Behaviour 140 (2018) 151e159

Contents lists available at ScienceDirect

Animal Behaviour

journal homepage: www.elsevier.com/locate/anbehav

Caste-dependent brood retrieval by workers in the exsecta

* Unni Pulliainen a, b, , Nick Bos a, b, Patrizia d'Ettorre c, Liselotte Sundstrom€ a, b a Centre of Excellence in Biological Interactions, Organismal and Evolutionary Biology Research Programme, Faculty of Biological and Environmental Sciences, University of Helsinki, Helsinki, b Tvarminne€ Zoological Station, Faculty of Biological and Environmental Sciences, University of Helsinki, Helsinki, Finland c Laboratory of Experimental and Comparative Ethology, University of Paris 13, Sorbonne Paris Cite, Paris, France article info abstract

Article history: The ability to distinguish friends from foe is a widespread phenomenon among social . In , Received 5 October 2017 recognition of intruders is important for the maintenance of colony integrity and survival. Intruders are Initial acceptance 4 December 2017 typically adult, but the acceptance of non-nestmate brood could result in severe fitness costs, depending Final acceptance 29 March 2018 on the caste of the brood. Accepting non-nestmate worker brood may not carry a cost, as they should not Available online 24 May 2018 drain resources of the adoptive colony but may instead add to the workforce. Sexual brood, however, MS. number: 17-00790R would typically not contribute to colony performance, yet require resources, and should thus be rejected. Here, we tested whether workers of the narrow-headed ant, Formica exsecta, which strongly discriminate Keywords: ants between adult nestmates and non-nestmates, also discriminate between nestmate and non-nestmate brood pupae. Furthermore, we investigated whether the caste of the brood (workers/sexuals) affects caste discrimination. We carried out analysis of surface chemicals to investigate whether the chemical distance cuticular hydrocarbons between colonies was associated with the propensity to accept non-nestmate pupae. We show that discrimination worker pupae were retrieved irrespective of their origin, whereas nestmate sexual pupae were retrieved inclusive fitness at a slightly higher rate than non-nestmates. Our chemical data, however, suggest that both the repro- pupa ductive and the worker brood carry sufficient chemical information for discrimination, as they both recognition express colony signatures. However, this information is acted upon only in the case of sexual brood. Our retrieval results thus suggest that workers selectively capitalize on the chemical information in agreement with social fitness predictions, albeit to a lower extent than during discrimination between adult individuals. © 2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

The ability to discriminate between friend and foe plays a crucial component, Mateo, 2004). Discrimination relates to an action (e.g. role in the evolution and maintenance of eusocial societies. acceptance, aggression), based on the information perceived during Through efficient recognition, altruistic behaviours can be directed recognition (Liebert & Starks, 2004; action component, Reeve, towards related individuals and the colony and its resources pro- 1989). Therefore, recognition does not necessarily imply action, tected from exploitation by unrelated intruders (d'Ettorre & Lenoir, which can be context dependent (Bos, Guerrieri, & d'Ettorre, 2010; 2011; Hepper, 1991). To this end, social insects have developed Chapuisat, 2004; Downs & Ratnieks, 2000; Nehring, Evison, precise discrimination abilities, and most are able to Santorelli, d'Ettorre, & Hughes, 2011; Sturgis & Gordon, 2012). discriminate between adult nestmates and non-nestmates very Even though social insects are extremely good at discriminating accurately (van Zweden & d'Ettorre, 2010). There is, however, an between adult group members and alien individuals, whether they essential difference between recognition and discrimination. can, or even need to, distinguish between nestmate and non- Recognition is the internal neural process of detecting and classi- nestmate brood is less clear. Given that the immature stages of fying another individual as a member of the group (the perception social insects are immobile, they are not potential intruders of colonies. Thus, the ability to distinguish between nestmate and non-nestmate brood could be under less stringent selection. None the less, the ability to discriminate against non-nestmate brood * Correspondence: U. Pulliainen, Centre of Excellence in Biological Interactions, may be favoured in species that are targeted by social parasites. Organismal and Evolutionary Biology Research Programme, Faculty of Biological Social parasites, whether conspecific or heterospecific, enter the and Environmental Sciences, University of Helsinki, Viikinkaari 1, Biocenter 3, P.O. fi Box 65, FI-00014, Helsinki, Finland. nest and exploit colony resources to their own bene t(Schmid- E-mail address: unni.pulliainen@helsinki.fi (U. Pulliainen). Hempel, 1995), thus severely reducing the fitness of the hosts https://doi.org/10.1016/j.anbehav.2018.04.015 0003-3472/© 2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. 152 U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159

