Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Coercion promotes alloparental care in cooperative breeders

Coercion promotes alloparental care in cooperative breeders The official journal of the Behavioral ISBE Ecology International Society for Behavioral Ecology Behavioral Ecology (2023), 34(3), 363–372. https://doi.org/10.1093/beheco/arac125 Original Article Coercion promotes alloparental care in cooperative breeders a,b, a a,c,d, Markus Zöttl, Tanja Schreier, and Michael Taborsky Division of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, CH-3032 Hinterkappelen, Switzerland, Department of Biology and Environmental Science, Linnaeus University, c d Kalmar, Sweden, Max Planck Institute of Animal Behavior, D-78467 Konstanz, Germany, and Institute for Advanced Study (Wissenschaftskolleg) Berlin, D-14193 Berlin, Germany Received 3 June 2022; revised 2 December 2022; editorial decision 6 December 2022; accepted 8 December 2022 Members of social groups may negotiate among each other about the exchange of goods and services. If this involves asymmet- ries between interacting partners, for instance in condition, power, or expected payoffs, coercion may be involved in the bargain. Cooperative breeders are excellent models to study such interactions, because asymmetries are inherent in the relationship between dominant breeders and subordinate helpers. Currently it is unclear whether punishment is used to enforce costly cooperation in such systems. Here we investigated experimentally in the cooperatively breeding cichlid Neolamprologus pulcher whether alloparental brood care provided by subordinates is contingent on enforcement by dominant breeders. We manipulated first the brood care beha- vior of a subordinate group member and then the possibility of the dominant breeders to punish idle helpers. When subordinates were prevented from providing brood care, breeders increased their attacks on them, which triggered increased alloparental brood care by helpers as soon as this was again possible. In contrast, when the possibility to punish helpers was prevented, energetically costly alloparental brood care did not increase. Our results confirm predictions of the pay-to-stay mechanism causing alloparental care in this species and they suggest more generally that coercion can play an important role in the control of cooperation. Key words: coercion, punishment, negotiation, cooperative breeding, pay-to-stay, Neolamprologus pulcher. INTRODUCTION direct fitness benefits from their group membership despite being reproductively suppressed (Taborsky 1984; Balshine-Earn et al. Social interactions are typically characterized by a conflict of 1998; Kingma et al. 2011; Leadbeater et al. 2011). They enjoy ac- interest between agents differing in power (Hammerstein 1981; cess to the cooperatively defended territory with resources such as Rubinstein 1982; Binmore 2010; Phillips 2018; Taborsky et al. food and shelters and may benefit from enhanced safety from pred- 2021). This asymmetry affects negotiations between social part- ators (Heg et al. 2004; Groenewoud et al. 2016). Later in life, they ners about access to resources, group membership, or other ad- may inherit the breeding position in their natal group and benefit vantages (McNamara et al. 1999; Melis et al. 2009). Negotiation from the presence of helpers that previously they had helped to processes, the interaction by which individuals adjust their beha- raise by alloparental care (Stiver et al. 2006, Kingma et al. 2011, vior to their partner’s behavior and signaling to achieve a ben- 2014, Leadbeater et al. 2011). However, the presence of subordi- eficial outcome, may encompass combinations of reciprocal nate group members can impose costs to dominants through re- exchanges and elements of coercion, such as punishment for source competition and by enhancing intra-sexual competition for cheating or withheld cooperation (Clutton-Brock and Parker reproduction (Webster et al. 2004; Mitchell et al. 2009a, 2009b; 1995; Raihani et al. 2012; Quinones et al. 2016; Taborsky et al. Hellmann et al. 2015). If these costs outweigh the benefits, the 2021). Currently, it is unclear to which extent coercion and en- dominant individuals will not tolerate subordinates as group mem- forcement affect mutual relationships and reciprocal exchanges in bers and eventually evict them from their territory (Dierkes et al. highly social animals. 1999; Taborsky 1985, 2016; Zöttl et al. 2013c; Thompson et al. In cooperative breeders, dominant individuals benefit from 2016). alloparental brood care performed by subordinate group members Theoretical models suggest that subordinates may provide help (Russell and Rowley 1988; Emlen and Wrege 1988; Brouwer et al. to compensate for the costs imposed to dominants, thereby re- 2005; Zöttl et al. 2013b). In return, the latter often gain substantial moving the incentive for forcible eviction (Gaston 1978; Kokko et al. 2002; Hamilton and Taborsky 2005). Such services provided Address correspondence to M. Zöttl. E-mail: markus.zottl@lnu.se. © The Author(s) 2023. