parthenogenetically produced gynes.
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e93
we then sampled individuals randomly without replacement, up to
a number dened by that colony's proportional contribution to the
parental colony relative to that of the largest offspring colony. For
instance, for a parental colony of 500 workers that ssioned into
two colonies of 200 and 300 workers and had 35 workers measured
from each, the 35 measured workers of the largest colony (which
comprised 60% of the parental colony) would be included in the
reconstructed colony while a proportional number (40%) would be
randomly sampled from the remaining colony (i.e. 23 workers for a
total sample of 58 workers). The composition of reconstructed
parental colonies was thus determined by the relative size of each
offspring colony even though our sample sizes for genetic and
morphological study were more or less consistent across nests and
limited by the size of the largest offspring colony.
We estimated the distribution of patrilines and worker size in
the parent colonies by generating for each metric a set of 1000
reconstructed colonies. We then considered whether observed
values from offspring colonies lay within the 95% condence in-
tervals of the distribution of their reconstructed parent colonies.
We used the same procedure for estimating the distribution of
relatedness in reconstructed parent colonies, but because of limi-
tations to processing time for relatedness calculations, our distri-
bution comprised estimates from only 100 reconstructed colonies.
Mean values for each metric analysed were also determined for
each reconstructed parent colony and compared to the mean of
their offspring colonies using paired Wilcoxon signed-rank tests.
For our morphometric analysis, we compared both mean worker
size and diversity of worker size between offspring colonies and
their reconstructed parent colony. The diversity of worker size was
estimated using the unbiased coefcient of variation (cv) for small
sample sizes (Sokal &Rohlf, 1995):
cv ¼1þ1
4nsd 100
m
Where nis the sample size (number of workers measured), and
mand sd are, respectively, the mean and standard deviation of the
tibia length. For the genetic analyses, we compared (1) the number
of patrilines, (2) the relatedness estimate and (3) the skew in pat-
riline distribution using the Nonacs Bindex (Nonacs, 2003). This
index ranges from 1 to 1, with more positive values indicating
higher bias than expected by chance (a value of 0) and more
negative values indicating a more even distribution than expected
by chance.
We next focused on the distribution of worker characteristics
among sister colonies, i.e. offspring colonies derived from the same
parent colony. To explore the inuence of inheritance of the nest of
the parent colony and presence of the mother queenwe performed
mixed-model analyses using the nlme package (Pinheiro &Bates,
2000) in R version 3.2.1 (R Core Team, 2014) with worker size as
the response variable, inheritance of the parent nest and presence/
absence of the mother queen as xed factors, and offspring colonies
within parent colonies as nested random factors. The mother queen
was found in the offspring colony that inherited the parent nest in
only two of the 14 cases (colonies 471A and 544A, Table 1). We
could thus test the two effects in the same model but could not test
for an interaction between the two factors as there was no case
where the mother queen survived but the parent nest was not
inherited. We also account for differences in worker size variation
by using the weight function to allow for heterogeneous variance
between offspring colonies that inherited the parent nest and those
that dispersed. We report results from log-likelihood ratio tests
between the full model and the model without the factor of in-
terest. For testing xed effects, models were tted with maximum
likelihood (Pinheiro &Bates, 2000). The heterogeneous variance
component was only retained in the full model if signicant.
To compare patriline distributions between offspring colonies,
we performed a Fisher exact test with the expectation that no bias
should be found if worker allocation is random. Comparisons be-
tween new offspring colonies from the same parent colony were
corrected for repeated tests following Benjamini and Hochberg
(1995). We assessed the potential for nepotism in the form of a
worker preference for gynes of the same patriline in the ve col-
onies in which gynes had been sexually produced (447C and 473B-
E; Table 1). We compared the proportion of workers of the focal
gynes' patriline across all sister colonies using chi-square tests.
To explore possible links between patriline and morphology, we
combined data for morphometric and genetic samples where
available (N¼20 offspring colonies from six parent colonies) and
analysed these with a mixed-effects model, with worker size as the
response variable, patriline nested within parent colony and parent
colony as random factors, and the number of workers of each
patriline as a xed covariate. We considered patriline as a random
factor in analyses as our aim was to estimate the part of variation
explained by patriline over the six colonies (as used for estimating
heritability in animal models; Wilson et al., 2010), rather than to
test whether specic patrilines within given colonies differed in
worker size. To do so we compared models tted with a restricted
maximum likelihood with and without the effect of patriline nested
within colony using a log likelihood ratio test (Pinheiro &Bates,
2000). The percentage of worker size variation explained by pat-
riline was calculated from the random parameters of the mixed-
effect model, and xed effects were tested as previously described.
Finally, in the cases for which worker patriline distribution
differed between sister colonies, the effects of worker size and
patriline could be intertwined if patrilines that are more prone to
leave the nests are also those with large workers. To test these
factors simultaneously, we used the three parent colonies in which
the parent nest was inherited by an offspring colony and tested
whether patriline and worker size affected the propensity of
workers to stay in the parent nest or leave to join a dispersing
colony using a binomial glm (binomial function and logit link), with
staying or leaving as the response variable. In this case, patriline
was considered as a xed factor, as we performed a separate
analysis for each colony and there were too few patrilines to
properly estimate the variance component. Factors were tested by
comparing models with and without the factor of interest using
chi-square tests (Crawley, 2007).
Ethical Note
Ants were observed without direct interference during the
monitoring part of our study (see Ch
eron, Cronin et al., 2011). All
specimens collected for morphological and genetic analysis were
placed directly into 90% alcohol with 10% Tris-EDTA in accordance
with standard sampling procedures.
RESULTS
Inuence of Fission on Worker Size Distribution
We rst compared worker size and coefcient of variation (cv)
between offspring colonies and their reconstructed parent colony.