(Buschinger, 2009; d'Ettorre & Heinze, 2001; Topoff, 1990). By do- such difference would be expected in the rate of retrieval of worker ing so, social parasites manipulate the workers of their host in one brood. Finally, we carried out chemical analysis to reveal possible way or another, causing host workers to care for parasite brood at differences between different castes and investigated whether the the expense of the colony's own brood (Davies, Bourke, & de L chemical distance between colonies has an effect on the rates of Brooke, 1989; Lenoir, d'Ettorre, Errard, & Hefetz, 2001). Thus, in brood retrieval. addition to discrimination against adult intruders, accurate discrimination against non-nestmate brood may prevent a social METHODS parasite from gaining foothold in the colony (Achenbach & Foitzik, 2009; Achenbach, Witte, & Foitzik, 2010; Johnson, Topoff, Vander Study Population Meer, & Lavine, 2005). Social parasite pressure may hone precise discrimination between species, which could also select for more Our study population of the narrow-headed ant, F. exsecta,isin precise recognition abilities within species, as has been suggested the Hanko peninsula of southwestern Finland. This population has in the ant Formica fusca (Chernenko, Helantera,€ & Sundstrom,€ been studied extensively with respect to demography, dispersal, 2011). relatedness and inbreeding (Chapuisat, Sundstrom,€ & Keller, 1997; Evidence for discrimination against non-nestmate brood in so- Haag-Liautard, Vitikiainen, Keller, & Sundstrom,€ 2009; Sundstrom,€ cial insects varies across species. Some species discriminate be- Chapuisat, & Keller, 1996; Sundstrom,€ Keller, & Chapuisat, 2003; tween nestmate and non-nestmate brood (Bonavita-Cougourdan, Vitikainen, Haag-Liautard, & Sundstrom,€ 2011; Vitikainen, Haag- Clement, & Lange, 1987; Fen eron & Jaisson, 1992; Helantera€ & Liautard, & Sundstrom,€ 2015). In addition, nestmate recognition Sundstrom,€ 2007; Klahn & Gamboa, 1983; Lenoir, 1984; Meunier, and the surface chemistry of these ants have been studied inten- Delaplace, & Chapuisat, 2010; Pirk, Neumann, & Hepburn, 2007; sively (Martin & Drijfhout, 2009; Martin, Helantera,€ & Drijfhout, Tan et al., 2009; Visscher, 1986), whereas others may readily 2008a; Martin, Shemilt, & Drijfhout, 2012; Martin, Vitikainen, accept brood of unfamiliar origin (Crosland, 1988; Goodloe & Drijfhout, & Jackson, 2012; Martin, Vitikainen, Shemilt, Drijfhout, Topoff, 1987; Holzer, Kümmerli, Keller, & Chapuisat, 2006; & Sundstrom,€ 2013). This species founds colonies by temporary Maeder, Freitag, & Cherix, 2005; Tan et al., 2009). Furthermore, on Serviformica species (Collingwood, 1979; Czechowski, some species can distinguish nestmate from non-nestmate brood, Radchenko, & Czechowska, 2002; Seifert, 2007), but may itself also but nevertheless adopt some non-nestmate brood (Souza, Della occasionally be targeted by social parasites, such as the mound- Lucia, Errard, Richard, & Lima, 2006; Viana et al., 2001). Species- building wood ants of the Formica s.str. group. Colonies can be specific differences in brood discrimination may be context either monogyne (one reproductive queen) or polygyne (multiple dependent, and thus reflect the evolutionary and ecological history queens). Based on previous work (Sundstrom€ et al., 1996) we chose of the species. only colonies known to be monogyne for this study. Discrimination stringency should depend on the costs and benefits of discrimination, and their fitness consequences (Reeve, Experimental Set-up 1989). Thus, worker brood from another colony may be accepted, as they could contribute to the colony workforce, once eclosed We conducted two sets of tests with identical set-ups, to mea- (Buschinger, Ehrhardt, & Winter, 1980; Fen eron & Jaisson, 1995; sure the propensity of workers to retrieve nestmate versus non- Isingrini, Lenoir, & Jaisson, 1985; Orivel, Errard, & Dejean, 1997). nestmate worker pupae (set 1) and sexual pupae (set 2), respec- Hence, they would not impose a cost but would contribute to the tively. We collected workers (ca. 300 workers each colony) and workforce of the adoptive colony. Conversely, sexual brood would pupae from 20 colonies for each of the experiments (four of which usually not contribute to colony performance, but instead require were used in both experiments), by collecting nest material from colony resources during maturation, yet provide no fitness returns inside the nest, and brought these to the laboratory. The collected to the resident colony members (but see Nehring, Boomsma, & nest fragments were housed in boxes (25 15 cm and 10 cm d'Ettorre, 2012). Thus, non-nestmate sexual brood should not be high), the walls of which were coated with Fluon and the bottoms accepted. Despite these relatively straightforward fitness pre- lined with peat as nest material. The ants were kept at room tem- dictions for acceptance of worker versus sexual brood, the effect of perature (ca. 21 C), fed BhathkareWhitcomb diet (Bhatkar & caste on the acceptance or rejection of brood remains largely un- Whitcomb, 1970), and watered daily. explored (Achenbach et al., 2010). Nestmate recognition in social insects is based on a colony- Behavioural Assays specific ‘chemical signature’, which consists of a characteristic mixture of cuticular hydrocarbons (Lenoir, Fresneau, Errard, & Two days after collection, 20 nurse ants were chosen, based on Hefetz, 1999). Most studies on recognition cues have been carried their proximity to the brood in the nestbox, and placed in small out on adult individuals (van Zweden & d'Ettorre, 2010), but surface experimental boxes (diameter 7 cm, height 5 cm; Fig. 1a) lined chemistry has been proposed to be the most likely basis also for with a thin layer of peat. In addition, five nestmate pupae brood recognition (Achenbach et al., 2010; Helantera€ & d'Ettorre, (depending on the set-up, worker or sexual pupae) were added, as 2014; Souza et al., 2006; Viana et al., 2001). None the less, few the lack of access to familiar brood may cause workers to decrease studies have combined behavioural assays on brood recognition their preference for nestmate brood (Fen eron & Jaisson, 1995; Hare with chemical analysis (Achenbach et al., 2010; Johnson et al., & Alloway, 1987; Lenoir, 1981). The ants were left to habituate 2005; Souza et al., 2006). overnight. The following day, each experimental nest was given the Here, we tested whether workers of the ant Formica exsecta can choice between 10 nestmate and 10 non-nestmate pupae (either all discriminate between nestmate and non-nestmate pupae. We also worker-destined [set 1] or all sexual [set 2]), with each colony investigated whether the caste of the brood (workers/sexuals) af- assigned as the donor of non-nestmate pupae for one recipient fects discrimination. Thus, if the caste of the brood does not influ- colony. The pupae were marked with either a red or a blue marker ence retrieval decisions, nestmate brood should be retrieved at a pen (YOKEN xylene-free, low-odour permanent marker) close to faster rate than non-nestmate brood, irrespective of the caste of the the meconium, with the colours assigned randomly for each trial. brood. Furthermore, non-nestmate sexual brood should be At the start of the trial the pupae were placed in a circle in an arena retrieved at a slower rate than nestmate sexual brood, whereas no (diameter 22 cm, height 5 cm), equidistant from the central U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 153