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 364 Behavioral Ecology in exchange of tolerance in the group have been referred to as METHODS “pay-to-stay” (PTS), which can lead to evolutionarily stable group Experimental subjects and group formation living and cooperation (Quinones et al. 2016). This mechanism Experimental subjects were taken from the laboratory stock popu- may involve a negotiation process and is particularly important lation originating from wild N. pulcher caught in Lake Tanganyika when relatedness between dominants and subordinates is low, the near Mpulungu, Zambia, in the years 1999, 2006, and 2009. The presence of subordinates inflicts costs on dominants, dispersal experimental groups were assembled following a standard proce- and independent breeding by subordinates is constrained, and dure and consisted of a male, a female, and a subordinate helper individuals are generally long-lived. The crucial difference from of unknown sex (Taborsky 1984). Resembling the natural group other mechanisms selecting for cooperative brood care, such as composition, males were the largest individuals in the group kin selection and benefits from group augmentation (García-Ruiz (standard length (SL), mean ± SD: 63.8 mm ± 8.3) followed by fe- et al. 2022), is that subordinates do not provide help voluntarily males (55.8 mm ± 5.5) and helpers (26.6 mm ± 3.0). The minimal but are forced by dominants to do so (Fischer et al. 2014; Naef size difference between the male and female pair members was and Taborsky 2020a, 2020b). PTS predicts that when subordin- 0.5  mm, whereas the female was at least 10  mm larger than the ates fail to produce sufficient benefits to the dominants they are respective helper. Each group was kept in a 100  L compartment punished and eventually evicted from the territory (Gaston 1978; of a subdivided 200 L aquarium (N = 17). All compartments con- Mulder and Langmore 1993). However, the predicted causal ef- tained a semi-transparent tube suspended below the water surface fect of punishment on increased helping behavior has hitherto serving as a retreat shelter, an air-driven biological filter suspended not been demonstrated in non-human animals (Raihani et al. in the water, and 4 flowerpot-halves serving as breeding shelters 2012). (see Figure 1). The bottom of the tank was covered with a mix- In the African cichlid fish Neolamplologus pulcher several key aspects ture of gravel and sand. The water temperature was held constant of the mechanism regulating the asymmetric negotiation between at 26–28°C and the light regime was set at a 13 h-light:11 h-dark dominants and subordinates that characterizes the pay-to-stay re- cycle, simulating natural conditions in Lake Tanganyika. The fish lationship have been experimentally demonstrated (Taborsky were fed six times a week with dried food and 2–4 times weekly 1985; Balshine-Earn et al. 1998; Bergmüller and Taborsky 2005; with additional krill. The experiments were conducted at the Bergmuller et al. 2005; Heg and Taborsky 2010; Zöttl et al. 2013a, Ethologische Station Hasli of the University of Bern, Switzerland 2013d; Fischer et al. 2014; Naef and Taborsky 2020a, 2020b). For between June and October 2012 under the ethical approval license instance, individuals prevented from defending or subordinates that of Veterinaeramt Bern 16/09. are temporarily removed from the group are subsequently attacked In the first days after group formation, the fish were habituated by more dominant helpers and breeders (Balshine-Earn et al. to a mesh cage (mesh size 10  ×  10  mm) that later in the experi- 1998; Fischer et al. 2014; Naef and Taborsky 2020a). Helpers pre- mental procedure was used to prevent breeders from entering the vented from defending the territory against intruders subsequently breeding shelter. Helpers learned that they could enter a shelter de- show increased levels of defense (Bergmuller et al. 2005; Naef and spite the mesh net put over it, whereas the breeders were unable to Taborsky 2020b), and a similar pattern emerged when temporarily pass the mesh. After this initial habituation, families were checked preventing helpers from providing alloparental care (Zöttl et al. for clutches every day. If eggs were found, the experiments took 2013b). Subsequent to the experimental manipulation of defense, place on the first and second day after spawning. helpers increased either cooperative defense or submissive beha- The helpers habituated quickly to the introduced mesh cage. All vior, suggesting that cooperative defense serves an appeasement helpers passed the mesh cage within a few days and our observa- function (Bergmuller and Taborsky, 2005). Similar results were tions confirmed that all helpers had learnt the task before the ex- obtained when helpers were experimentally prevented from dig- periment started. During the time between training and breeding ging out the territory shelter (Naef and Taborsky 2020a). In ad- the groups were exposed to the mesh net cage intermittently to en- dition, subordinates are only accepted in the territory when help sure that they remained habituated to the experimental procedures. is needed (Taborsky 1985; Zöttl et al. 2013c). However, all these experiments manipulated the behavior or presence of subordinate Experimental design and procedure helpers rather than the behavior of dominant breeders, so that the potentially underlying negotiation process between both parties is We started the experiment when the group produced their first still poorly understood. clutch (Figure 1). At that point the fish were habituated to the ex- Here we tested experimentally whether breeders force non- perimental setup and the subordinates had learned that they could breeders to provide alloparental care by direct physical attacks and pass freely through the mesh net temporarily covering the breeding very close proximity, which may signal the threat of imminent ag- shelter (Supplementary Figure 1 and Supplementary Table 1). After gression in N. pulcher. First, we created a situation in which helpers a baseline behavior recording (“initial phase”), the group was ex- either showed normal levels of alloparental brood care or were pre- posed to one of two treatments in the experience phase. In the vented to care for the dominant breeders’ clutch. Then we either experience treatment “help prevented” the whole group was pre- prevented physical social interaction between breeders and helpers vented from caring for the brood for 30  min by