The mean size of workers in offspring colonies differed from that
expected if workers had been randomly allocated in 74% (40 of 54)
of cases (i.e. was outside the 95% condence intervals (CIs) of the
reconstructed parent colonies; Fig. 1a). There was no signicant
difference in worker size between offspring colonies and the mean
value of reconstructed parent colonies overall (Wilcoxon paired
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e94
test: z¼1.54, N¼14, P¼0.124). None the less, workers in
offspring colonies that inherited the parent nest were on average
smaller than in the reconstructed parent colony (z¼2.22, N¼11,
P¼0.026) whereas workers in offspring colonies that dispersed to
new nesting sites tended to be larger than those in the recon-
structed parent colony (z¼1.96, N¼14, P¼0.050; Fig. 1a).
Similarly, the cv in worker size differed from the distribution of
reconstructed parent colonies in 70% (38 of 54) of offspring colonies
(Fig. 1b). While the pattern among offspring colonies was again
nonsignicant (Wilcoxon paired test: z¼1.66, N¼14, P¼0.096),
offspring colonies that inherited the parent nest tended to have a
higher cv (z¼1.78, N¼11, P¼0.075) while those that dispersed
had a lower cv than that of the reconstructed parent colony
(z¼2.23, N¼14 , P¼0.026).
We then focused on differences between sister colonies. There
was no heterogeneity of variance in worker size linked to the
presence of the mother queen (likelihood ratio ¼0.038, df ¼1,
P¼0.85) whereas there was signicant heterogeneity of variance
linked to the inheritance of the parent nest (likelihood ratio ¼17. 57,
df ¼1, P<0.0001), with variance being 1.4 times higher in offspring
colonies that dispersed than in those that inherited the parent nest
(Fig. 1b). The presence of the mother queen had no effect on mean
worker size (likelihood ratio ¼1.88, df ¼1, P¼0.17). However,
workers in the offspring colony that inherited the parent nest were
smaller on average than those in offspring colonies that dispersed
(likelihood ratio ¼16.15, df ¼1, P<0.0001; Fig. 1a).
Inuence of Fission on Genetic Characteristics
Our genetic analysis indicated that six offspring colonies con-
tained a total of 10 foreign workers (Table 1). Following removal of
these workers from the genetic data set, we obtained a total of 735
genotyped workers from 20 offspring colonies produced by six
parent colonies. We identied between three and 12 patrilines per
parent colony. Two of these patrilines contained a single individual
each, but in both cases the putative father differed by at least two
loci from the other putative fathers of the colony so that they were
considered real patrilines and not results of PCR error.
The number of patrilines in reconstructed parent colonies
(6.7 ±3.3, range 3e12; based on median values of reconstructed
colonies) was in most cases no different to that measured in
offspring colonies (mean ¼6.2 ±2.5, range 3e9; Fig. 2a) and there
was no difference in mean values between parent and offspring
colonies overall (Wilcoxon paired test: z¼1.60, N¼6, P¼0.109).
The frequency of workers of different patrilines was highly skewed
in the 20 offspring colonies, as well as in their reconstructed parent
colonies (Nonacs skew index B>0; P<0.001 in all cases; Fig. 2b).
Skew in 40% (eight of 20) of offspring colonies differed signicantly
from that of their reconstructed parent colony (Fig. 2b), and was
higher in six of these cases, representing a signicant pattern of
increased skew overall following colony ssion (Wilcoxon paired
test: z¼2.20, N¼6, P¼0.028). Relatedness in reconstructed
parent colonies ranged from 0.279 to 0.575 (median values; Fig. 2c)
and in offspring colonies from 0.268 to 0.592. Relatedness in the
latter differed from the distribution of the 100 reconstructed parent
colonies in 40% (eight of 20) of cases, and was in all instances higher
than expected if workers were randomly allocated, representing a
pattern of increased relatedness following ssion overall (Wilcoxon
paired test: z¼2.20, N¼6, P¼0.028; Fig. 2c). In contrast to our
results for morphology there was no clear pattern of differences
between colonies that inherited the original nest and those that
dispersed, although differences between parent and offspring col-
onies were more common in cases in which one colony continued
to occupy the original nest (30%, 70% and 60% for patrilines, relat-
edness and skew respectively; N¼10) than those in which all
colonies dispersed (0%, 10% and 20% respectively; N¼10).
The distribution of worker patrilines differed signicantly (i.e.
was not random) between sister colonies in only two of the six
cases (Fig. 3). For these two cases, pairwise tests revealed signi-
cant differences between the colonies that inherited the parent
nest and three dispersing colonies (following adjustments for
multiple tests; Benjamini &Hochberg, 1995;Fig. 3) as well as be-
tween dispersing colonies 511B and 511D. In the latter case the
difference resulted from the absence of four patrilines in colony
511B relative to 511D, and a close look at the genetic data strongly
suggests the occurrence of a queen turnover in the parent colony in
the previous season (as the current mother queen had already
produced workers) or the fusion of two colonies headed by two
clonal queens. Indeed, for four patrilines, only one of the two alleles
of the current mother queen was found in the workers, suggesting
the mother of these patrilines was homozygous at these loci, unlike
the current mother queen. Interestingly, these four patrilines were
absent from colony 511B where the current mother queen was
found.
Evidence for Nepotism
We investigated the propensity for workers to associate with a
gyne of their own patriline in the ve offspring colonies with
sexually produced gynes (447C, 473B-E; Table 1). In each case the
proportion of workers of the same patriline as the gyne was no
different in the colony where the gyne was located than in sister
1.8
1.9
2
2.1
2.2
2.3
2.4
Mean tibia length (mm)
(a)
426
447
456
464
471
473
511
544
571
581
598
634
641
667
Colon
y
6
8
10
12
14
Mean cv
(b)
426
447
456
464
471
473
511
544
571
581
598
634
641
667
Figure 1. (a) Mean worker tibia length and (b) coefcient of variation in worker tibia
length in offspring colonies that inherited the parent nest (triangles) and offspring
colonies that dispersed (circles) compared to reconstructed parent colonies (dark-grey
bars). Bars represent 95% CIs of 1000 reconstructed colonies. Black markers lie outside
95% CIs whereas white markers lie within 95% CIs. Note that in some cases, all colonies
dispersed and no colony inherited the parent nest.
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e95
colonies (chi-square tests: P>0.28 in all cases), suggesting workers
are not disproportionately associated with gynes of the same pat-
riline during ssion.
