In many taxa, the subsocial route is considered the main pathway to permanent sociality, but the relative contribution of offspring interactions and parental care to the maintenance of cohesion and tolerance at advanced developmental stages remains poorly studied. Spiders are relevant models for this question because they all show a transient gregarious phase before dispersal, and the transition to permanent sociality, which concerns approximately 20 of the ∼50,000 species, is assumed to rely on the subsocial route. Using spiderlings of the solitary species Agelena labyrinthica, we manipulated the social context to demonstrate that tolerance in aggressive juveniles can be restored when exposed to siblings after moulting. We propose that moulting can reopen closed critical periods and renew the imprinting to social cues and thus lead to the reacquisition of tolerance. Our study highlights the critical role of contacts between juveniles in the expression of tolerance, which opens novel avenues for understanding social transitions.

Sociality represents a crucial step in the evolution of the complexity of life, and has occurred repeatedly in vertebrates and invertebrates (Maynard Smith and Szathmary, 1997). In a broad sense, sociality refers to group living that does not depend on a common attraction for feeding or nesting sites but on the mutual attraction between group members (Grassé, 1952; Costa, 2018). Within this general framework, sociality covers an incredible variety of forms, from the undifferentiated aggregations of cockroaches to the most integrated colonies of social insects. Two alternative routes have historically been proposed to explain the evolutionary transition to sociality (Lin and Michener, 1972). The parasocial route posits that complex forms of sociality arise from aggregations of unrelated individuals of the same generation. The subsocial route, which is considered the main route in many taxa, suggests that advanced sociality evolved from an intermediate ancestor whose offspring stayed together and received parental care. The subsocial route implicitly combines two aspects with relative contributions that are not easy to disentangle, namely the effect of parental care and the role of interactions between siblings. One hypothesis is that parental influence is limited to reducing the propensity of juveniles to disperse by providing them with the resources they need (e.g. food, shelter), but that the driving force behind permanent sociality is the prolonged interactions between juveniles (Wickler and Seibt, 1993).

Spiders are a relevant model to explore this question. The vast majority of the approximately 50,000 known species are solitary at adulthood, but transient gregariousness at the earliest developmental stages is a universal life-history trait in spiders (Foelix, 2011; Chiara et al., 2019). Interestingly, 19 species have developed permanent sociality characterised by cooperative brood care, web construction or hunting, and at least 16 independent derivations of sociality have been identified in spiders (Avilés and Guevara, 2017). The transition to permanent sociality in spiders is thought to derive from the subsocial pathway, as several highly social lineages are nested within clades with ancestral maternal care (Viera and Agnarsson, 2017). Subsocial spiders are defined as species where offspring stay together with the parent beyond the age at which they begin to feed (Yip and Rayor, 2014). Dispersal of juveniles in non-subsocial species usually occurs before the first moult after hatching from the egg case, whereas it happens after one or more moults in subsocial species (Krafft and Horel, 1980). By providing food, mothers can ensure the trophic supply necessary for the development and moulting of juveniles, which do not need to disperse to capture their own prey to meet their nutritional needs as solitary species typically do.

In this study, our approach was to manipulate the social context in a solitary spider species showing no maternal care to reveal the role of spiderling interactions on the expression of tolerance. In Agelena labyrinthica, we have previously shown that cohesion in juveniles relies on mutual attraction and that the onset of aggression results from social isolation after dispersal (Mougenot et al., 2012; Chiara et al., 2019; Mauduit and Jeanson, 2023). Here, we hypothesised that tolerance can be restored when spiders are exposed to their siblings after moulting, even if they were previously aggressive. A validation of this hypothesis would identify the role of social contacts as a unique behavioural mechanism that could explain both the induction of tolerance – a trait common to all species – early in ontogeny, and the maintenance of cohesion at later stages in subsocial species.

Model species and animal collection

Agelena labyrinthica (Clerck 1757) (Agelenidae) is a solitary web-building spider widely distributed across southern and central Europe. Females usually lay one to three egg cases containing an average of 100 eggs (typically between 20 and 200 eggs) in late summer, and die in autumn. Hatching occurs inside the egg case after a period of incubation of approximately 20 days. At this stage, larvae show little mobility and obtain their nutrients from egg yolk. This first instar lasts 1 week and ends with a first moult in the egg case (second instar). Spiderlings then enter a winter diapause until early spring, when they emerge from the egg case and remain gregarious for approximately 1 week before they disperse (Mougenot et al., 2012; Chiara et al., 2019). Second-instar spiderlings are mobile, produce silk and can hunt small prey (Lesne et al., 2016a; Chiara and Jeanson, 2020).