placing a close- to inhibit the possibility of punishment, or we allowed unrestricted meshed cage over the breeding shelter so that no fish could pass. physical contact. We predicted that if physical social interaction Previous experiments have shown that individuals react to time between breeders and helpers is required to enforce help, more periods where they were unable to conduct brood care by increasing help should be provided in the treatment where social interaction their brood care investment subsequently (Zottl et al 2013d). In the is possible. In contrast, if the mere presence of dominant breeders experience treatment “help allowed” a broad-meshed cage was put suffices to induce help, increased levels of help should be observed over the breeding shelter for 30 min that allowed the helper to pass, after experimentally imposed idleness, regardless of whether phys- but the breeders were excluded from entering the breeding shelter ical interactions were possible of not. to provide egg care. Zöttl et al. · Coercion promotes alloparental care 365 Initial phase Experience phase Test phase Manipulation of possibility to Behavioural observations of brood care conduct brood care Physical contact Help allowed: No need prevented for compensation Breeder male Physical contact Subordinate allowed Breeder female helper Helper can conduct brood care. Physical contact prevented Help prevented: Need for compensation from helper Breeding shelter Flower pot halves as containing eggs shelters Physical contact allowed All individuals are excluded from conducting brood care. Figure 1 Sketch of the experimental set-up. The experiment started one or two days after the dominant female had spawned for the first time. In the “Initial phase”, fish behavior was recorded without experimental manipulation. In the experience phase, for 30 min either all group members were prevented from entering the breeding shelter (impassable mesh cage put over the breeding shelter; i.e., experience treatment “help prevented”), or the helper could access the breeding shelter to care for the eggs (mesh cage was passable for the helper; i.e., experience treatment “help allowed”). In the test phase, for 20  min the treatments either allowed physical contact between the breeders and helpers enabling punishment by physical attacks (ramming or biting; i.e., experimental treatment “physical contact allowed”), or the breeders were confined in clear perspex cylinders preventing bodily contact (i.e., experimental treatment “physical contact prevented”). All groups underwent all four possible combinations of experience and test phases in random order yielding a within subject, full factorial experimental design. In the test phase immediately following the experience phase, contact allowed” condition the cylinders were inserted in the same each group was assigned to a treatment either preventing social way, but the breeders remained outside and were unconstrained interaction between breeders and helpers (“physical contact pre- in their ability to physically attack helpers (Figure 1). All experi- vented”) or allowing social interaction (“physical contact allowed”, mental groups underwent all treatments in the experience and test see Figure 1). In both situations, the broad-meshed cage was put phases in random order, so that each group was tested four times. over the breeding shelter which allowed helpers to care for the eggs The minimum time between two consecutive trials was 1 hour. In whereas the breeders were excluded from entering the shelter. In total, 17 groups were used in the experiment and a total of 68 trials the experimental treatment “physical contact prevented” physical of the test phase were conducted. In 11 of these 17 groups, the interactions were prevented by restricting the breeders to an area experimental trials were conducted on two consecutive days, with in the tank using transparent perspex cylinders. Hence, they re- two trials per day separated by at least an interval of one hour. In mained visually present and could interact with the helper, but they 6 groups we conducted all 4 trials on the same day, again with one were unable to exert any overt physical aggression. In the “physical hour minimum time between consecutive trials. 366 Behavioral Ecology To answer the question whether breeders increased aggression Brood care and behavioral observations towards the subordinate and whether the subordinate increased In the test phase, we conducted continuous focal observation of submission to the breeders in response to preventing the possibility 15  min on the helper individual (Altmann 1974) focusing on two to provide brood care we used the behavioral data collected in the alloparental care behaviors that have contrasting energetic im- experience phase and specified the frequency of overt aggression, plications. We recorded the time individuals engaged in fanning. This behavior serves to create a water current around the clutch and increases the supply of oxygen for the clutch. It is energet- 25 (a) ically costly because it involves a forward swimming movement Help treatment * contact treatment: p = 0.001 that is compensated by intense beats with the pectoral fins to keep the fish on the spot (Grantner and Taborsky 1998; Taborsky and Grantner 1998). Second, we quantified egg cleaning by counting the frequency of micro nips towards the eggs, which serves a hygi- enic function by removing fungi and other microorganisms. This brood care behavior probably involves less energy expenditure (Taborsky and Grantner 1998). Additionally, in each phase of the experiment, we recorded social interactions between breeders and helpers during the 15 min focal observations. We present frequen- cies of overt physical aggression from both breeders towards the helper (ramming