Effect of Patriline on Worker Size
In the 20 offspring colonies for which we had both genetic and
morphological data, the number of workers of each patriline was
not associated with worker size (likelihood ratio ¼2.28, df ¼1,
P¼0.13) and this factor was thus removed from the model.
Workers of different patrilines varied in size (likelihood
ratio ¼21.13, df ¼1, P<0.001), with patriline explaining 7.7% of the
variation in worker size. None the less, in the cases for which we
detected differences in worker patriline distribution between sister
colonies, the effects of worker size and patriline could be inter-
twined if patrilines that are more prone to leave the nests are also
those with large workers. For the three cases in which an offspring
colony inherited the nest, we tested whether patriline and worker
size affected the propensity of workers to stay in the inheriting
colony or leave with a dispersing colony using a binomial model
with staying or leaving as the response variable. In the two cases
(colonies 511 and 634) with a signicant difference in patriline
distribution, worker propensity to leave was signicantly associ-
ated with both patriline (
c
2
8
¼12.35, P<0.0001 for colony 511 and
c
2
2
¼20.58, P<0.0001 for colony 634) and worker size (
c
2
1
¼11.99,
P<0.001 for colony 511 and
c
2
1
¼7.38, P¼0.007 for colony 634).
On the other hand, in colony 471 for which we previously did not
detect signicant variation in patriline distribution between sister
colonies, there was no effect of patriline (
c
2
8
¼12.35, P¼0.14) but a
highly signicant effect of worker size on the propensity to leave
(
c
2
1
¼12.95, P<0.001). These results indicate that both patriline
and worker size can independently inuence the propensity to
leave the parent nest.
DISCUSSION
Workers were not allocated randomly during the process of
colony ssion in the ant C. cursor, but according to worker size, and
in some cases patriline. Specically, larger workers were more
likely to be allocated to offspring colonies that dispersed than
remain in the offspring colony that inherited the parent nest. We
also showed by comparing offspring colonies and reconstructed
parent colonies that in many cases ssion led to a reduction in the
diversity of worker size and genetic diversity in offspring colonies.
Finally, workers of different patrilines differed in size, in contrast to
two previous studies on this species (Eyer et al., 2013; Fournier
et al., 2008) but in support of studies on other ants (Fraser et al.,
2000; Hughes et al., 2003; Jaffe et al., 2007). Our ndings there-
fore indicate that allocation of workers is not only quantitatively
biased (Ch
eron, Cronin et al., 2011), but also qualitatively biased,
and this is linked with nest inheritance and dispersal. These nd-
ings align with data on ssion in honeybees, Apis mellifera, that
show that the departing swarm containing the mother queen has a
quantitative advantage in incorporating the majority of the work-
force, while a qualitative advantage extends to the nonswarming
daughter colony in the form of the inherited material resources
represented by the nest comb, its brood and provisions (Rangel &
Seeley, 2012).
During colony ssion in C. cursor, resources (adults and brood)
are carried individually from the parent colony to offspring colonies
by a proportion of workers acting as scouts (Cagniant, 1976; Lenoir
et al., 1988). As in other social insects employing DCF, these scouts
also act as decision makers in determining to which offspring col-
ony resources will be allocated (Seeley &Morse, 1977). Thus,
whereas the passive majority of individuals may have some control
over whether they stay in the colony that inherits the parent nest or
depart to join a dispersing colony (for example by placing them-
selves at the nest entrance), they have little choice over the desti-
nation. Bias between the offspring colony inheriting the nest and
dispersing colonies can thus arise in two ways: (1) differential
dispersal propensities between workers, with some ants actively
(or incidentally) adopting behaviours that increase their chances of
being carried to a dispersing colony and (2) active discrimination by
transporting scouts, in selecting workers to be carried to dispersing
colonies on the basis of particular characteristics. These modes of
generating bias may act in concert or independently, although we
Colon
y
426
447
471
473
511
634
Relatedness
0.3
0.4
0.5
0.6 (c)
B index
0.1
0.2
(b)
426
447
471
473
511
634
(a)
Patrilines
5
10
426
447
471
473
511
634
Figure 2. (a) Number of patrilines, (b) mean skew (Bindex) in patriline distribution
and (c) relatedness values for offspring colonies that inherited the parent nest (tri-
angles) and offspring colonies that dispersed (circles) compared to distributions for
reconstructed parent colonies (dark-grey bars). Bars represent 95% CIs of 1000
reconstructed colonies (or 100 reconstructed colonies for relatedness). Black markers
lie outside 95% CIs whereas white markers lie within 95% CIs. Note that in some cases,
all colonies dispersed and no colony inherited the parent nest.
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e96
argue here that interindividual differences in dispersal propensity
provide the most parsimonious explanation for the majority of the
patterns observed for reasons explained below.
Much of the bias in allocation among sister colonies was related
to worker size, with larger workers more frequently found in
offspring colonies that had dispersed. Larger workers are likely to
be more highly valued because they are more costly to produce
(H
olldobler &Wilson, 1990; Oster &Wilson, 1978), and because of
the benets of large size (Cerd
a, 2001; Heinze et al., 2003; Porter &
Tschinkel, 1985). In Cataglyphis ants, larger workers are able to
withstand higher temperatures (Cl
emencet et al., 2010) and have
higher immune levels (Bocher, Tirard, &Doums, 2007), which may
provide important foraging advantages (Cerd
a, 2001; Cerd
a&
Retana, 1997). These advantages may be particularly important to
dispersing colonies, which could gain a two-fold benet: starting
with sufcient large workers to constitute an accomplished work-
force for nest excavation and foraging (which may be crucial for the
rapid development of dispersing colonies and for scramble com-
petitors as in C. cursor) and circumventing the cost of the produc-
tion of large workers (and thereby gaining a head start towards
colony development). At the same time, the loss of predominantly
large workers may represent an acceptable cost to the offspring
colony that inherits the parent nest, both because of the dispro-
portionate advantage conferred to dispersing sister colonies and
because the inheriting colony does not have to excavate a nest and
is hence better positioned to manufacture new large workers to
replace those lost. Alternatively, colonies may simply have no
control over which workers are allocated to which offspring colony.