In this study, we used nine egg cases collected between September and October 2020 in southwest France (Table S1). The egg cases were put in a cooled incubator at 4°C in the dark on 29 October 2020, where they stayed until the beginning of the experiments, to mimic winter diapause. The egg cases were placed at room temperature after spending 22 or 23 weeks at 4°C. Before being used in experiments, the egg cases were manually opened (rather than allowing them to emerge naturally) so that all siblings may experience the same social context. Egg cases used in this study contained a total of 101±34 spiderlings (mean±s.d., N=9). Hereafter, we refer to the day of opening as the day of emergence or day 0.

Influence of social isolation on aggressiveness

This experiment was intended to confirm our previous findings that spiderlings reared in isolation develop aggressive interactions (Chiara et al., 2019; Mauduit and Jeanson, 2023). Spiderlings were reared either in isolation (N=16) in small breeding dishes made of white polylactic acid (PLA) (diameter: 17 mm, height: 12 mm) or in pairs (N=25 pairs) in large PLA dishes (diameter: 25 mm, height: 12 mm) (Fig. 1A). All spiderlings reared in pairs were siblings. The size of the arenas differed for isolated and paired spiders to ensure that the available area per individual was similar between pre-treatments. Breeding dishes were sprayed with water every 2 days and never fed to prevent second-instar spiderlings from moulting. After 30 days, spiderlings that were maintained alone or in groups were paired with a second-instar sibling (unfamiliar to those reared in groups) that was reared socially and unfed. Both individuals were introduced simultaneously into a PVC assay arena (diameter: 22 mm, height: 4 mm) containing no silk to avoid any residency bias. We then recorded mortality for 5 days (see below).

Fig. 1.

Experimental design to test the influence of social isolation and moulting on aggressiveness. Spiderlings reared alone (in orange) or socially (in blue) were tested (A) on day 30 without moulting (i.e. during second instar) or (B) within 24 h after moulting (i.e. during third instar). All spiders were tested with a second-instar sibling (in black) reared socially.

Fig. 1.

Experimental design to test the influence of social isolation and moulting on aggressiveness. Spiderlings reared alone (in orange) or socially (in blue) were tested (A) on day 30 without moulting (i.e. during second instar) or (B) within 24 h after moulting (i.e. during third instar). All spiders were tested with a second-instar sibling (in black) reared socially.

Influence of moulting on aggressiveness

This experiment aimed to evaluate the influence of moulting on aggressive behaviour. It was not possible to reuse the individuals assayed before moulting for aggressiveness because most spiders cannibalised their conspecifics after being isolated for 30 days. If we had used the same individuals, we would have selected only those that had survived to moult, which would have introduced a considerable bias. Therefore, we had to use another batch of spiderlings that were reared either in isolation (N=26) in small breeding PLA dishes (diameter: 17 mm, height: 12 mm) or in pairs (N=38 pairs) in large PLA dishes (diameter: 25 mm, height: 12 mm) before moulting (Fig. 1B).

Breeding dishes were sprayed with water every 2 days. As moulting requires spiders to be fed, each spider received at least three fruit flies every other day starting on day 8 (two flies were introduced simultaneously into dishes containing pairs). When we presumed that spiders in pairs were ready to moult (as indicated by their swollen opisthosoma), they were placed individually in a breeding dish until they completed their moult. At this age, moulting lasts approximately 20 min. In 62% of paired spiderlings, one spider moulted before being isolated to complete its moult. Time to moult varied among spiders (mean±s.d.: 21±4 days, N=60; minimum: 14 days, maximum: 34 days) and spiderlings reared socially moulted earlier (mean±s.d.: 20±3 days) than those reared alone (mean±s.d.: 23±4 days) (ANOVA: F1,56=12.7, P<0.001), a phenomenon found in other species such as cockroaches (Holbrook and Schal, 1998). Spiderlings that had just completed their moult to the third instar within a maximum of 24 h (reared socially N=32, reared in isolation N=26, we therefore discarded six pairs of spiders) were then paired in a new PLA arena (diameter: 25 mm, height: 12 mm) containing no silk with a second-instar sibling that had been maintained in group previously and never fed. In these tests, we used larger test arenas than before to account for the larger size of third-instar spiders than second-instar spiders (cephalothorax length: 1.45 mm versus 1.17 mm, Lesne et al., 2016b). For spiderlings reared socially that had already moulted without being previously isolated (N=20 pairs), they were also transferred within the 24 h after the moult to a new PLA arena (diameter: 25 mm, height: 12 mm) with a non-familiar second-instar sibling. In summary, spiders that moulted in the third instar after being reared alone or in groups were paired with a second-instar sibling. In each test, both spiderlings were introduced simultaneously to an arena with no silk. We monitored mortality for 5 days (see below) (Fig. 1B).