and biting), frequencies of restrained aggres- sion (threat signals without physical contact), and the frequencies of submissive behavior (tail quivering) of the helper towards the (b) breeders. We did not record other submissive behaviors (submis- sive postures and hook swimming) because they are usually shown less frequently and it is unclear whether they involve energetic costs comparable to tail quivering. Statistical analyses To model fanning duration, we fitted the log transformed duration of fanning behavior as response variable and added the experi- mental condition of the experience phase (factor with two levels: help prevented/allowed) and the current experimental treatment (factor with two levels: physical interactions prevented/allowed) as fixed factors. We included the interaction between these fixed factors and specified a random effect that identified the experi- Prevented contact Contact Prevented contact Contact mental group. We modeled the data specifying a Gaussian distri- Help allowed Help prevented bution using a Linear Mixed Effect Model (LMM). To model egg cleaning behavior we fitted the number of egg cleaning events as Figure 2 the response variable and again added the experimental condition Fanning duration (A) and egg cleaning frequency (B) of helpers in the test of the experience phase and the current experimental treatment as phase, as measures of alloparental care in response to the experimental fixed factors. We included the interaction between the fixed factors treatments. The boxes depict the median and the interquartile ranges, and and the identity of the test group as random effect. Because the the whiskers show the range of the raw data from N = 17 observations egg cleaning count data showed signs of overdispersion we mod- in each condition. The dots represent data points of each replicate. Stars eled the data specifying a negative binomial distribution using a indicate significance at the level of p < 0.05. Full model details are shown in Table 1. Generalized Linear Mixed Effect Model (GLMM). Table 1 Brood care behavior of the helper in the test phase. Shown are the estimates and incidence rate ratios, 95% confidence intervals, and p-values of an LMM and a GLMM modeling the fanning duration and the frequency of egg cleaning during 15 min observations, respectively. As predictors the model included the treatments of the experience and test phases, and their interaction. Both models included the group identity as random effect (N = 17 groups, N = 68 observations). Significant p -values are printed in bold. Fanning Egg cleaning Predictors Estimates CI p Incidence Rate Ratios CI p (Intercept) 1.02 0.52–1.53 <0.001 44.89 30.36–66.37 <0.001 Treatment experience phase 0.25 −0.30 to 0.80 0.367 1.7 1.18–2.46 0.005 Treatment test phase -0.39 -0.94–0.16 0.166 1.52 1.05–2.20 0.027 Experience * test phase 1.33 0.55–2.11 0.001 1 0.63–1.61 0.991 2 2 Marginal R /Conditional R 0.234/0.548 0.188/0.625 Frequency of egg cleaning Fanning duration (s) Zöttl et al. · Coercion promotes alloparental care 367 Breeder male to helper Breeder female to helper Initial phase Experience phaseTest phase Initial phase Experience phase Test phase 15 (a) 15 (b) ** 10 10 5 5 0 0 (c) (d) 30 30 * * 20 20 10 10 0 0 Figure 3 Frequency of overt aggression (A & B) and restrained aggression (C, D) against the helper from the male (A & C) and the female (B & D) breeder during the different phases and treatments. The boxes depict the median and the interquartile ranges, and the whiskers show the range of the raw data from N = 17 observations in each condition. Stars indicate significance at the level of p < 0.05. Full model details are shown in Tables 2 and 3. restrained aggression and of submission as response variable in sep- they had been prevented from conducting alloparental care in arate models. We fitted the treatment as factor with 2 levels (help the experience phase and physical contact between breeders and allowed/ help prevented) and the group identity as random factor. helpers was allowed in the test phase (Table 1, Figure 2A). We also modeled the distribution of aggressive and submissive be- Helper egg cleaning in the test phase varied in dependence of havior longitudinally throughout the experiment when the groups both the help prevented in the experience phase and the manipula- moved from the initial phase to the experience phase (help pre- tion of physical contact in the test phase (Table 1, Figure 2B). Help vented) and then to the test phase (contact allowed) by fitting the prevented during the experience phase triggered increased levels of phase as factor with 3 levels and group identity as random effect. egg cleaning behavior of helpers during the test phase. The same To conduct all pairwise comparisons between the 3 phases we sub- effect was produced by the treatment in which dominants could sequently used a Tukey post-hoc analysis with adjusted significance exert overt aggression on the subordinates during the test phase level for multiple comparisons. Here too, we used a negative bino- (Figure 2B, Table 1B). In contrast to fanning behavior, the inter- mial distribution to account for overdispersion of the data. action between these two factors was not significant suggesting that All analyses were conducted in R Version 4.2 (R Core Team this form of energetically cheap alloparental brood care did not 2022) using the packages lme4 and glmmTMB (Bates et al. 2011; only depend on overt aggression from breeders (Table 1B). Brooks et al. 2017). All figures were produced from raw, untrans- Breeder aggression against the helper formed data. Overt attacks (including physical contact) depended on the treat- ment and the phase of the experiment. During the experience RESULTS phase treatment “help prevented”, breeders