(a) (b) 22 cm

Arena 5 cm

5 cm

7 cm Nestmate pupa Experimental Non-nestmate pupa nestbox

Figure 1. The pupae retrieval set-up. (a) Side view of the set-up with a hole in the middle of the arena to allow access from the experimental nestbox. (b) Top view of the arena with alternating nestmate and non-nestmate pupae placed in a circle (adapted from Rosengren & Cherix, 1981; Maeder et al., 2005). opening, and alternating between nestmate and non-nestmate software (National Institute of Standards and Technology, Gai- pupae (Fig. 1a and b). Once the pupae had been placed in their thersburg, MD, U.S.A.). The software was also used to detect and respective positions, the arena was placed on top of the experi- subtract the background noise from small peaks. mental nestbox. A stick was placed in the central opening of the arena, providing the ants access to the arena to retrieve the pupae Statistical Analyses (Fig. 1a and b). Once the ants started to retrieve the pupae, the origins (nestmate/non-nestmate) and the order of retrieval were Behavioural data recorded. To avoid nonindependence of the data, and to measure Pupa retrieval was analysed separately for the two sets (worker retrieval rate, instead of measuring whether all pupae would be and sexual pupae) using a Cox proportional hazards model (‘coxph’ collected, only the first 10 of the 20 pupae retrieved were consid- in R package survival, R Core Team, 2013), with the order of retrieval ered. The sex of the sexual pupae (set 2) was determined after the of the pupae as the dependent variable and the origin of pupae experiment, based on morphological characteristics (shape of (nestmate/non-nestmate) as a factor. The pupal sex ratios varied in mandibles and abdomen as well as eye size). We used 800 pupae in the colonies used for the retrieval of sexual pupae. Therefore, we total, 10 nestmate and 10 non-nestmate pupae in each of the 20 analysed possible effects of colony sex ratio, using those produced replicates in both set 1 and set 2. by both the donor and recipient colony as covariates for the sexual pupae Cox proportional hazards model. A test for the proportional Chemical Analysis of Surface Hydrocarbons hazards assumption (command ‘cox.zph’ in R package survival), indicated that these assumptions were not violated (all Additional adult workers, worker pupae and sexual pupae from P ¼ 0.26e0.96). As only the first 10 of 20 retrieval events were each colony were used for chemical analysis of their surface recorded, the remaining 10 pupae were entered as censored data. chemistry (five of each caste, except in mixed-sex broods, where three males and three females were used). The cocoons were Chemical data removed with clean sharp forceps under a Leica Wild MZ6/M3B We collected chemical data from seven groups: (1) worker co- stereomicroscope (2.0 magnification), after which the pupal cases coons, (2) developing worker individuals, (3) queen cocoons, (4) (henceforth ‘cocoons’) and the developing individuals inside the developing queen individuals, (5) male cocoons, (6) developing cocoons (henceforth ‘developing individuals’) were placed in male individuals and (7) adult workers. As no previous data were separate 2 ml glass vials (Sigma Aldrich, France). available on surface compounds of pupae in this species, we To extract the surface chemicals, 120 ml of pentane (HPLC included in the analysis all compounds consistently present in at grade) was added to the sample vials for 10 min, after which the least one group (Appendix Table A1). This differs from earlier extract was transferred to a clean 200 ml glass insert (Sigma studies on adult F. exsecta workers (Martin et al., 2008a; Martin, Aldrich), and allowed to evaporate under a fume hood for 2 h. The Helantera,€ & Drijfhout, 2008b; Martin, Helantera,€ & Drijfhout, samples were then rediluted in 50 ml of pentane containing an 2011; Martin, Schemilt et al., 2012; Martin, Trontti, Schemilt, internal standard (5 ng/mlofn-C18), and 2 ml of the liquid was Drijfhout, Butlin & Jackson, 2012; Martin, Vitikainen et al., 2012; injected into an Agilent 7890A gas chromatograph (capillary col- Martin, Vitikainen, Helantera,€ & Drijfhout, 2008; Martin et al., umn: Agilent HP-5MS, 30 m 25 mm 0.25 mm; split- 2013), which generally included only the main alkanes and al- splitless injector; carrying helium gas at 1 ml/min), coupled kenes. The areas of the chosen peaks were standardized by calcu- with an Agilent 5975c mass spectrometer with 70 eV electron lating the ln(Pi, /g(P)) (Aitchison,1986), where Pi is the area of a peak impact ionization. The temperature program started from 70 Cto and g(P) is the geometric mean of all the peak areas of the indi- 200 Cat30 C/min, then from 200 C to 320 Cat4 C/ vidual. We then reduced the number of variables to be included in min, finishing with a hold at 320 C for 5 min. For every 10 the analyses by running a principal component analysis (PCA) on samples, a solvent-only control was included to check for possible the standardized peak areas, and retained the first seven principal contamination; all control runs were clean. The compounds were components (PCs), which together explained nearly 90% of the total identified based on the diagnostic ions in their mass spectra and the variance (Appendix Table A1). retention times and were integrated using MSD Chemstation To assess whether the cocoons and/or the developing in- (Agilent Technologies, Santa Clara, CA, U.S.A.). As they had similar dividuals carry colony-dependent odour profiles, we ran a retention times, some compounds co-eluted into the same peak. permutational MANOVA set to 1000 permutations. We included the These peaks were deconvoluted and identified using the AMDIS 32 Euclidean distances for each of the brood categories, as well as the 154 U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 cocoons and the developing individuals, separately, and used the rate than non-nestmate pupae, leaving a larger proportion of non- PCs as response variables and colony as a factor. The analysis was nestmate pupae in the arena (Fig. 2a; Cox hazard ratio, run using the command ‘Adonis’ in the R package vegan (Oksanen, HR ¼ 0.68, P ¼ 0.01). However, when offered the same choice, Blanchet, Kindt, Legendre, & Minchin, 2013). For queens and males, but with worker pupae, the workers showed no significant prefer- we also included sex, and the interaction between colony and sex, ence for either category of pupa (set 1; Fig. 2b; HR ¼ 0.82, P ¼ 0.15). to test for sex-specific differences. The sex ratio of neither the recipient (nestmate: HR ¼ 0.99, P ¼ 0.86) Finally, we tested whether the chemical distance between nor the donor colony (non-nestmate: HR ¼ 0.99, P ¼ 0.99) had a nestmate workers and non-nestmate pupae was associated with significant effect on the retrieval of the pupae. Samples included both the propensity of workers to accept non-nestmate pupae. To ach- female- and male-biased sex ratios (range 0e1) with an average sex ieve this, we first estimated the position of each colony in a ratio of 0.55 across the colonies. multidimensional space by calculating colony (chemical) averages The chemical profile of F. exsecta workers and pupae (both co- for each PC separately. These averages were calculated for each coons and developing individuals) typically consisted of 32 hydro- group within a colony (workers, cocoons, developing individuals). carbons encompassing alkanes, alkenes and methyl-branched We then used these colony averages to calculate a chemical alkanes (Table A1). For pupae, both the cocoons and the developing Euclidean distance matrix between each colony pair of nestmate individuals of each caste (queens, males and workers) showed a workers, and non-nestmate pupae (with developing individuals colony-specific chemical profile (Table A2). The surface hydrocar- and cocoons separately), and finally ran a binomial generalized bons of queen and male pupae differed significantly from each other, linear model (GLM), to investigate whether the propensity to with respect to both the cocoons and the developing individuals accept non-nestmate pupae was affected by the chemical distance (Table A3). The interaction between sex and colony, however, was not between colonies. We used successes and failures in retrieval significant (Table A3). This suggests that queen and male pupae have (nestmate versus non-nestmate pupae) as a combined response specific odours, regardless of colony origin. (For percentages of each variable and the chemical distance as a covariate. We ran the of the compounds within groups, see Table A4.) analysis using the chemical data for both cocoons and developing The rate at which non-nestmate sexual pupae were retrieved individuals and tested the following assumptions: independence of decreased as the chemical distance between the retrieving workers each data point, distribution of the residuals, variance structure and and the non-nestmate pupae offered increased (GLM: Z ¼ 1.926, the relationship between the response and the predictor, with none P ¼ 0.05; Fig. 3a). This only applied to the developing sexual in- of the assumptions violated. All data were analysed using R version dividuals, not their cocoons (GLM: Z ¼ 0.09, P ¼ 0.92). There was no 3.0.2 (R Core Team, 2013). association between the rate at which non-nestmate worker pupae were retrieved and the chemical distance between the colonies Ethical Note (developing individuals: GLM: Z ¼ 1.441, P ¼ 0.15; Fig. 3b; cocoons: GLM: Z ¼ 0.95, P ¼ 0.952). We confirm that our methods adhere to the ASAB/ABS Guide- lines for the Use of Animals in Research. No rules govern the DISCUSSION treatment of our study organism. To ensure colony survival, only a small fraction of the worker ants and pupae were collected from In agreement with our predictions, we found that workers of any colony, leaving the resident reproductive queen and the ma- F. exsecta retrieved worker pupae irrespective of their colony of jority of workers untouched. Ants were housed in boxes with peat origin, whereas they preferentially retrieved nestmate sexual pu- as nest material, kept at room temperature, and fed and watered pae. This preference was not complete, as non-nestmate sexual daily. Individuals were treated carefully, and chemical extraction pupae were also retrieved, but it was statistically significant. We was done after anaesthesia by freezing. also found that both the cocoons and the developing individuals of worker, queen and male pupae carried colony-specific odours RESULTS (R2 ¼ 0.38e0.51 across castes), which suggests that these ants have the information they need to discriminate between brood from When offered the choice between nestmate and non-nestmate different origins, and that they act upon this information selectively sexual pupae (set 2), workers retrieved nestmate pupae at a higher (Nehring et al., 2011).

(a) (b) 1 HR = 0.68, P = 0.01 1 HR = 0.82, P = 0.15

0.8 0.8

0.6 Nestmate 0.6 95% CI Proportion of pupae left Non-nestmate 0.4 95% CI 0.4

0 2468100 246810 Retrieval event

Figure 2. Proportions of nestmate and non-nestmate (a) sexual pupae and (b) worker pupae not retrieved across the first 10 retrieval events, including the 95% confidence intervals indicated by the dotted lines and shaded areas. U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 155

0.8 (a) Z = 1.93 , P = 0.05 0.8 (b) Z = 1.44 , P = 0.15

0.6 0.6

0.4 0.4 pupae retrieved 0.2 0.2 Proportion of non-nestmate

0 0 4 6 8 10 12 14 4 6 8 10 12 14 Chemical distance between retrieving workers and retrieved pupae (developing individuals)

Figure 3. Proportion of non-nestmate pupae (developing individuals) retrieved against chemical distance of (a) sexual and (b) worker pupae.