Monitoring mortality in pairs

For both experiments, all arenas were checked daily for five consecutive days to record death events. Based on our previous work (Chiara et al., 2019), mortality was considered the result of aggressive interactions. As it was not possible to control precisely the age of the individuals at the time of the moult, we checked that age at moulting did not affect the proportion of dead spiders when in pairs (GLM: χ²1=0.45, P=0.5). Also, we wanted to ensure that any lack of aggressive or cannibalistic behaviour between spiders immediately after moulting would not be due to their temporary inability to express aggressive behaviour. Thus, we isolated 10 spiders and fed them until they moulted. After moulting, we gave them a fruit fly that all spiders killed within 24 h, confirming their ability to hunt rapidly after their ecdysis.

The number of spiderling deaths was compared between treatments using a Cox proportional hazard regression model allowing for censored data. All statistical analyses were performed with the software R 4.0.2 (http://www.r-project.org/).

Second- and third-instar spiderlings reared alone or socially were all tested with second-instar spiderlings reared socially. Mortality was lower when pairs contained a third-instar spiderling (Cox model: χ²1=6.44, P=0.011), but was not influenced by earlier social experience (Cox model: χ²1=2.87, P=0.09), with the interaction between both factors being non-significant (Cox model: χ²1=0.63, P=0.43) (Fig. 2). Moulting reduced mortality from 28% (N=25) to 16% (N=32) in pairs of spiders reared in a group and from 50% (N=16) to 19% (N=26) when pairs contained one spiderling reared alone (Fig. 2). When mortality was observed in pairs involving a third-instar spider, the dead individual was always the second-instar spider.

Fig. 2.

Survival curves of the latencies before one spiderling died in pairs containing a second-instar spider and either another second-instar spider or a third-­instar spider reared alone or socially before moulting. The coloured bands give the 95% confidence intervals. The number of replicates is indicated in parentheses.

Fig. 2.

Survival curves of the latencies before one spiderling died in pairs containing a second-instar spider and either another second-instar spider or a third-­instar spider reared alone or socially before moulting. The coloured bands give the 95% confidence intervals. The number of replicates is indicated in parentheses.

This study showed that exposing spiderlings to siblings after moulting restored social tolerance. It could be argued that the restoration of tolerance is a by-product of nutritional status, independent of the moult as such, as spiders need to be fed in order to moult, and therefore that feeding is responsible for the reduction in aggression. However, we have previously shown that feeding does not affect the tolerance of spiderlings reared in groups but increases the aggression of spiders reared in isolation (Mauduit and Jeanson, 2023). Therefore, our finding that tolerance was renewed after ecdysis in isolated spiderlings, whereas feeding is reported to increase their aggressiveness, further supports our conclusion of the positive influence of moulting on the restoration of amicable social interactions between juveniles. It could also be objected that the observed tolerance is not limited to conspecifics but relies on a general reduction in aggressiveness after moulting, as spiders may not be able to deploy their full behavioural repertoire or may minimise their risk-taking until their new cuticle hardens after ecdysis. This explanation could possibly account for the lack of cannibalism during the very first few hours, but is unlikely to explain the low level of cannibalism for 5 days, as we have observed that spiderlings of A. labyrinthica can kill prey within 24 h of moulting. This argument is consistent with the resumption of feeding reported in the brown recluse spider Loxosceles reclusa, whose juveniles begin to catch prey as early as 20 h after ecdsysis and whose ability to hunt is well established at 48 h (Vetter and Rust, 2010).

Our finding that social contacts after moulting are sufficient for the expression of tolerant behaviours supports the hypothesis that interactions between juveniles could have been more decisive than previously assumed in the transition to permanent sociality. An additional and indirect argument is that two of the 19 known species of social spiders belong to the genus Agelena (A. consociata and A. republican; Chauvin and Denis, 1965; Darchen, 1967), which has a total of 45 species, but none, to our knowledge, have been described as subsocial. All these elements strengthen the idea that the juveniles could be the central driver of social transitions and that the mother, although helping to maintain cohesion, would only play a secondary role.