Worker size can provide the basis for a distribution bias if trans-
porters preferentially select larger ants and/or larger ants show a
higher dispersal propensity. In the latter case this could simply arise
from the fact that larger ants are more prone to forage, defend and
maintain the nest and therefore more likely to be found in the
chambers near the entrance. Both of these scenarios can explain the
allocation bias in large workers to dispersing colonies relative to
inheriting colonies, although dispersal propensity has the advan-
tage of not requiring us to invoke active discrimination abilities in
transporting workers and avoids additional complications such as
delays generated by a selection process each time transporters re-
turn to the source nest, and a potential trade-off between nding an
optimal target and search time.
In two of six cases, patrilines were not randomly distributed
among sister colonies and, as for the distribution of worker size,
this pattern could arise if workers of different patrilines exhibit
different dispersal propensities. In line with previous studies on
social insects (Heinze et al., 1997; Rangel et al., 2009; Solís et al.,
1998), we found no evidence of nepotism during ssion in the
form of a greater bias in colonies containing sexually produced
gynes, and indeed, the two colonies in which signicant differences
in the distribution of patrilines were detected contained only
parthenogenetically produced gynes. Furthermore, in the two cases
in which the parent colony produced gynes sexually, we found no
evidence of workers being overrepresented in offspring colonies
containing gynes of their own patriline, or higher bias in sister
426 (P)
511 (P)
0.6
0.4
0.2
1
0
0.8
0.6
0.4
0.2
1
0
0.8
R BCD R BCD
471 (P)
RABC
473 (S)
RBCDE
634 (P)
RABCD
RAB D
Proportion of the workforce
447 (SP)
*** ***
***
***
Figure 3. Stacked bar graphs showing proportions of different patrilines in the reconstructed parent colony (R) and in offspring colonies (AeE). Offspring colonies Ainherited the
parent nest while offspring colonies BeEdispersed. Differently patterned bars represent different patrilines. Asterisks indicate signicance levels (corrected for multiple tests) of
pairwise Fisher exact tests on patriline distributions among sister colonies (i.e. excluding reconstructed parent colonies): **P<0.01; ***P<0.001. Types of gynes produced are
indicated following colony number (S ¼sexually produced, P ¼parthenogenetically produced, SP ¼both forms present). Data for reconstructed parent colonies (R) were taken as
the median number of workers of each patriline over the 1000 simulated colonies (see Methods).
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e97
colonies than in colonies containing only parthenogenetically
produced gynes. Patriline bias may arise from one additional source
in C. cursor regardless of the type of gyne present, although at
present it is difcult to estimate its importance. Workers in C. cursor
can produce new gynes via thelytoky in the event of colony
orphaning, and abundant patrilines are more likely to mother
replacement gynes (e.g. Ch
eron, Monnin et al., 2011). Workers of
the same patrilines could thus benet from initiating new colonies
together if the probability of queen replacement was sufciently
high, although at present the frequency of queen replacement in
the eld is unknown. We did nd that patrilines were not randomly
distributed among sister colonies in one case in which all gynes
were parthenogenetically produced (colonies 511A, B and D). In this
case, workers of four patrilines that appear to have been mothered
by a previous mother queen were absent from the colony con-
taining the current mother queen. This may indicate allocation of
individuals along kin lines, although genetic evidence also sug-
gested a mother queen turnover had occurred in the paternal col-
ony, and it is difcult to anticipate what inuence this may have
had on the ssion process. Thus, while differing dispersal pro-
pensities between patrilines are sufcient to explain the majority of
the genetic patterns observed during colony ssion, we cannot rule
out a possible occasional role of nepotism in the case of queen
turnover.
Our combined results thus indicate that although we cannot
rule out a role of active discrimination by transporters and occa-
sional nepotistic behaviour, the majority of the allocation bias
observed among offspring colonies can be explained by differing
dispersal propensities between workers based on size and, in some
cases, patriline. We stress, however, that our genetic data comprise
only six mother colonies, reconstructed from the 20 offspring col-
onies they ssioned into, and should be treated accordingly with
some caution. Previous studies have suggested that patrilineal
differences in dispersal propensity can explain distribution pat-
terns during ssion in honeybees (Estoup, Solignac, &Cornuet,
1994; Getz, Brückner, &Parisian, 1982; Kryger &Moritz, 1997),
and the one study that demonstrated evidence of assortative dis-
tribution of workers along kin lines during ssion in ants (Sepp
a
et al., 2008) could not rule out this possibility. Workers of
different patrilines are known to vary in their tendency to perform
various tasks in C. cursor (Eyer et al., 2013) and it seems reasonable
to assume that this also applies to dispersal. However, no previous
study, to our knowledge has provided evidence of size-based
variation in dispersal propensity during ssion. Although size var-
ied with patriline in this study, patriline explained only 8% of the
variation in worker size and this is insufcient to account for the
strong morphometric patterns observed among sister colonies.
Our ndings have implications for our understanding of the
evolution and ecology of reproductive and mating strategies. By
comparing the reconstructed parent colony and offspring colonies,
we showed that ssion can decrease both morphological and ge-
netic diversity in offspring colonies. This may represent an inherent
cost of ssion given the benets of a diverse workforce in social
insects (e.g. Baer &Schmid-Hempel, 1999; Cerd
a&Retana, 1997;
Hughes &Boomsma, 2004; Reber, Castella, Christe, &Chapuisat,
2008; Ugelvig et al., 2010). The reduction in worker genetic di-
versity resembles that associated with a population genetic
bottleneck (Nei, Maruyama, &Chakraborty, 1975). However, it is
likely to be short-lived relative to a genetic bottleneck because (1)
the genetic diversity of the breeding population (i.e. queens) is
unaffected by ssion, and thus diversity will be rapidly reinjected
by the production of new brood through multiple mating and (2)
brood items transported to the new site during ssion are pre-
sumably exempt from bias, and will thus help restore diversity on
maturity. Furthermore, qualitative investment bias may act to
erode or augment any quantitative bias in investment, and both
these components may need to be incorporated into investment
strategy models (e.g. Crozier &Pamilo, 1996; Pamilo, 1991)to
properly understand reproductive investment under DCF. We close
with one caveat to the above discussion in that age is likely to be an
important factor in determining worker value, and could poten-
tially inuence worker allocation patterns and dispersal propensity
during ssion. However, it is at present not possible to reliably age
ants and this question must remain open. None the less, because
ants do not grow after eclosion as adults (H
olldobler &Wilson,
1990) any effect of age is unlikely to inuence the size-linked ef-
fects we demonstrate.