We found that non-nestmate worker pupae were readily 2010), while others show acceptance (Crosland, 1988; Haskins & retrieved by workers, yet adult workers in this species are rejected Haskins, 1950; Souza et al., 2006). Conspecific pupae have also (Martin, Vitikainen et al., 2008, 2012). This suggests that rejection is been shown to be either accepted (Haskins & Haskins, 1950; context dependent. Indeed, acceptance of worker pupae should not Maeder et al., 2005; Souza et al., 2006) or rejected (Carlin & impose fitness costs on the colony, but rather may be beneficial Schwartz, 1989; Fouks et al., 2011). These discriminatory rules (Buschinger et al., 1980; Fen eron & Jaisson, 1995; Isingrini et al., may depend on a variety of factors and the studies cited above have 1985; but see Achenbach & Foitzik, 2009). As demonstrated by different conditions and set-ups; hence it is difficult to conclude a the case of slave-making ants (d'Ettorre & Heinze, 2001; Foitzik, general pattern of acceptance according to brood stage. However, DeHeer, Hunjan, & Herbers, 2001; Topoff, 1990), adopting non- according to fitness predictions, discrimination against non- nestmate brood of especially later developmental stages can nestmate brood may be stronger at earlier developmental stages, significantly promote the growth of the adopting colony. This is as these would drain colony resources, compared to pupae, which because the adopted brood does not require resources for devel- would not require additional colony resources for their develop- opment yet may add significantly to the workforce available. Colony ment. Our results provide support for the hypotheses that weak size is a major factor contributing to both worker and queen fitness selection has shaped the caste-specific differences in brood through increased survivorship, productivity and competition acceptance and that the workers selectively capitalize on available (Haag-Liautard et al., 2009; Sundstrom,€ 1995). We note, however, chemical information in agreement with expectations under that we do not know whether the non-nestmate pupae would have adaptive behavioural decisions. been raised into adults or, for example, used as a food resource. Pupal sex ratios had no effect on retrieval, although female and As predicted, when retrieving sexual pupae, workers showed a male pupae differed in their surface chemistry. Hence, the observed preference for nestmates. None the less, non-nestmate sexual pu- retrieval patterns are unlikely to be influenced by the sex of the pae were also retrieved. With sexual pupae, the risks and possible pupae. However, we found that the greater the chemical distance fitness costs are determined by both acceptance and rejection er- between the retrieving workers and the retrieved non-nestmate rors (Downs & Ratnieks, 2000; Reeve, 1989). Accepting non- pupae (the developing individual inside the cocoon), the lower nestmate sexual pupae (acceptance error) is likely to incur costs, the proportion of non-nestmate sexual pupae retrieved. Such a as these individuals do not usually contribute to the performance of pattern did not emerge for the retrieval of worker pupae, yet both the adopting colony yet consume colony resources. Conversely, worker and sexual brood contained significant variation in their rejecting nestmate sexual brood (rejection error) also incurs fitness chemical profile that allows the extraction of information about costs, as the sexual offspring represent the contributions made to colony origin. This provides further support for conditional future generations. Thus, the fact that non-nestmate sexual brood discrimination, where both sexual and worker brood express was also retrieved suggests that the selection pressure from non- colony-specific signatures, but only in the case of sexual brood is nestmate sexual brood is not pervasive, and that the acceptance this information acted upon. Interestingly, we found no correlation threshold is rather adjusted to minimize rejection errors. In natural between retrieval rate and the chemical distance of retrieving conditions invasion of sexual brood is probably not a major threat, workers and cocoons (pupal case analysed alone), for either sexual as the pupae are immobile, and would need transport between or worker brood. nests. However, species that are subject to colony take-over by Pupae have been shown to express colony-specific chemical social parasites, as is occasionally the case in F. exsecta, may have a profiles (e.g. Richard, Poulsen, Drijfhout, Jones, & Boomsma, 2007), stricter acceptance threshold towards intruders (Chernenko et al., and we found that both the developing individuals and the cocoons 2011). of F. exsecta carried a colony-specific chemical signature. A study on Earlier studies have shown that unrelated eggs of F. exsecta are newly emerged callows in Formica truncorum (Johnson & accepted, at least in multiqueen and multinest societies (Holzer Sundstrom,€ 2012) showed that these also carry a colony-specific et al., 2006). Studies on other ants have shown that some species profile, whereas another study on callows in F. exsecta (Martin, discriminate against conspecifics of earlier brood stages, eggs and Trontti, et al., 2012) showed that the profiles diverged from those larvae (Fouks, d'Ettorre, & Nehring, 2011; Helantera€ & Ratnieks, of their natal colony following isolation. In our study, and in the 2009; Helantera€ & Sundstrom,€ 2007; Lenoir, 1981; Meunier et al., study by Johnson and Sundstrom€ (2012), the pupae and the callows, 156 U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 respectively, were not kept in isolation before analysis, so their Aitchison, J. (1986). The statistical analysis of compositional data. London, U.K.: & chemical profiles would not have diverged from those of the natal Chapman Hall. Bhatkar, A., & Whitcomb, W. H. (1970). Artificial diet for rearing various species of colony in the same manner. ants. Florida Entomologist, 53, 229e232. Although both the cocoons and developing individuals had Bonavita-Cougourdan, A., Clement, J. L., & Lange, C. (1987). Nestmate recognition: colony-specific chemical profiles, only the chemical profiles of the The role of cuticular hydrocarbons in the ant Camponotus vagus Scop. Journal of Entomological Science, 22(1), 1e10. developing sexuals (not the cocoons) showed a positive correlation Bos, N., Guerrieri, F. J., & d'Ettorre, P. (2010). Significance of chemical recognition with retrieval probability. This indicates that odours on the devel- cues is context dependent in ants. Animal Behaviour, 80(5), 839e844. oping individual can be detected through the cocoon, a result Boysen, S. T., & Berntson, G. G. (1989). Conspecific recognition in the chimpanzee (Pan troglodytes): Cardiac responses to significant others. Journal of Comparative consistent with those in other species of ants and honeybees in Psychology, 103(3), 215e220. which workers are able to assess the health of brood through the Brandstaetter, A. S., Endler, A., & Kleineidam, C. J. (2008). Nestmate recognition in cocoon or cell cap, respectively (Pull et al., 2018; Rotherbuhler, ants is possible without tactile interaction. Naturwissenschaften, 95(7), 601e608. https://doi.org/10.1007/s00114-008-0360-5. 