attacked the helper Alloparental care more often than during the treatment “help allowed” (Figure 3A and B, Table 2A). The longitudinal comparison between the three The subordinates provided different amounts of alloparental brood different phases also suggested an effect of the experimental manip- care depending on the behavioral manipulation treatment. The ulation “help prevented” during the experience phase on breeder helpers showed more fanning than in all other treatments when Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Frequency overt aggression Frequency restrained aggression 368 Behavioral Ecology Helper submission to breeders aggression: overt attacks increased from the initial phase to the ex- Initial phase Experience phase Test phase perience phase (when helping was prevented) and decreased there- * * after in the test phase (when helpers resumed brood care, Figure 3A and B, Table 2B). This pattern was the same for both the male and female breeders, with the exception that female breeders did not show a significant decrease of aggression between experience and test phases (Figure 3A and B; Table 2B). Restrained aggression (threat signals) was not equally affected by the treatment, but also changed throughout the phases of the experiment. During the experience phase treatment “help pre- vented”, breeders showed restrained aggression toward the helper at similar rates as when the helper had the opportunity to con- duct brood care (Figure 3C and D, Table 3A). The longitudinal comparison between the three different phases again showed an effect of the manipulation “help prevented” during the experi- ence phase on breeder restrained aggression. Restrained aggres- sion increased from the initial phase to the experience phase in males when helping was prevented (Figure 3C, Table 3B). In fe- males the increase in restrained aggression was significant when the initial phase was compared with the test phase (Figure 3D; Figure 4 Table 3B). Frequency of tail quivering exhibited by the helper toward the male and female breeders during the different phases and treatments of the Helper submission towards breeders experiment. The boxes depict the median and the interquartile ranges, and the whiskers show the range of the raw data from N = 17 replicates in each Helpers showed more tail quivering towards the breeders in the ex- condition. Stars indicate significance at the level of p < 0.05. Full model perience phase treatment “help prevented” than when they were details are shown in Table 4. Table 2 Overt aggressive behavior of the male and female breeders in A) the experience phase and B) longitudinally throughout the experimental phases when help was prevented during the experience phase and contact was allowed during the test phase. Shown are the incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression during 15 min observations. As predictors the model in A) included the treatment of the experience phase and B) the three different stages of the experiment. Both models included the group identity as random effect (N = 17 groups and N = 68 or 51 observations, respectively). Significant p -values are printed in bold A) Female overt aggression Predictors Male overt aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI (Intercept) 0.66 0.31 – 1.38 0.271 1.59 0.85 – 2.98 0.147 Help prevented 2.88 1.51 – 5.50 0.001 1.93 1.16 – 3.20 0.011 2 2 Marginal R /Conditional R 0.154/ 0.433 0.068/ 0.438 B) Male overt aggression Female overt aggression Predictors Incidence Rate Ratios CI p Incidence Rate Ratios CI p (Intercept) 0.14 0.04 – 0.50 0.002 0.77 0.28 – 2.10 0.611 Experience Phase 7.17 3.05 – 16.84 <0.001 4.1 1.55 – 10.90 0.005 Test Phase 0.83 0.25 – 2.73 0.763 2.37 0.83 – 6.72 0.106 Tukey pairwise comparisons (adjusted for multiple comparisons) Estimate Standard Error p Estimate Standard Error p Initial vs Experience -1.969 0.436 <0.001 -1.412 0.498 0.018 Initial vs Test 0.182 0.606 0.951 -0.861 0.533 0.249 Experience vs Test 2.152 0.472 <0.001 0.551 0.385 0.335 2 2 Marginal R /Conditional R 0.220/0.728 0.191/0.369 Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Frequency of tail quiver Zöttl et al. · Coercion promotes alloparental care 369 Table 3 Restrained aggressive behavior of the male and female breeders in A) the experience phase and B) longitudinally throughout the experimental phases when help was prevented during the experience phase and contact was allowed during the test phase. Shown are the incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression during 15 min observations. As predictors the model in A) included the treatment of the experience phase and B) the three different stages of the experiment. Both models included the group identity as random effect (N = 17 groups and N = 68 or 51 observations, respectively). Significant p -values are printed in bold A) Predictors Male restrained aggression Female restrained aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI p Intercept 6.63 4.13–10.65 <0.001 13.61 9.08–20.40 <0.001 Help prevented 0.83 0.51–1.37 0.476 0.93 0.54–1.62 0.808 2 2 Marginal R /Conditional R 0.006/0.323 0.001/0.092 B) Predictors Male restrained aggression Female restrained aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI p (Intercept) 2.3 1.08–4.89 0.031 5.71 3.02–10.81 <0.001 Experience Phase 2.27 1.18–4.39 0.015 1.64 0.86–3.12 0.135 Test Phase 1.94 0.92–4.10 0.083 2.39 1.19–4.79 0.015 Tukey pairwise comparisons (adjusted for multiple comparisons) Estimate Standard Error p Estimate Standard Error p Initial vs Experience −0.82 0.34 0.048 −0.5 0.33 0.3 Initial vs Test −0.66 0.38 0.21 −0.87 0.35 0.048 Experience vs Test 0.16 0.32 0.87 −0.38 0.38 0.5 2 2 Marginal R /Conditional R 0.090/0.426 0.124/0.219 allowed to care for the brood (Figure 4, Table 3A). The