Acknowledgments
This work was supported by grant ANR-06-BLAN-0268 to C.D.,
P.F. and T.M., and grant ANR-11-IDEX-004-02 from Sorbonne Uni-
versities to C.D. and T.M. We thank Elphi Hamdi and Darja
Dubravcic for preliminary results, and Rapha
el Boulay and several
anonymous referees for helpful comments on the manuscript.
References
Amor, F., Ortega, P., Jowers, M. J., Cerd
a, X., Billen, J., Lenoir, A., et al. (2011). The
evolution of workerequeen polymorphism in Cataglyphis ants: Interplay be-
tween individual and colony-level selections. Behavioral Ecology and Sociobio-
logy, 65,1473e1482.
Baer, B., & Schmid-Hempel, P. (1999). Experimental variation in polyandry affects
parasite load and tness in a bumble bee. Nature, 397,151e154 .
Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: A practical
and powerful approach to multiple testing. Journal of the Royal Statistical So-
ciety: Series B, 57, 289e300.
Bernardo, J. (1996). The particular maternal effect of propagule size, especially egg
size: Patterns, models, quality of evidence and interpretations. American Zool-
ogist, 36,216e236.
Bocher, A., Tirard, C., & Doums, C. (2007). Phenotypic plasticity of immune defence
linked with foraging activity in the ant Cataglyphis velox.Journal of Evolutionary
Biology, 20, 2228e2234.
Boomsma, J., & Ratnieks, F. L. W. (1996). Paternity in social hymenoptera. Philo-
sophical Transactions of the Royal Society B: Biological Sciences, 351,947e975.
Breed, M. D. (2014). Kin and nestmate recognition: The inuence of WD Hamilton
on 50 years of research. Animal Behaviour, 92,271e279.
Briese, D. T. (1983). Different modes of reproductive behaviour (including a
description of colony ssion) in a species of Chelaner (Hymenoptera: For-
micidae). Insectes Sociaux, 30, 308e316.
Bulmer, M. G. (1983). Sex ratio theory in social insects with swarming. Journal of
Theoretical Biology, 100, 329e339.
Burgess, S. C., Bode, M., & Marshall, D. J. (2013). Costs of dispersal alter optimal
offspring size in patchy habitats: Combining theory and data for a marine
invertebrate. Functional Ecology, 27,757e765.
Cagniant, H. (1976). Cycle biologique de la fourmi Cataglyphis cursor Fonscolombe.
Hym
enopt
eres, Formicidae. Vie et Milieu, 26,277e281.
Cagniant, H. (1983). La parth
enogen
ese th
elytoque et arrh
enotoque des ouvri
eres
de la fourmi Cataglyphis cursor Fonscolombe (Hym
enopt
eres Formicidae). Etude
biom
etrique des ouvri
eres et de leurs potentialit
es reproductrices. Insectes
Sociaux, 30,241e254.
Cerd
a, X. (2001). Behavioural and physiological traits to thermal stress tolerance in
two Spanish desert ants. Etologia, 9,15e27.
Cerd
a, X., & Retana, J. (1997). Links between worker polymorphism and thermal
biology in a thermophilic ant species. Oikos, 78, 467e474.
Cerd
a, X., & Retana, J. (2000). Alternative strategies by thermophilic ants to cope
with extreme heat: Individual versus colony level traits. Oikos, 89,155e163.
Ch
eron, B., Cronin, A. L., Doums, C., F
ed
erici, P., Haussy, C., Tirard, C., et al. (2011).
Unequal resource allocation among colonies produced by ssion in the ant
Cataglyphis cursor.Ecology, 92, 1448e1458.
Ch
eron, B., Monnin, T., F
ed
erici, P., & Doums, C. (2011). Variation in patriline
reproductive success during queen production in orphaned colonies of the
thelytokous ant Cataglyphis cursor.Molecular Ecology, 20,2011e2022.
Cl
emencet, J., Cournault, L., Odent, A., & Doums, C. (2010). Worker thermal tolerance
in the thermophilic ant Cataglyphis cursor (Hymenoptera, Formicidae). Insectes
Sociaux, 57,11e15.
Cl
emencet, J., & Doums, C. (2007). Habitat-related microgeographic variation of
worker size and colony size in the ant Cataglyphis cursor.Oecologia, 152,
211e218.
Crawley, M. J. (2007). The R book. Chichester, U.K.: Wiley.
Cronin, A. L., F
ed
erici, P., Doums, C., & Monnin, T. (2012). The inuence of intra-
specic competition on resource allocation during dependent colony founda-
tion in a social insect. Oecologia,361e369.
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e98
Cronin, A. L., Loeuille, N., & Monnin, T. (2016). Strategies of offspring investment and
dispersal in a spatially structured environment: A theoretical study using ants.
BMC Ecology, 16,4.
Cronin, A. L., Molet, M., Doums, C., Monnin, T., & Peeters, C. (2013). Recurrent
evolution of dependent colony foundation across eusocial insects. Annual Re-
view of Entomology, 58,37e55.
Crozier, R. H., & Pamilo, P. (1996). Evolution of social insect colonies: Sex allocation
and kin selection. Oxford, U.K.: Oxford University Press.
De Fraipont, M., Clobert, J., John, H., & Meylan, S. (2000). Increased pre-natal
maternal corticosterone promotes philopatry of offspring in common lizards
Lacerta vivipara.Journal of Animal Ecology, 69,404e413.
Doums, C., Cronin, A. L., Ruel, C., F
ed
erici, P., Haussy, C., Tirard, C., et al. (2013).
Facultative use of thelytokous parthenogenesis for queen production in the
polyandrous ant Cataglyphis cursor.Journal of Evolutionary Biology, 26,
1431 e1444 .