1964). Ants can also detect nestmate cues from a distance Buschinger, A. (2009). Social parasitism among ants: A review (: without tactile contact (Brandstaetter, Endler, & Kleineidam, 2008), Formicidae). Myrmecological News, 12,219e235. so it should be possible to assess the odours through the cocoon. Buschinger, A., Ehrhardt, W., & Winter, U. (1980). The organization of slave raids in e We assume, however, that the odour profiles of both the cocoon dulotic ants a comparative study (Hymenoptera; Formicidae). Zeitschrift für Tierpsychologie, 53(3), 245e264. https://doi.org/10.1111/j.1439-0310.1980.tb01 and the developing individual inside are detected, probably as a 053.x. blend. Carlin, N. F., & Schwartz, P. H. (1989). Pre-imaginal experience and nestmate brood fl As far as we know, the caste of the brood has not been consid- recognition in the carpenter ant, Camponotus oridanus. Animal Behaviour, 38(1), 89e95. https://doi.org/10.1016/S0003-3472(89)80068-5. ered in previous studies on conspecific brood recognition. Here, we Chapuisat, M. (2004). Nestmate recognition in the unicolonial ant Formica paral- found evidence for some caste-specificity in the responses, a result ugubris. Behavioral Ecology, 16(1), 15e19. https://doi.org/10.1093/beheco/arh128. € that echoes earlier results, which show caste-specific acceptance of Chapuisat, M., Sundstrom, L., & Keller, L. (1997). Sex-ratio regulation: The eco- fi nomics of fratricide in ants. Proceedings of the Royal Society B: Biological Sciences, heterospeci c pupae in the ants Temnothorax and Protomognathus 264(1385), 1255e1260. (Achenbach et al., 2010), and adult conspecifics in the ants Sol- Chernenko, A., Helantera,€ H., & Sundstrom,€ L. (2011). Egg recognition and social enopsis invicta (Vander Meer & Porter, 2001) and Pachycondyla parasitism in Formica ants. Ethology, 117, 1081e1092. & Collingwood, C. A. (1979). The Formicidae (Hymenoptera) of Fennoscandia and luteipes (Kikuchi, Tsuji, Ohnishi, Le Breton, 2007). Caste-biased Denmark. Klampenborg, Denmark: Scandinavian Science Press. acceptance of non-nestmate conspecifics is expected to be based Crosland, M. W. J. (1988). Inability to discriminate between related and unrelated on inclusive fitness benefits, and our results agree with those cited larvae in the ant Rhytidoponera confusa (Hymenoptera: Formicidae). Annals of the Entomological Society of America, 81,844e850. above to suggest that this is the case also within species. Czechowski, W., Radchenko, A., & Czechowska, W. (2002). Ants of Poland. Warsaw, The ability to distinguish between group members and alien Poland: Museum & Institute of Zoology. Retrieved from https://books.google.fi/ individuals is common among social animals, such as birds (Krause, books/about/The_Ants_of_Poland.html?id¼FDS9AAAACAAJ&redir_esc¼y. Krüger, Kohlmeier, & Caspers, 2012; Sharp, McGowan, Wood, & Davies, N. B., Bourke, A. F., & de L Brooke, M. (1989). Cuckoos and parasitic ants: Interspecific brood parasitism as an evolutionary arms race. Trends in Ecology & Hatchwell, 2005; Wilkinson, Specht, & Huber, 2010) and mam- Evolution, 4(9), 274e278. mals (Boysen & Berntson, 1989; Mateo, 2003, 2010). It is even more d'Ettorre, P., & Heinze, J. (2001). Sociobiology of slave-making ants. Acta Ethologica, e crucial for eusocial societies, as not only adults but also brood 3(2), 67 82. https://doi.org/10.1007/s102110100038. d'Ettorre, P., & Lenoir, A. (2011). Nestmate recognition. In L. Lach, C. Parr, & K. Abbott is important for the maintenance of the colony's integrity and (Eds.), Ant Ecology (pp. 194e208). USA: Oxford University Press. survival. However, brood recognition appears to be context Downs, S. G., & Ratnieks, F. L. W. (2000). Adaptive shifts in honey bee (Apis mellifera dependent, with discriminatory rules depending on factors such as L.) guarding behavior support predictions of the acceptance threshold model. Behavioral Ecology, 11(3), 326e333. https://doi.org/10.1093/beheco/11.3.326. brood stage (Lenoir, 1981), parasite pressure (Chernenko et al., Fen eron, R., & Jaisson, P. (1992). Nestmate-brood recognition among workers of 2011), colony structure (Meunier, Delemont, & Lucas, 2011) and different social status in Ectatomma tuberculatum Olivier (Formicidae, Ponerinae). the presence of one or more queens (Helantera€ & Sundstrom,€ 2007; Behavioural Processes, 27(1), 45e51. https://doi.org/10.1016/0376-6357(92) € 90039-G. Sundstrom, 1997). Here we have shown that the caste of the brood Fen eron, R., & Jaisson, P. (1995). Ontogeny of nestmate brood recognition in a can also affect discrimination. Recognition consists of different in- primitive ant, Ectatomma tubercalutum Olivier (Ponerinae). Animal Behaviour, ternal processes, while discrimination is the result of an interplay 50(1), 9e14. Foitzik, S., DeHeer, C. J., Hunjan, D. N., & Herbers, J. M. (2001). Coevolution in host- between both perception and action, and we have shown that the parasite systems: Behavioural strategies of slave-making ants and their hosts. information, which is likely to be perceived, does not necessarily Proceedings of the Royal Society of London B,1139e1146. translate into action. This confirms the importance of combining Fouks, B., d'Ettorre, P., & Nehring, V. (2011). Brood adoption in the leaf-cutting ant chemical data with behavioural experiments to better understand Acromyrmex echinatior: Adaptation or recognition noise? Insectes Sociaux, 58(4), 479e485. https://doi.org/10.1007/s00040-011-0167-9. this complex process. Goodloe, L. P., & Topoff, H. (1987). Pupa acceptance by slaves of the social-parasitic ant, Polyergus (Hymenoptera: Formicidae). Psyche, 94(3e4), 293e302. Haag-Liautard, C., Vitikiainen, E., Keller, L., & Sundstrom,€ L. (2009). Fitness and the Acknowledgments level of homozygosity in a social insect. Journal of Evolutionary Biology, 22(1), 134e142. https://doi.org/10.1111/j.1420-9101.2008.01635.x. Hare, J. F., & Alloway, T. M. (1987). Early learning and brood discrimination in We thank Siiri Fuchs and Annu Tertsonen for their help with the leptothoracine ants (Hymenoptera: Formicidae). Animal Behaviour, 35(6), behavioural assays and Chloe Leroy for technical assistance with 1720e1724. Haskins, C. P., & Haskins, E. F. (1950). Notes on the biology and social behavior of the chemical analyses. This study was funded by LUOVA Graduate archaic ponerine ants of the genera Myrmecia and Promyrmecia. Annals of the School, the Academy of Finland (projects 251337, 252411, 284666 Entomological Society of America, 43(4), 461e491. https://doi.org/10.1093/aesa/ and 289731) and the University of Helsinki. 43.4.461. Helantera,€ H., & d'Ettorre, P. (2014). A comparative study of egg recognition signature mixtures in Formica ants. Evolution, 69, 520e529. https://doi.org/ 10.1111/evo.12590. References Helantera,€ H., & Ratnieks, F. L. W. (2009). Two independent mechanisms of egg recognition in worker Formica fusca ants. Behavioral Ecology and Sociobiology, Achenbach, A., & Foitzik, S. (2009). First evidence for slave rebellion: Enslaved ant 63(4), 573e580. € € workers systematically kill the brood of their social parasite Protomognathus Helantera, H., & Sundstrom, L. (2007). Worker policing and nest mate recognition in americanus. Evolution, 63(4), 1068e1075. https://doi.org/10.2307/25483655. the ant Formica fusca. Behavioral Ecology and Sociobiology, 61(8), 1143e1149. Achenbach, A., Witte, V., & Foitzik, S. (2010). Brood exchange experiments and https://doi.org/10.1007/s00265-006-0327-5. chemical analyses shed light on slave rebellion in ants. Behavioral Ecology, 21(5), Hepper, P. G. (1991). Kin Recognition. Cambridge, U.K.: Cambridge University Press. 948e956. https://doi.org/10.1093/beheco/arq008. U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 157