longitu- that dominant breeders punish idle subordinates, which should dinal comparison revealed that the submission rate of the helpers in turn enhance their helping effort. We manipulated both the increased from the initial phase to the experience phase (when helpful behavior of subordinates and the respective response of helping was prevented) and decreased thereafter significantly in the dominants, and we measured the behaviors resulting from the test phase (when helpers reassumed brood care; Figure 4, Table these manipulations. Our data reveal that indeed, idle subordi- 2B). nate group members were punished by enhanced overt attacks from dominant breeders, which in turn caused the affected sub- ordinate to increase its helping effort. In contrast, when pun- DISCUSSION ishment was experimentally prevented, helpers did not increase There are many indications in a wide range of taxa that energetically costly fanning behavior. This confirms the predic- interacting agents negotiate among each other about the ex- tions of punishment and its functionality, which is the basis of change of goods and services (reviewed in Taborsky et al. 2021). negotiations in a pay-to-stay relationship, thereby confirming Pertinent theoretical studies have outlined the conditions and the causality between social enforcement and helping behavior. functionality of such trade and suggest that when social partners Such feedback control and social regulation of cooperation has negotiate among each other, power asymmetry among them may been suggested to operate also in some other social species, for lead to punishment and enforcement of cooperative behavior example in paper wasps (Polistes fuscatus; [Reeve and Gamboa (Rubinstein 1982; Clutton-Brock and Parker 1995; McNamara et 1987]), and we suggest that this may be more widespread than al. 1999; Gardner and West 2004; McNamara and Leimar 2010; currently assumed. Press and Dyson 2012; Hellmann and Hamilton; 2018). However, In asymmetric relationships, the answer to the question how the prediction that aggressive behavior can induce cooperation much agents that are more powerful than their social partners in idle subordinates has rarely been tested. A critical test of this can demand from them depends on the alternatives, or “out- hypothesis must involve experimental manipulations of the re- side options”, of the latter (Reeve 2000; Cant et al. 2012; Cant sponses of both partners to each other’s actions. This was the aim and Johnstone 2009). In N. pulcher, the outside options for sub- of our study. ordinate group members are typically poor because of the high Based on the hypothesis that social partners with greater mortality risk when leaving a group (Taborsky and Limberger power can enforce service by less powerful partners we predicted 1980; Taborsky 1984; Heg et al. 2004; Jungwirth et al. 2015). 370 Behavioral Ecology Table 4 Zöttl et al. 2013b). Hence, dominants should not raise the price too Tail quivering of helper displayed toward the male and female high for the tolerance of subordinate group members, but this should breeders as measure of submission in A) the experience phase depend on the local conditions determining the outside options of and B) longitudinally throughout the experimental phases when subordinates. The remarkable variation of group composition and help was prevented and contact was allowed. Shown are the group size between populations exposed to different levels of preda- incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression tion risk supports this conjecture: in populations with enhanced pre- during 15 min observations. As predictors, the model in A) dation risk the groups contained comparatively more large helpers, included the treatment of the experience phase and in B) the which may suggest that the breeders can demand more help from three different stages of the experiment. Both models included their subordinates when predation pressure is high (Groenewoud the group identity as random effect (N = 17 groups and N = et al. 2016). However, our study worked with standardized groups 68 or 51 observations, respectively). Significant p -values are printed in bold of three individuals and it is currently unclear how group size and composition would affect the results of this experiment. Helper submission In contrast to the evidence for PTS, kin selection is apparently of minor importance concerning the relationship among group Incidence p members in this species, as large helpers are often unrelated to the Predictors Rate Ratios CI dominant breeders (Dierkes et al. 2005; Stiver et al.; 2005). In a A) laboratory experiment, unrelated helpers exhibited higher levels of (Intercept) 15.72 11.50–21.49 <0.001 alloparental brood care than related helpers, confirming an impor - Help prevented 1.67 1.31–2.11 <0.001 2 tant prediction of PTS and refuting kin selection as an alternative Marginal R  / 0.119/0.572 explanation (Zöttl et al. 2013d). Furthermore, only unrelated indi- Conditional R B) viduals increased alloparental care when the costs of their presence (Intercept) 14.85 10.46–21.08 <0.001 in the territory were experimentally increased (Zöttl et al. 2013d). Experience Phase 1.84 1.32–2.58 <0.001 In conclusion, our study demonstrates the causal relationship Test Phase 0.8 0.53–1.20 0.277 between punishment for withheld help and subsequent compen- sation by enhanced cooperation, which reflects an important part Tukey pairwise comparisons (adjusted for multiple of the negotiation process among individuals in asymmetric social comparisons) relationships. Estimate Standard Error p SUPPLEMENTARY MATERIAL Initial vs Experience −0.61 0.171 0.0024 Initial vs Test 0.224 0.206 0.5266 Supplementary material can be found at http://www.beheco. Experience vs Test 0.834 0.186 <0.001 oxfordjournals.org/ Observations 51 Marginal R / 0.220/0.543 2 We thank all members of the Ethologische Station Hasli for help during Conditional R the data collection and Stefan Fischer and Arne Jungwirth for helpful discussions. Hence, subordinates should invest as much cooperative effort as FUNDING is required to be tolerated in a safe territory. A previous experi- This work was supported by grants from the Swiss National Science ment investigating the shelter digging effort of helpers, which is Foundation to Michael Taborsky (SNSF grants 310030B_138660, an energetically very expensive behavior (Grantner and Taborsky 31003A_156152, and 31003A 176174). 