Estoup, A., Solignac, M., & Cornuet, J.-M. (1994). Precise assessment of the number
of patrilines and of genetic relatedness in honeybee colonies. Proceedings of the
Royal Society B: Biological Sciences, 258,1e7.
Evison, S. E., & Hughes, W. O. (2011). Genetic caste polymorphism and the evolution
of polyandry in Atta leaf-cutting ants. Naturwissenschaften, 98, 643e649.
Eyer, P.-A., Freyer, J., & Aron, S. (2013). Genetic polyethism in the polyandrous desert
ant Cataglyphis cursor.Behavioral Ecology, 24,144e151.
Fern
andez-Escudero, I., Sepp
a, P., & Pamilo, P. (2001). Dependent colony founding in
the ant Proformica longiseta.Insectes Sociaux, 48,80e82.
Fournier, D., Battaille, G., Timmermans, I., & Aron, S. (2008). Genetic diversity,
worker size polymorphism and division of labour in the polyandrous ant Cat-
aglyphis cursor.Animal Behaviour, 75, 151e15 8.
Fraser, V. S., Kaufmann, B., Oldroyd, B. P., & Crozier, R. H. (2000). Genetic inuence
on caste in the ant Camponotus consobrinus.Behavioral Ecology and Sociobiology,
47,188e194.
Gardner, A., & Grafen, A. (2009). Capturing the superorganism: A formal theory of
group adaptation. Journal of Evolutionary Biology, 22, 659e671.
Getz, W. M., Brückner, D., & Parisian, T. R. (1982). Kin structure and the swarming
behavior of the honey bee Apis mellifera.Behavioral Ecology and Sociobiology, 10,
265e270.
Gil, D. (2008). Hormones in avian eggs: Physiology, ecology and behavior. Advances
in the Study of Behavior, 38,337e398.
Gotwald, W. H. J. (1995). Army ants: The biology of social predation. Ithaca, NY:
Cornell University Press.
Groothuis, T. G., Müller, W., von Engelhardt, N., Carere, C., & Eising, C. (2005).
Maternal hormones as a tool to adjust offspring phenotype in avian species.
Neuroscience &Biobehavioral Reviews, 29, 329e352.
Heinze, J., Elsishans, C., & H
olldobler, B. (1997). No evidence for kin assortment
during colony propagation in a polygynous ant. Naturwissenschaften, 84,
249e250.
Heinze, J., Foitzik, S., Fischer, B., Wanke, T., & Kipyatkov, V. E. (2003). The signi-
cance of latitudinal variation in body size in a holarctic ant, Leptothorax acer-
vorum.Ecography, 26, 349e355.
H
olldobler, B., & Wilson, E. O. (1990). The ants. Berlin, Germany: Springer Verlag.
H
olldobler, B., & Wilson, E. O. (2009). The super-organism. The beauty, elegance and
strangeness of insect societies. London, U.K.: W. W. Norton.
Hughes, W. O., & Boomsma, J. J. (2004). Genetic diversity and disease resistance in
leaf-cutting ant societies. Evolution, 58, 1251e1260.
Hughes, W. O. H., Sumner, S., Van Borm, S., & Boomsma, J. J. (2003). Worker caste
polymorphism has a genetic basis in Acromyrmex leaf-cutting ants. Proceedings
of the National Academy of Science, USA, 100, 9394e9397.
Jaffe, R., Kronauer, D. J. C., Kraus, F. B., Boomsma, J. J., & Moritz, R. F. A. (2007).
Worker caste determination in the army ant Eciton burchellii.Biology Letters, 3,
513e516.
Krist, M. (2011). Egg size and offspring quality: A meta-analysis in birds. Biological
Reviews, 86, 692e716.
Kryger, P., & Moritz, R. F. A. (1997). Lack of kin recognition in swarming honeybees
(Apis mellifera). Behavioral Ecology and Sociobiology, 40, 271e276.
Lenoir, A., Qu
erard, L., Pondicq, N., & Berton, F. (1988). Reproduction and dispersal in
the ant Cataglyphis cursor (Hymenoptera, Formicidae). Psyche, 95,21e44.
Marshall, D. J., & Keough, M. J. (2008). The evolutionary ecology of offspring size in
marine invertebrates. Advances in Marine Biology, 53,1e60.
McGraw, K., Adkins-Regan, E., & Parker, R. (2005). Maternally derived carotenoid
pigments affect offspring survival, sex ratio, and sexual attractiveness in a
colorful songbird. Naturwissenschaften, 92,375e380.
Moilanen, A., Sundstr
om, L., & Pedersen, J. S. (2004). Matesoft: A program for
deducing parental genotypes and estimating mating system statistics in hap-
lodiploid species. Molecular Ecology Notes, 4, 795e797.
Mousseau, T. A., & Fox, C. W. (1998). The adaptive signicance of maternal effects.
Trends in Ecology &Evolution, 13, 403e407.
Nager, R. G., Monaghan, P., & Houston, D. C. (200 0). Within-clutch trade-offs be-
tween the number and quality of eggs: Experimental manipulations in gulls.
Ecology, 81, 1339e1350.
Nei, M., Maruyama, T., & Chakraborty, R. (1975). The bottleneck effect and genetic
variability in populations. Evolution,1e10.
Nonacs, P. (2003). Measuring the reliability of skew indices: Is there one best index?
Animal Behaviour, 65,615e627.
Oster, G. F., & Wilson, E. O. (1978). Caste and ecology in the social insects. Princeton,
NJ: Princeton University Press.
Pamilo, P. (1991). Evolution of colony characteristics in social insects. I. Sex allo-
cation. American Naturalist, 137,83e107.
Pearcy, M., Aron, S., Doums, C., & Keller, L. (2004). Conditional use of sex and
parthenogenesis for worker and queen production in ants. Science, 306,
1780e1783.
Peeters, C., & Molet, M. (2010). Colonial reproduction and life histories. In L. Lach,
C. Parr, & K. Abbott (Eds.), Ant ecology (pp. 159e176). Oxford, U.K.: Oxford
University Press.