Moulting in spiders is regulated by ecdysteroids (Bonaric and De Reggi, 1977), whose functions are largely conserved in arthropods (Chang, 1993). In many taxa, changes in ecdysteroid titres are associated with behavioural changes, particularly in relation to the expression of social interactions. For example, injection of 20-hydroxyecdysone increases aggression in the lobster Homarus americanus (Bolingbroke and Kass-Simon, 2001) and stimulates dominance behaviours in the wasp Polistes gallicus (Röseler et al., 1984), but reduces sexual cannibalism in the spider Tegenaria atrica (Trabalon et al., 2005).

It has also been shown in the fruit fly Drosophila melanogaster that ecdysteroid signalling plays a critical role in learning (Ishimoto et al., 2009). We can speculate that hormonal fluctuations related to moulting events could delineate sensitive or critical periods when individuals could imprint on the social cues from conspecifics to develop, maintain or restore a high level of social tolerance. Indeed, social isolation in spiders is thought to cause the forgetting of social cues whose presence normally ensures the maintenance of tolerance (Chiara et al., 2019; Chiara and Jeanson, 2020). Thus, the reopening of a critical period triggered by the moult could allow the spiders to become tolerant again after having been kept alone for several days and having developed aggressive behaviour before the moult. In the same line, it has been shown in the context of filial imprinting in chicks that hormonal treatment can reopen a previously closed sensitive period and allow learning of the cues associated with the imprinted object again (Yamaguchi et al., 2012).

The succession of moulting events during ontogeny in spiders, and perhaps in other arthropod species, may provide repeated opportunities for the reopening of a critical period, leading to repetitive imprinting during development. Therefore, the hormonal changes associated with moulting outside the maternal cocoon may allow the recapitulation of what happens in the egg case where spiderlings, after an initial moult, stay in close contact and are tolerant until emergence and dispersal, which may take place weeks or months later. This would be reminiscent of what has been proposed for some bird species where song learning occurs throughout an individual's life (open-ended learners) and could be seen as the repeated extension or reopening of the sensitive period for learning (Rundstrom and Creanza, 2021).

In conclusion, our results suggest that social interactions between juveniles in conjunction with moulting events may play an unsuspected role in the development of social behaviour in juveniles and possibly in the evolution of sociality by promoting the maintenance of tolerance at later stages of ontogeny.

We thank Gérard Latil for his help with the experiments.

Author contributions

Conceptualization: E.M., R.J.; Methodology: E.M., R.J.; Validation: E.M., R.J.; Formal analysis: E.M.; Writing - original draft: E.M., R.J.; Supervision: R.J.

Funding

E.M. was supported by a PhD grant from the French Ministry of Higher Education and Research. Funding was provided by Centre national de la recherche scientifique (CNRS) (www.cnrs.fr) and Université Toulouse III (www.univ-tlse3.fr) to R.J. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Data availability

The dataset is available on the repository Zenodo (https://doi.org/10.5281/zenodo.7591126).