1998; Taborsky and Grantner 1998), suggested that this is in- deed the case. When large, reproductively mature helpers were DATA AVAILABILITY given the chance to disperse and reproduce elsewhere in a safe Analyses reported in this article can be reproduced using the data provided environment, they reduced the shelter digging effort in their by Zöttl et al 2022. group (Bergmuller et al. 2005) as predicted by pay-to-stay theory (Kokko et al. 2002). In the field, helpers reduced their cooper - Handling Editor: Amanda Ridley ative antipredator defense effort shortly before dispersing (Zöttl et al. 2013a). Similarly, in paper wasps, Polistes dominula, subordi- REFERENCES nate group members reduced their cooperative foraging effort in Altmann J. 1974. Observational study of behavior: sampling methods. their group when outside options were experimentally enhanced Behavior. 49:227–265. (Grinsted and Field 2017). The extent to which outside options Balshine-Earn S, Neat FC, Reid H, Taborsky M. 1998. Paying to stay or affect the negotiation process between helpers and breeders would paying to breed? Field evidence for direct benefits of helping behavior in be a worthwhile subject for future studies. a cooperatively breeding fish. Behav Ecol. 9:432–438. In cooperatively breeding animals dominants typically also rely Balshine S, Leach B, Neat F, Reid H, Taborsky M, Werner N. 2001. Correlates of group size in a cooperatively breeding cichlid fish on the presence and cooperation of subordinates. In N. pulcher, (Neolamprologus pulcher). Behav Ecol Sociobiol. 50:134–140. depending on the local predation pressure breeders may rarely be Bates D, Maechler M, Bollinge B. 2011. lme4: linear mixed-effects models able to raise young on their own (Brouwer et al. 2005; Groenewoud using S4 classes. R package version 0.999375-42 ed. et al. 2016). In addition, they benefit from the presence of subordi- Bergmuller R, Heg D, Taborsky M. 2005. Helpers in a cooperatively breeding cichlid stay and pay or disperse and breed, depending on eco- nate brood care helpers by load lightening and enhanced produc- logical constraints. Proc R Soc B Biol Sci. 272:325–331. tivity (Taborsky 1984; Balshine et al. 2001; Taborsky et al. 2007; Zöttl et al. · Coercion promotes alloparental care 371 Bergmuller R, Taborsky M. 2005. Experimental manipulation of helping Mcnamara JM, Gasson CE, Houston AI. 1999. Incorporating rules for re- in a cooperative breeder: helpers “pay to stay” by pre-emptive appease- sponding into evolutionary games. Nature. 401:368–371. ment. Anim Behav. 69:19–28. Mcnamara JM, Leimar O. 2010. Variation and the response to variation Binmore K. 2010. Bargaining in biology? J Evol Biol. 23:1351–1363. as a basis for successful cooperation. Philos Trans R Soc B Biol Sci. Brooks ME, Kristensen K, Van Benthem KJ, Magnusson A, Berg CW, 365:2627–2633. Nielsen A, Skaug HJ, Machler M, Bolker BM. 2017. glmmTMB balances Melis AP, Hare B, Tomasello M. 2009. Chimpanzees coordinate in a nego- speed and flexibility among packages for zero-inflated generalized linear tiation game. Evol Hum Behav. 30:381–392. mixed modeling. R J. 9:378–400. Mitchell JS, Jutzeler E, Heg D, Taborsky M. 2009a. Dominant members of Brouwer L, Heg D, Taborsky M. 2005. Experimental evidence for helper cooperatively-breeding groups adjust their behaviour in response to the effects in a cooperatively breeding cichlid. Behav Ecol. 16:667–673. sexes of their subordinates. Behaviour. 146:1665–1686. Cant MA. 2012. Suppression of social conflict and evolutionary transitions Mitchell JS, Jutzeler E, Heg D, Taborsky M. 2009b. Gender differences in to cooperation. Am Nat. 179:293–301. the costs that subordinate group members impose on dominant males in Cant MA, Johnstone RA. 2009. How threats influence the evolutionary res- a cooperative breeder. Ethology. 115:1162–1174. olution of within-group conflict. Am Nat. 173:759–771. Mulder RA, Langmore NE. 1993. Dominant males punish helpers Clutton-Brock TH, Parker GA. 1995. Punishment in animal societies. for temporary defection in superb fairy-wrens. Anim Behav. Nature. 373:209–216. 45:830–833. Dierkes P, Heg D, Taborsky M, Skubic E, Achmann R. 2005. Genetic re- Naef J, Taborsky M. 2020a. Commodity-specific punishment for experi- latedness in groups is sex-specific and declines with age of helpers in a mentally induced defection in cooperatively breeding fish. R Soc Open cooperatively breeding cichlid. Ecol Lett. 8:968–975. Sci. 7:191808. Dierkes P, Taborsky M, Kohler U. 1999. Reproductive parasitism of Naef J, Taborsky M. 2020b. Punishment controls helper defence against broodcare helpers in a cooperatively breeding fish. Behav Ecol. egg predators but not fish predators in cooperatively breeding cichlids. 10:510–515. Anim Behav. 168:137–147. Emlen ST, Wrege PH. 1988. The role of kinship in helping decisions Phillips T. 2018. The concepts of asymmetric and symmetric power can among white-fronted bee-eaters. Behav Ecol Sociobiol. 