Pinheiro, J. C., & Bates, D. M. (200 0). Mixed-effects models in S and S-PLUS. New York,
NY: Springer.
Porter, S. D., & Tschinkel, W. R. (1985). Fire ant polymorphism (Hymenoptera:
Formicidae): Factors affecting worker size. Annals of the Entomological Society of
America, 78, 381e386.
Queller, D. C., & Goodnight, K. F. (1989). Estimating relatedness using genetic
markers. Evolution, 43, 258e275.
R Core Team. (2014). R: A language and environment for statistical computing. Vienna,
Austria: R Foundation for Statistical Computing.
Rangel, J., Mattila, H. R., & Seeley, T. D. (2009). No intracolonial nepotism during
colony ssioning in honey bees. Proceedings of the Royal Society B: Biological
Sciences, 276, 3895e3900.
Rangel, J., Reeve, H. K., & Seeley, T. D. (2013). Optimal colony ssioning in social
insects: Testing an inclusive tness model with honey bees. Insectes Sociaux, 60,
445e452.
Rangel, J., & Seeley, T. (2012). Colony ssioning in honey bees: Size and signicance
of the swarm fraction. Insectes Sociaux, 59, 453e462.
Reber, A., Castella, G., Christe, P., & Chapuisat, M. (2008). Experimentally increased
group diversity improves disease resistance in an ant species. Ecology Letters, 11,
682e689.
Rheindt, F. E., Strehl, C. P., & Gadau, J. (2005). A genetic component in the deter-
mination of worker polymorphism in the Florida harvester ant Pogonomyrmex
badius.Insectes Sociaux, 52,163e168.
Seeley, T. D. (1996). The wisdom of the hive: The social physiology of honey bee col-
onies. Cambridge: Massachusetts: Harvard University Press.
Seeley, T. D., & Morse, R. A. (1977). Dispersal behavior of honey bee swarms. Psyche,
84,199e209.
Seeley, T., & Tarpy, D. (2007). Queen promiscuity lowers disease within honeybee
colonies. Proceedings of the Royal Society B: Biological Sciences, 274,67
e72.
Sepp
a, P., Fern
andez-Escudero, I., Gyllenstrand, N., & Pamilo, P. (2008). Colony
ssion affects kinship in a social insect. Behavioral Ecology and Sociobiology, 62,
589e597.
Smith, C. C., & Fretwell, S. D. (1974). Optimal balance between size and number of
offspring. American Naturalist, 108, 499e506.
Sokal, R. R., & Rohlf, F. J. (1995). Biometry (3rd ed.). New York, NY: W. H. Freeman.
Solís, C. R., Hughes, C. R., Klingler, C. J., Strassmann, J. E., & Queller, D. C. (1998). Lack
of kin discrimination during wasp colony ssion. Behavioral Ecology, 9,172e176 .
Stearns, S. C., & Hoekstra, R. F. (2005). Evolution. An introduction. Second edition.
Oxford, U.K.: Oxford University Press.
Ugelvig, L. V., Kronauer, D. J., Schrempf, A., Heinze, J., & Cremer, S. (2010). Rapid anti-
pathogen response in ant societies relies on high genetic diversity. Proceedings
of the Royal Society B: Biological Sciences, 277, 2821e2828.
van Veen, J. W., & Sommeijer, M. J. (2000). Colony reproduction in Tetragonisca
angustula (Apidae, Meliponini). Insectes Sociaux, 47,70e75.
Wilson, A. J., Reale, D., Clements, M. N., Morrissey, M. M., Postma, E., Walling, C. A.,
et al. (2010). An ecologist's guide to the animal model. Journal of Animal Ecology,
79,13e26.
A. L. Cronin et al. / Animal Behaviour 119 (2016) 1e99
... Thus, across multiple Cataglyphis species, there were apparently no significant paternal genetic effects on worker head width. These results are consistent with those from previous species-specific studies, which found that the association between worker size variation and within-colony genetic relatedness or patriline identity was weak to absent in the ants C. piliscapa and C. bombycina (Leniaud et al. 2015) (but see Cronin et al. 2016b, who discovered that patriline affected worker size in C. cursor). Overall, it seems there is weak selection on polyandry to increase the morphological diversity of workers in Cataglyphis ant species. ...
Article
Full-text available
In eusocial insects, worker polymorphism is shaped by several factors, including colony size, queen mating frequency, and the timing of queen-worker differentiation during larval development. In a comparative study of 18 species of Cataglyphis desert ants representing a wide range of worker sizes, we used phylogenetically controlled analyses to examine correlations between worker head width variation (i.e., worker polymorphism) and multiple social traits, namely, mature colony size, mean worker head width, queen head width, queen-worker head width dimorphism, and within-colony genetic relatedness, resulting from multiple mating by queens. We found that worker polymorphism was positively correlated with mature colony size, mean worker head width, and queen head width. In contrast, worker polymorphism was not correlated with queen-worker dimorphism and within-colony genetic relatedness. These results underscore that evolution of worker polymorphism and social traits are correlated. They also illustrate that additional research using multivariate approaches is needed to further clarify the evolution of insect societies. Significance statement In eusocial insects, worker morphological variation (i.e., worker polymorphism) is tightly linked to division of labor. Multiple factors are supposed to shape the evolution of worker polymorphism. Using phylogenetically controlled analyses of worker head width variation from 18 species of Cataglyphis desert ants, we show that worker polymorphism positively correlates with mature colony size, mean worker head width, and queen size. These results highlight that the evolution of worker polymorphism and social traits are correlated. Identifying the mechanisms underlying these relationships could provide major insights into the development and evolution of insect societies.
... Indeed, large mature colonies produce large offspring colonies but not more offspring colonies (40), and mature colonies seem to produce larger offspring colonies when colony density (competition) is higher (47), a trend similar to that found in the thermophilous ant Aphaenogaster senilis whereby colonies produce smaller colonies when colony density is low (24). In addition, workers allocated to offspring colonies are on average larger than those that remain in the mother colony (50), probably to increase the competitiveness and stress tolerance of offspring colonies because large foragers are more robust and efficient (see Section 2.2). Overall, these data suggest that desert ant species reproducing through DCF maximize the competitiveness and tolerance of their offspring colony. ...