Holzer, B., Kümmerli, R., Keller, L., & Chapuisat, M. (2006). Sham nepotism as a Nehring, V., Boomsma, J. J., & d'Ettorre, P. (2012). Wingless virgin queens assume result of intrinsic differences in brood viability in ants. Proceedings of the Royal helper roles in Acromyrmex leaf-cutting ants. Current Biology, 22,671e673. Society B: Biological Sciences, 273(1597), 2049e2052. https://doi.org/10.1098/ Nehring, V., Evison, S. E. F., Santorelli, L. A., d'Ettorre, P., & Hughes, W. O. H. (2011). Kin- rspb.2006.3553. informative recognition cues in ants. Proceedings of the Royal Society B: Biological Isingrini, M., Lenoir, A., & Jaisson, P. (1985). Preimaginal learning as a basis of Sciences, 278(1714), 1942e1948. https://doi.org/10.1098/rspb.2010.2295. colony-brood recognition in the ant Cataglyphis cursor. Proceedings of the Na- Oksanen, J., Blanchet, G., Kindt, R., Legendre, P., & Minchin, P. R. (2013). Vegan: tional Academy of Sciences of the United States of America, 82(24), 8545e8547. Community Ecology Package. R package version 2.0-9 https://cran.r-project.org/ Johnson, C. A., & Sundstrom,€ L. (2012). Cuticular chemistry of two social forms in a package¼vegan. facultatively polygyne ant (Hymenoptera: Formicidae: Formica truncorum). Orivel, J., Errard, C., & Dejean, A. (1997). Ant gardens: Interspecific recognition in Annales Zoologici Fennici, 49,1e17. Retrieved from http://www.sekj.org/PDF/ parabiotic ant species. Behavioral Ecology and Sociobiology, 40,87e93. anz49-free/anz49-001.pdf. Pirk, C. W. W., Neumann, P., & Hepburn, R. (2007). Nestmate recognition for eggs in Johnson, C. A., Topoff, H., Vander Meer, R. K., & Lavine, B. (2005). Do these eggs smell the honeybee (Apis mellifera L.). Behavioral Ecology and Sociobiology, 61(11), funny to you?: an experimental study of egg discrimination by hosts of the 1685e1693. https://doi.org/10.1007/s00265-007-0399-x. social parasite Polyergus breviceps (Hymenoptera: Formicidae). Behavioral Pull, C. D., Ugelvig, L. V., Wiesenhofer, F., Tragust, S., Schmitt, T., Brown, M. J. F., et al. Ecology and Sociobiology, 57(3), 245e255. https://doi.org/10.1007/s00265-004- (2018). Destructive disinfection of infected brood prevents systemic disease 0851-0. spread. eLife, 7, e32073. https://doi.org/10.7554/eLife.32073. Kikuchi, T., Tsuji, K., Ohnishi, H., & Le Breton, J. (2007). Caste-biased acceptance of R Core Team. (2013). R: A language and environment for statistical computing. Vienna, non-nestmates in a polygynous ponerine ant. Animal Behaviour, 73(4), Austria: R Foundation for Statistical Computing. 559e565. https://doi.org/10.1016/j.anbehav.2006.04.015. Reeve, H. (1989). The evolution of conspecific acceptance thresholds. American Klahn, J. E., & Gamboa, G. J. (1983). Social wasps: Discrimination between kin and Naturalist, 133,407e435. nonkin brood. Science, 221(4609), 482e484. https://doi.org/10.1126/science.221. Richard, F.-J., Poulsen, M., Drijfhout, F., Jones, G., & Boomsma, J. J. (2007). Specificity 4609.482. in chemical profiles of workers, brood and mutualistic fungi in Atta, Acro- Krause, E. T., Krüger, O., Kohlmeier, P., & Caspers, B. A. (2012). Olfactory kin myrmex, and Sericomyrmex fungus-growing ants. Journal of Chemical Ecology, recognition in a songbird. Biology Letters, 8(3), 327e329. https://doi.org/ 33(12), 2281e2292. https://doi.org/10.1007/s10886-007-9385-z. 10.1098/rsbl.2011.1093. Rosengren, R., & Cherix, D. (1981). The pupa-carrying test as a taxonomic tool in the Lenoir, A. (1981). Brood retrieving in the ant, Lasius niger L. Sociobiology, 6,153e178. Formica rufa group. In P. E. Howse, & J.-L. Clement (Eds.), Biosystematics of social Lenoir, A. (1984). Brood-colony recognition in Cataglyphis cursor worker ants (Hy- insects (pp. 263e282). London, U.K.: Academic Press. menoptera: Formicidae). Animal Behaviour, 32(3), 942e944. https://doi.org/ Rotherbuhler, W. C. (1964). Behavior genetics of nest cleaning in honey bees. IV. 10.1016/S0003-3472(84)80189-X. Responses of F1 and backcross generations to disease-killed brood. American Lenoir, A., d'Ettorre, P., Errard, C., & Hefetz, A. (2001). Chemical ecology and social Zoologist, 4(2), 111e123. parasitism in ants. Annual Review of Entomology, 46(1), 573e599. https:// Schmid-Hempel, P. (1995). Parasites and social insects. Apidologie. Princeton, NJ: doi.org/10.1146/annurev.ento.46.1.573. Princeton University Press. https://doi.org/10.1051/apido:19950307. Lenoir, A., Fresneau, D., Errard, C., & Hefetz, A. (1999). Individuality and colonial Seifert, B. (2007). Die Ameisen Mittel- und Nordeuropas. Tauer, Germany: Lutra- identity in ants: The emergence of the social representation concept. In Verlags-und-Vertriebsges. C. Detrain, J. L. Deneubourg, & J. M. Pasteels (Eds.), Information processing in Sharp, S. P., McGowan, A., Wood, M. J., & Hatchwell, B. J. (2005). Learned kin social insects (pp. 219e237). Basel: Birkhauser.€ https://doi.org/10.1007/978-3- recognition cues in a social bird. Nature, 434(7037), 1127e1130. https://doi.org/ 0348-8739-7_12. 10.1038/nature03522. Liebert, A. E., & Starks, P. T. (2004). The action component of recognition systems: A Souza, D. J., Della Lucia, T. M. C., Errard, C., Richard, F.-J., & Lima, E. R. (2006). focus on the response. Annales Zoologici Fennici, 41(6), 747e764. Behavioural and chemical studies of discrimination processes in the leaf-cutting Maeder, A., Freitag, A., & Cherix, D. (2005). Species- and nestmate brood discrimi- ant Acromyrmex laticeps nigrosetosus (Forel, 1908). Brazilian Journal of Biology, nation in the sibling wood ant species Formica paralugubris and Formica 66(3), 863e871. lugubris. Annales Zoologici Fennici, 42(3), 201e212. Sturgis, S. J., & Gordon, D. M. (2012). Nestmate recognition in ants (Hymenoptera: Martin, S., & Drijfhout, F. (2009). A review of ant cuticular hydrocarbons. Journal of Formicidae): A review. Myrmecological News, 16,101e110. Chemical Ecology, 35(10), 1151e1161. https://doi.org/10.1007/s10886-009-9695-4. Sundstrom,€ L. (1995). Sex allocation and colony maintenance in monogyne and Martin, S. J., Helantera,€ H., & Drijfhout, F. P. (2008a). Colony-specific hydrocarbons polygyne colonies of Formica truncorum (Hymenoptera: Formicidae): The identify nest mates in two species of Formica ant. Journal of Chemical Ecology, impact of kinship and mating structure. American Naturalist, 146(2), 182e201. 34(8), 1072e1080. https://doi.org/10.1007/s10886-008-9482-7. Sundstrom,€ L. (1997). Queen acceptance and nestmate recognition in monogyne Martin, S. J., Helantera,€ H., & Drijfhout, F. P. (2008b). Evolution of species-specific and polygyne colonies of the ant Formica truncorum. Animal Behaviour, 53(3), cuticular hydrocarbon patterns in Formica ants. Biological Journal of the Linnean 499e510. https://doi.org/10.1006/anbe.1996.0300. Society, 95(1), 131e140. https://doi.org/10.1111/j.1095-8312.2008.01038.x. Sundstrom,€ L., Chapuisat, M., & Keller, L. (1996). Conditional manipulation of sex Martin, S. J., Helantera,€ H., & Drijfhout, F. P. (2011). Is parasite pressure a driver of ratios by ant workers: A test of theory. Science, 274(5289), chemical cue diversity in ants? Proceedings of the Royal Society B: Biological 993e995. https://doi.org/10.1126/science.274.5289.993. Sciences, 278(1705), 496e503. Sundstrom,€ L., Keller, L., & Chapuisat, M. (2003). Inbreeding and sex-biased gene Martin, S. J., Shemilt, S., & Drijfhout, F. P. (2012). Effect of time on colony odour flow in the ant Formica exsecta. Evolution, 57(7), 1552e1561. https://doi.org/ stability in the ant Formica exsecta. Die Naturwissenschaften, 99(4), 327e331. 10.1554/02-490. https://doi.org/10.1007/s00114-012-0898-0. Tan, K., Yang, M., Radloff, S. E., Yu, Y., Pirk, C. W. W., & Hepburn, H. R. (2009). Intra- Martin, S. J., Trontti, K., Shemilt, S., Drijfhout, F., Butlin, R., & Jackson, D. (2012). Weak and interspecific brood recognition in pure and mixed-species honeybee col- patriline effects are present in the cuticular hydrocarbon profiles of isolated onies, Apis cerana and A. mellifera. Apidologie, 40(2), 184e191. Formica exsecta ants but they disappear in the colony environment. Ecology and Topoff, H. (1990). Slave-making ants. American Naturalist, 78(6), 520e528. Evolution, 2(9), 2333e2346. https://doi.org/10.1002/ece3.319. Vander Meer, R. K., & Porter, S. D. (2001). Fate of newly mated queens introduced in Martin, S. J., Vitikainen, E., Drijfhout, F. P., & Jackson, D. (2012). Conspecific ant monogyne and polygyne Solenopsis invicta (Hymenoptera: Formicidae) col- aggression is correlated with chemical distance, but not with genetic or spatial onies. Annals of the Entomological Society of America, 94, 289e297. distance. Behavior Genetics, 42(2), 323e331. https://doi.org/10.1007/s10519- Viana, A. M. M., Frezard, A., Malosse, C., Della Lucia, T. M. C., Errard, C., & Lenoir, A. 011-9503-0. (2001). Colonial recognition of fungus in the fungus-growing ant Acromyrmex Martin, S. J., Vitikainen, E., Helantera,€ H., & Drijfhout, F. P. (2008). Chemical basis of subterraneus subterraneus (Hymenoptera: Formicidae). Chemoecology, 11(1), nest-mate discrimination in the ant Formica exsecta. Proceedings of the Royal 29e36. https://doi.org/10.1007/PL00001829. Society B: Biological Sciences, 275(1640), 1271e1278. https://doi.org/10.1098/ Visscher, K. P. (1986). Kinship discrimination in queen rearing by honey bees (Apis rspb.2007.1708. mellifera). Behavioral Ecology and Sociobiology, 18(6), 453e460. https://doi.org/ Martin, S. J., Vitikainen, E., Shemilt, S., Drijfhout, F. P., & Sundstrom,€ L. (2013). Sources of 10.1007/BF00300521. variation in cuticular hydrocarbons in the ant Formica exsecta. Journal of Chemical Vitikainen, E., Haag-Liautard, C., & Sundstrom,€ L. (2011). Inbreeding and repro- Ecology, 39(11e12), 1415e1423. https://doi.org/10.1007/s10886-013-0366-0. ductive investment in the ant Formica exsecta. Evolution, 65(7), 2026e2037. Mateo, J. M. (2003). Kin recognition in ground squirrels and other rodents. Journal of https://doi.org/10.1111/j.1558-5646.2011.01273.x. Mammalogy, 84(4), 1163e1181. https://doi.org/10.1644/BLe-011. Vitikainen, E. I. K., Haag-Liautard, C., & Sundstrom,€ L. (2015). Natal dispersal, mating Mateo, J. M. (2004). Recognition systems and biological organization: The percep- patterns, and inbreeding in the ant Formica exsecta. American Naturalist, 186(6), tion component of social recognition. Annales Zoologici Fennici, 41(6), 729e745. 716e727. https://doi.org/10.1086/683799. Mateo, J. M. (2010). Self-referent phenotype matching and long-term maintenance Wilkinson, A., Specht, H. L., & Huber, L. (2010). Pigeons can discriminate group of kin recognition. Animal Behaviour, 80(5), 929e935. https://doi.org/10.1016/ mates from strangers using the concept of familiarity. Animal Behaviour, 80(1), j.anbehav.2010.08.019. 109e115. https://doi.org/10.1016/j.anbehav.2010.04.006. Meunier, J., Delaplace, L., & Chapuisat, M. (2010). Reproductive conflicts and egg van Zweden, J. S., & d'Ettorre, P. (2010). Nestmate recognition in social insects and discrimination in a socially polymorphic ant. Behavioral Ecology and Sociobio- the role of hydrocarbons. In G. J. Blomquist, & A. Bagneres (Eds.), Insect hy- logy, 64(10), 1655e1663. https://doi.org/10.1007/s00265-010-0979-z. drocarbons: biology, biochemistry and chemical ecology (pp. 222e243). Cam- Meunier, J. M., Delemont, O., & Lucas, C. (2011). Recognition in ants: Social origin bridge, U.K.: Cambridge University Press. matters. PLoS One, 6(5), e19347. 158 U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159