Avilés
,
L.
and
Guevara
,
J.
(
2017
).
Sociality in spiders
. In
Comparative Social Evolution
(ed.
D.
Rubenstein
and
P.
Abbot
), pp.
188
-
223
.
Cambridge University Press
.
Bolingbroke
,
M.
and
Kass-Simon
,
G.
(
2001
).
20-hydroxyecdysone causes increased aggressiveness in female American lobsters, Homarus americanus
.
Horm. Behav.
39
,
144
-
156
.
Bonaric
,
J. C.
and
De Reggi
,
M.
(
1977
).
Changes in ecdysone levels in the spider Pisaura mirabilis nymphs (Araneae, Pisauridae)
.
Experientia
33
,
1664
-
1665
.
Chang
,
E. S.
(
1993
).
Comparative endocrinology of molting and reproduction: insects and crustaceans
.
Annu. Rev. Entomol.
38
,
161
-
180
.
Chauvin
,
R.
and
Denis
,
J.
(
1965
).
Une araignée sociale du Gabon
.
Biologia Gabonica
1
,
93
-
99
.
Chiara
,
V.
and
Jeanson
,
R.
(
2020
).
Influence of past and current social contexts on hunting behaviour in spiderlings
.
Behav. Ecol. Sociobiol.
74
,
87
.
Chiara
,
V.
,
Ramon Portugal
,
F.
and
Jeanson
,
R.
(
2019
).
Social intolerance is a consequence, not a cause, of dispersal in spiders
.
PLoS Biol.
17
,
e3000319
.
Costa
,
J. T.
(
2018
).
The other insect societies: overview and new directions
.
Curr. Opin. Insect. Sci.
28
,
40
-
49
.
Darchen
,
R.
(
1967
).
Une nouvelle araignée sociale du Gabon Agelena republicana Darchen
.
Biologia Gabonica
3
,
31
-
42
.
Foelix
,
R. F.
(
2011
).
Biology of Spiders
, 3rd edn.
Oxford University Press
.
Grassé
,
P. P.
(
1952
).
Le fait social : Ses critères biologiques, ses limites
. In
Structure et Physiologie des Sociétés Animales
(ed.
P. P.
Grassé
), pp.
7
-
17
.
Paris
,
France
:
CNRS
.
Holbrook
,
G. L.
and
Schal
,
C.
(
1998
).
Social influences on nymphal development in the cockroach, Diploptera punctata
.
Physiol. Entomol.
23
,
121
-
130
.
Ishimoto
,
H.
,
Sakai
,
T.
and
Kitamoto
,
T.
(
2009
).
Ecdysone signaling regulates the formation of long-term courtship memory in adult Drosophila melanogaster
.
Proc. Natl. Acad. Sci. USA
106
,
6381
-
6386
.
Krafft
,
B.
and
Horel
,
A.
(
1980
).
Maternal behaviour and mother–young relationships in spiders
.
Reprod. Nutr. Dev.
20
,
747
-
758
.
Lesne
,
P.
,
Cazalé-Debat
,
L.
,
Portugal
,
F. R.
,
Trabalon
,
M.
and
Jeanson
,
R.
(
2016a
).
Early experience and social performance in spiderlings
.
Anim. Behav.
122
,
9
-
16
.
Lesne
,
P.
,
Trabalon
,
M.
and
Jeanson
,
R.
(
2016b
).
Cannibalism in spiderlings is not only about starvation
.
Behav. Ecol. Sociobiol.
70
,
1669
-
1678
.
Lin
,
N.
and
Michener
,
C. D.
(
1972
).
Evolution of sociality in insects
.
Q. Rev. Biol.
47
,
131
-
159
.
Mauduit
,
E.
and
Jeanson
,
R.
(
2023
).
Ontogenetic change in social context as a cue for a behavioural switch in spiderlings
.
Anim. Behav.
195
,
9
-
18
.
Maynard Smith
,
J.
and
Szathmary
,
E.
(
1997
).
The Major Transitions in Evolution
.
Oxford University Press
.
Mougenot
,
F.
,
Combe
,
M.
and
Jeanson
,
R.
(
2012
).
Ontogenesis and dynamics of aggregation in a solitary spider
.
Anim. Behav.
84
,
391
-
398
.
Röseler
,
P. F.
,
Röseler
,
I.
,
Strambi
,
A.
and
Augier
,
R.
(
1984
).
Influence of insect hormones on the establishment of dominance hierarchies among foundresses of the paper wasp, Polistes gallicus
.
Behav. Ecol. Sociobiol.
15
,
133
-
142
.
Rundstrom
,
P.
and
Creanza
,
N.
(
2021
).
Song learning and plasticity in songbirds
.
Curr. Opin. Neurobiol.
67
,
228
-
239
.
Trabalon
,
M.
,
Niogret
,
J.
and
Legrand-Frossi
,
C.
(
2005
).
Effect of 20-hydroxyecdysone on cannibalism, sexual behavior, and contact sex pheromone in the solitary female spider, Tegenaria atrica
.
Gen. Comp. Endocrinol.
144
,
60
-
66
.
Vetter
,
R. S.
and
Rust
,
M. K.
(
2010
).
Influence of spider silk on refugia preferences of the recluse spiders Loxosceles reclusa and Loxosceles laeta (Araneae: Sicariidae)
.
J. Econ. Entomol.
103
,
808
-
815
.
Viera
,
C.
and
Agnarsson
,
I.
(
2017
).
Parental care and sociality
. In
Behaviour and Ecology of Spiders. Contributions from the Neotropical Region
(ed.
C.
Viera
and
M. O.
Gonzaga
), pp.
351
-
381
.
Springer
.
Wickler
,
W.
and
Seibt
,
U.
(
1993
).
Pedogenetic sociogenesis via the ‘sibling-route’ and some consequences for Stegodyphus spiders
.
Ethology
95
,
1
-
18
.
Yamaguchi
,
S.
,
Aoki
,
N.
,
Kitajima
,
T.
,
Iikubo
,
E.
,
Katagiri
,
S.
,
Matsushima
,
T.
and
Homma
,
K. J.
(
2012
).
Thyroid hormone determines the start of the sensitive period of imprinting and primes later learning
.
Nat. Commun.
3
,
1081
.
Yip
,
E. C.
and
Rayor
,
L. S.
(
2014
).
Maternal care and subsocial behaviour in spiders
.
Biol. Rev.
89
,
427
-
449
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information