Article
Full-text available
Unlike most desert-dwelling animals, Cataglyphis ants do not attempt to escape the heat; rather, they apply their impressive heat tolerance to avoid competitors and predators. This thermally defined niche has promoted a range of adaptations both at the individual and colony levels. We have also recently discovered that within the genus Cataglyphis there are incredibly diverse social systems, modes of reproduction, and dispersal, prompting the tantalizing question of whether social diversity may also be a consequence of the harsh environment within which we find these charismatic ants. Here we review recent advances regarding the physiological, behavioral, life-history, colony, and ecological characteristics of Cataglyphis and consider perspectives on future research that will build our understanding of organic adaptive responses to desertification. Expected final online publication date for the Annual Review of Entomology Volume 62 is January 07, 2017. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.
Article
Full-text available
Background: Offspring investment strategies vary markedly between and within taxa, and much of this variation is thought to stem from the trade-off between offspring size and number. While producing larger offspring can increase their competitive ability, this often comes at a cost to their colonization ability. This competition–colonization trade- off (CCTO) is thought to be an important mechanism supporting coexistence of alternative strategies in a wide range of taxa. However, the relative importance of an alternative and possibly synergistic mechanism—spatial structuring of the environment—remains the topic of some debate. In this study, we explore the influence of these mechanisms on metacommunity structure using an agent-based model built around variable life-history traits. Our model combines explicit resource competition and spatial dynamics, allowing us to tease-apart the influence of, and explore the interaction between, the CCTO and the spatial structure of the environment. We test our model using two reproductive strategies which represent extremes of the CCTO and are common in ants. Results: Our simulations show that colonisers outperform competitors in environments subject to higher temporal and spatial heterogeneity and are favoured when agents mature late and invest heavily in reproduction, whereas competitors dominate in low-disturbance, high resource environments and when maintenance costs are low. Varying life-history parameters has a marked influence on coexistence conditions and yields evolutionary stable strategies for both modes of reproduction. Nonetheless, we show that these strategies can coexist over a wide range of life-history and environmental parameter values, and that coexistence can in most cases be explained by a CCTO. By explicitly considering space, we are also able to demonstrate the importance of the interaction between dispersal and landscape structure. Conclusions: The CCTO permits species employing different reproductive strategies to coexist over a wide range of life-history and environmental parameters, and is likely to be an important factor in structuring ant communities. Our consideration of space highlights the importance of dispersal, which can limit the success of low-dispersers through kin competition, and enhance coexistence conditions for different strategies in spatially structured environments.
Article
The common approach to the multiplicity problem calls for controlling the familywise error rate (FWER). This approach, though, has faults, and we point out a few. A different approach to problems of multiple significance testing is presented. It calls for controlling the expected proportion of falsely rejected hypotheses — the false discovery rate. This error rate is equivalent to the FWER when all hypotheses are true but is smaller otherwise. Therefore, in problems where the control of the false discovery rate rather than that of the FWER is desired, there is potential for a gain in power. A simple sequential Bonferronitype procedure is proved to control the false discovery rate for independent test statistics, and a simulation study shows that the gain in power is substantial. The use of the new procedure and the appropriateness of the criterion are illustrated with examples.
Article
A new method is described for estimating genetic relatedness from genetic markers such as protein polymorphisms. It is based on Grafen's (1985) relatedness coefficient and is most easily interpreted in terms of identity by descent rather than as a genetic regression. It has several advantages over methods currently in use: it eliminates a downward bias for small sample sizes; it improves estimation of relatedness for subsets of population samples; and it allows estimation of relatedness for a single group or for a single pair of individuals. Individual estimates of relatedness tend to be highly variable but, in aggregate, can still be very useful as data for nonparametric tests. Such tests allow testing for differences in relatedness between two samples or for correlating individual relatedness values with another variable.
Article
Multiple mating by females (polyandry) remains hard to explain because, while it has substantial costs, clear benefits have remained elusive. The problem is acute in the social insects because polyandry is probably particularly costly for females and most material benefits of the behavior are unlikely to apply. It has been suggested that a fitness benefit may arise from the more genetically diverse worker force that a polyandrous queen will produce. One leading hypothesis is that the increased genetic diversity of workers will improve a colony's resistance to disease. We investigated this hypothesis using a polyandrous leaf-cutting ant and a virulent fungal parasite as our model system. At high doses of the parasite most patrilines within colonies were similarly susceptible, but a few showed greater resistance. At a low dose of the parasite there was more variation between patrilines in their resistance to the parasite. Such genetic variation is a key prerequisite for polyandry to result in increased disease resistance of colonies. The relatedness of two hosts did not appear to affect the transmission of the parasite between them, but this was most likely because the parasite tested was a virulent generalist that is adapted to transmit between distantly related hosts. The resistance to the parasite was compared between small groups of ants of either high or low genetic diversity. No difference was found at high doses of the parasite, but a significant improvement in resistance in high genetic diversity groups was found at a low dose of the parasite. That there is generic variation for disease resistance means that there is the potential for polyandry to produce more disease-resistant colonies. That this genetic variation can improve the resistance of groups even under the limited conditions tested suggests that polyandry may indeed produce colonies with improved resistance to disease.
Article
In Mediterranean open habitats, dominant ant species are heat-intolerant and risk-averser, foraging very far from their critical thermal limits (CTM). Subordinate are heat-tolerant (thermophilic) and risk-proner, foraging very near their CTM, running a high heat mortality risk, but having better performance at high temperatures. Thermal tolerance allows a far greater dominance in the ecosystem by subordinates than might be expected from their relative abundance and fighting abilities. Foraging of subordinates is more influenced by temperature than by competition of dominants. The mutual exclusion between dominant and subordinate species seems the result of physiological specialization to different temperature ranges. Focussing on two Cataglyphis species (subordinate and thermophilic), two alternative mechanisms facing extreme heat are described: C. velox polymorphism (large workers are more resistant than small ones), and physiological and behavioural adaptations of C. rosenhaueri workers (all of them of small size).