APPENDIX

Table A1 Identification of the compounds found on pupae of F. exsecta as well as their contributions to the principal components (PC)

Peak no. Compound PC1 PC2 PC3 PC4 PC5 PC6 PC7

1 5.9-DiMeC17 0.22 0.12 0.18 0.04 0.05 0.20 0.06 2 x.y-DiMeC17 0.23 0.13 0.24 0.03 0.00 0.08 0.03 3 x.y-DiMeC17 0.23 0.14 0.24 0.04 0.00 0.09 0.03 4 x.y-DiMeC19 0.24 0.10 0.23 0.07 0.01 0.02 0.00 5 x.y-DiMeC19 0.23 0.12 0.24 0.08 0.02 0.04 0.01 6C21 0.13 0.01 0.13 0.24 0.17 0.27 0.50 7 x.y-DiMeC21 0.24 0.03 0.18 0.14 0.01 0.08 0.01 8 x.y-DiMeC21 0.23 0.12 0.23 0.10 0.04 0.01 0.00 9C23:1 0.16 0.20 0.02 0.02 0.39 0.36 0.05 10 C23 0.22 0.17 0.15 0.12 0.01 0.09 0.20 11 3-MeC23 0.15 0.18 0.05 0.11 0.38 0.18 0.09 12 C24:1 0.18 0.23 0.03 0.01 0.15 0.26 0.08 13 3.11-DiMeC23 0.23 0.10 0.24 0.05 0.03 0.01 0.01 14 C24 0.25 0.03 0.04 0.12 0.14 0.15 0.02 15 C25:1 0.20 0.26 0.05 0.05 0.03 0.10 0.06 16 C25 0.20 0.11 0.14 0.12 0.24 0.25 0.04 17 3-MeC25 0.19 0.12 0.08 0.06 0.27 0.06 0.23 18 C26:1 0.16 0.23 0.09 0.02 0.32 0.15 0.17 19 C26 0.11 0.14 0.32 0.25 0.14 0.12 0.30 20 C27:1 0.17 0.24 0.10 0.09 0.28 0.20 0.08 21 C27 0.13 0.19 0.30 0.18 0.14 0.10 0.20 22 5.13DiMe C27 þ 5.11DiMeC27 0.21 0.13 0.17 0.01 0.14 0.17 0.16 23 C28 0.09 0.31 0.24 0.02 0.01 0.06 0.11 24 C29:1 0.17 0.22 0.13 0.20 0.18 0.20 0.02 25 C29 0.11 0.31 0.16 0.23 0.00 0.08 0.06 26 11-MeC29 0.08 0.03 0.05 0.33 0.21 0.36 0.41 27 9-MeC29 0.04 0.10 0.23 0.24 0.36 0.31 0.29 28 7-MeC29 0.06 0.17 0.30 0.29 0.20 0.16 0.08 29 5-MeC29 0.05 0.19 0.08 0.32 0.02 0.26 0.37 30 C30 0.13 0.28 0.14 0.22 0.05 0.07 0.06 31 C31:1 0.16 0.03 0.13 0.37 0.08 0.17 0.07 32 C31 0.12 0.26 0.08 0.33 0.01 0.08 0.16 Variance (%) 37.5 18.7 12.8 7.2 5.4 3.7 2.9 Cumulative variance (%) 37.5 56.3 69.1 76.3 81.6 85.3 88.3

The internal standard peak (C18) was not included in the analyses.

Table A2 Permutational MANOVA calculating the chemical variance between colonies

df Sum of squares Mean square FR2 Pr(>F)

Queens Cocoons Colony 12 162.37 13.5308 3.1499 0.54941 0.002 Residuals 31 133.16 4.2956 0.45059 Total 43 295.53 1.00000 Developing individuals Colony 12 187.32 15.6099 2.143 0.45342 0.004 Residuals 31 225.81 7.2841 0.54658 Total 43 413.13 1.00000 Males Cocoons Colony 14 545.37 38.955 2.6855 0.47835 0.004 Residuals 41 594.74 14e506 0.52165 Total 55 1140.12 1.00000 Developing individuals Colony 13 401.06 30.851 2.0577 0.39483 0.013 Residuals 41 614.72 14.993 0.60517 Total 54 1015.78 1.00000 Workers Cocoons Colony 19 562.12 29.585 2.5106 0.37353 0.001 Residuals 80 942.74 11.748 0.62647 Total 99 1504.86 1.0000 Developing individuals Colony 19 531.06 27.951 2.4841 0.38003 0.001 Residuals 77 866.38 11.252 0.61997 Total 96 1397.44 1.00000

The analysis was run using the command ‘Adonis’ in the R package vegan (pupae~sex)colony, permutations ¼ 1000, method ¼ ‘Euclidean’). Permutation: free. Terms added sequentially (first to last). U. Pulliainen et al. / Animal Behaviour 140 (2018) 151e159 159

Table A3 Permutational MANOVA calculating the chemical variance between the sexes in sexuals, as well as the interaction of colony and sex

df Sum of squares Mean square FR2 Pr(>F)

Cocoons Sex 1 92.64 92.645 9.1639 0.06062 0.004 Colony 18 641.35 35.631 3.5244 0.41965 0.001 Sex)colony 8 66.39 8.299 0.8209 0.04344 0.600 Residuals 72 727.91 10.110 0.47629 Total 99 1528.29 1.00000 Developing individuals Sex 1 105.74 105.745 9.0581 0.06890 0.001 Colony 18 470.25 26.125 2.2379 0.30642 0.004 Sex)colony 7 118.13 16.875 1.4455 0.07697 0.157 Residuals 72 840.53 11.674 0.54770 Total 98 1534.65 1.0000

The analysis 'was run using the command ‘Adonis’ in the R package vegan (pupae~sex)colony, permutations ¼ 1000, method ¼ ‘Euclidean’). Permutation: free. Terms added sequentially (first to last).

Table A4 Average percentage of each of the compounds of the chemical profile for each of the groups

Compound Queens Males Workers

Cocoons Developing Cocoons Developing Cocoons Developing Adults individuals individuals individuals

%SD%SD%SD%SD%SD%SD%SD

5.9-DiMeC17 0.37 0.17 0.27 0.13 0.48 0.32 0.32 0.47 1.47 0.81 1.09 0.39 0.25 0.18 x.y-DiMeC17 0.39 0.19 0.31 0.13 0.48 0.32 0.29 0.32 1.63 0.92 1.47 0.39 0.29 0.23 x.y-DiMeC17 0.78 0.38 0.65 0.26 0.95 0.64 0.62 0.61 3.23 1.87 2.96 0.75 0.58 0.43 x.y-DiMeC19 0.33 0.17 0.29 0.13 0.39 0.28 0.26 0.27 1.40 0.80 1.45 0.39 0.30 0.26 x.y-DiMeC19 0.66 0.35 0.61 0.26 0.78 0.54 0.56 0.54 2.77 1.58 2.80 0.76 0.56 0.47 C21 0.05 0.04 0.11 0.17 0.08 0.13 0.22 0.39 0.11 0.08 0.23 0.37 0.09 0.06 x.y-DiMeC21 0.23 0.12 0.25 0.10 0.27 0.18 0.23 0.21 0.94 0.51 1.05 0.32 0.29 0.26 x.y-DiMeC21 0.55 0.29 0.54 0.22 0.68 0.51 0.48 0.44 2.20 1.24 2.33 0.68 0.51 0.45 C23:1 0.82 0.72 0.04 0.08 0.82 0.89 0.22 0.46 3.86 4.35 0.58 1.40 4.78 5.28 C23 1.04 0.47 0.93 0.62 1.00 0.68 1.90 1.66 1.78 0.52 1.36 1.17 2.95 1.07 3-MeC23 0.03 0.06 0.00 0.01 0.03 0.08 0.02 0.06 0.16 0.20 0.01 0.07 0.12 0.19 C24:1 0.05 0.06 0.00 0.00 0.04 0.07 0.01 0.02 0.17 0.21 0.01 0.04 0.36 0.27 3.11-DiMeC23 0.39 0.23 0.32 0.14 0.46 0.36 0.27 0.29 1.59 0.94 1.54 0.54 0.30 0.27 C24 0.51 0.16 0.48 0.13 0.73 1.07 0.57 0.22 1.24 0.51 1.22 0.29 0.75 0.29 C25:1 3.43 2.54 0.11 0.21 3.27 3.40 0.69 1.69 10.75 7.72 1.06 1.70 17.09 5.73 C25 7.22 1.48 8.95 3.39 6.54 2.14 11.32 4.11 11.72 3.01 10.04 2.69 19.77 6.81 3-MeC25 0.10 0.08 0.07 0.03 0.08 0.10 0.10 0.08 0.22 0.17 0.05 0.07 0.11 0.12 C26:1 0.10 0.15 0.00 0.00 0.07 0.12 0.02 0.06 0.16 0.17 0.01 0.04 0.65 0.38 C26 0.51 0.27 2.11 0.78 0.90 1.38 2.23 0.68 0.38 0.18 0.85 0.21 0.65 0.31 C27:1 6.20 6.78 0.12 0.20 5.31 5.80 0.73 1.66 12.12 8.33 1.13 2.28 23.78 9.83 C27 17.00 2.86 44.74 4.03 18.46 4.72 40.90 6.54 12.24 3.83 29.65 3.83 14.28 5.53 5.13-DiMeC27 0.20 0.09 0.30 0.12 0.25 0.27 0.41 0.17 0.54 0.37 0.27 0.10 0.44 0.15 þ5.11-DiMeC27 C28 2.12 0.48 3.45 0.51 2.56 0.81 2.88 0.63 0.84 0.41 2.39 0.48 0.28 0.14 C29:1 1.94 2.07 0.03 0.07 2.09 5.01 0.18 0.31 2.45 1.72 0.13 0.53 4.44 2.14 C29 44.74 8.88 28.88 6.01 44.60 11.95 27.44 6.81 20.99 7.90 31.28 4.67 4.28 1.71 11-MeC29 0.70 0.65 0.00 0.00 0.60 0.74 0.09 0.35 0.57 1.53 0.04 0.26 0.26 0.85 9-MeC29 0.30 0.29 0.77 1.26 0.31 0.54 0.99 1.37 0.16 0.30 0.12 0.36 0.04 0.11 7-MeC29 0.95 0.63 3.37 1.45 0.92 0.88 3.86 1.92 0.76 0.58 1.53 0.93 0.29 0.32 5-MeC29 0.44 0.25 0.00 0.02 0.46 0.49 0.07 0.20 0.07 0.22 0.09 0.26 0.02 0.06 C30 0.50 0.24 0.29 0.13 0.53 0.59 0.22 0.11 0.16 0.16 0.18 0.21 0.04 0.06 C31:1 1.76 1.17 0.16 0.21 1.30 1.40 0.36 0.40 1.04 0.63 0.47 0.62 0.84 0.57 C31 5.59 1.74 1.85 0.98 4.56 2.25 1.54 0.63 2.29 1.05 2.60 1.10 0.60 0.31