23:305–315. help resolve the puzzle of altruistic and cooperative behaviour. Biol Rev. Fischer S, Zöttl M, Groenewoud F, Taborsky B. 2014. Group-size- 93:457–468. dependent punishment of idle subordinates in a cooperative breeder Press WH, Dyson FJ. 2012. Iterated Prisoner’s Dilemma contains strategies where helpers pay to stay. Proc R Soc B Biol Sci. 281:20140184. that dominate any evolutionary opponent. Proc Natl Acad Sci USA. García-Ruiz I, Quiñones A, Taborsky M. 2022. The evolution of cooper- 109:10409–10413. ative breeding by direct and indirect fitness effects. Sci Adv. 8:eabl7853. Quinones AE, Van Doorn GS, Pen I, Weissing FJ, Taborsky M. Gardner A, West SA. 2004. Cooperation and punishment, especially in hu- 2016. Negotiation and appeasement can be more effective drivers mans. Am Nat. 164:753–764. of sociality than kin selection. Philos Trans R Soc B Biol Sci. Gaston AJ. 1978. The evolution of group territorial behavior and coopera- 371:20150089. tive breeding. Am Nat. 112:1091–1100. R Core Team. 2022. R: a language and environment for statistical com- Grantner A, Taborsky M. 1998. The metabolic rates associated with resting, puting. Vienna: R Foundation for Statistical Computing. and with the performance of agonistic, submissive and digging behav- Raihani NJ, Thornton A, Bshary R. 2012. Punishment and cooperation in iours in the cichlid fish Neolamprologus pulcher (Pisces: Cichlidae). J Comp nature. Trends Ecol Evol. 27:288–295. Physiol B Biochem Syst Environ Physiol. 168:427–433. Reeve HK. 2000. A transactional theory of within-group conflict. Am Nat. Grinsted L, Field J. 2017. Market forces influence helping behaviour in co- 155:365–382. operatively breeding paper wasps. Nat Commun. 8:1–8. Reeve HK, Gamboa GJ. 1987. Queen regulation of worker foraging Groenewoud F, Frommen JG, Josi D, Tanaka H, Jungwirth A, Taborsky in paper wasps: a social feedback control system (Polistes fuscatus, M. 2016. Predation risk drives social complexity in cooperative breeders. Hymenoptera: Vespidae). Behaviour. 102:147–167. Proc Natl Acad Sci USA. 113:4104–4109. Rubinstein A. 1982. Perfect equilibrium in a bargaining model. Econom J Hamilton IM, Taborsky M. 2005. Unrelated helpers will not fully compen- Econom Soc. 50:97–109. sate for costs imposed on breeders when they pay to stay. Proc R Soc B Russell E, Rowley I. 1988. Helper contributions to reproductive success Biol Sci. 272:445–454. in the splendid fairy-wren (Malurus splendens). Behav Ecol Sociobiol. Hammerstein P. 1981. The role of asymmetries in animal contests. Anim 22:131–140. Behav. 29:193–205. Stiver KA, Dierkes P, Taborsky M, Gibbs HL, Balshine S. 2005. Relatedness Heg D, Bachar Z, Brouwer L, Taborsky M. 2004. Predation risk is an eco- and helping in fish: examining the theoretical predictions. Proc R Soc B logical constraint for helper dispersal in a cooperatively breeding cichlid. Biol Sci. 272:1593–1599. Proc R Soc Lond Ser B Biol Sci. 271:2367–2374. Stiver KA, Fitzpatrick J, Desjardins JK, Balshine S. 2006. Sex differences Heg D, Taborsky M. 2010. Helper Response to Experimentally Manipulated in rates of territory joining and inheritance in a cooperatively breeding Predation Risk in the Cooperatively Breeding Cichlid Neolamprologus cichlid fish. Anim Behav. 71:449–456. pulcher. PloS One 5:e10784. Taborsky B, Skubic E, Bruintjes R. 2007. Mothers adjust egg size to Hellmann JK, Hamilton IM. 2018. Dominant and subordinate outside op- helper number in a cooperatively breeding cichlid. Behav Ecol. tions alter help and eviction in a pay-to-stay negotiation model. Behav 18:652–657. Ecol. 29:553–562. Taborsky M. 1984. Broodcare helpers in the cichlid fish Lamprologus brichardi Hellmann JK, Ligocki IY, O’connor CM, Reddon AR, Garvy KA, Marsh- - their costs and benefits. Anim Behav. 32:1236–1252. Rollo SE, Gibbs HL, Balshine S, Hamilton IM. 2015. Reproductive Taborsky M. 1985. Breeder-helper conflict in a cichlid fish with broodcare sharing in relation to group and colony-level attributes in a cooperative helpers - an experimental analysis. Behaviour. 95:45–75. breeding fish. Proc R Soc B Biol Sci. 282:20150954. Taborsky M. 2016. Cichlid fishes: a model for the integrative study of social Jungwirth A, Walker J, Taborsky M. 2015. Prospecting precedes dispersal behavior. Cooperative breeding in vertebrates. Cambridge: Cambridge and increases survival chances in cooperatively breeding cichlids. Anim University Press, p 272–293. Behav. 106:107–114. Taborsky M, Cant MA, Komdeur J. 2021. The evolution of social behav- Kingma SA, Hall ML, Peters A. 2011. Multiple benefits drive helping be- iour. Cambridge: Cambridge University Press. havior in a cooperatively breeding bird: an integrated analysis. Am Nat. Taborsky M, Grantner A. 1998. Behavioural time-energy budgets of co- 177:486–495. operatively breeding Neolamprologus pulcher (Pisces: Cichlidae). Anim Kingma SA, Santema P, Taborsky M, Komdeur J. 2014. Group augmen- Behav. 56:1375–1382. tation and the evolution of cooperation. Trends Ecol Evol. 29:476–484. Taborsky M, Limberger D. 1980. Lamprologus brichardi, a cichlid fish with fac- Kokko H, Johnstone RA, Wright J. 2002. The evolution of parental and ultative polygyny and helpers at brood care and territory maintenance. alloparental effort in cooperatively breeding groups: when should helpers Newsl IAFE. 3:63–64. pay to stay? Behav Ecol. 13:291–300. Thompson FJ, Marshall HH, Sanderson JL, Vitikainen EI, Nichols HJ, Leadbeater E, Carruthers JM, Green JP, Rosser NS, Field J. 2011. Nest in- Gilchrist JS, Young AJ, Hodge SJ, Cant MA. 2016. Reproductive com- heritance is the missing source of direct fitness in a primitively eusocial petition triggers mass eviction in cooperative banded mongooses. Proc R insect. Science. 333:874–876. Soc B Biol Sci. 283:20152607. 372 Behavioral Ecology Webster MS, Tarvin KA, Tuttle EM, Pruett-Jones S. 2004. Reproductive Zöttl M, Frommen JG, Taborsky M. 2013c. Group size adjustment to promiscuity in the splendid fairy-wren: effects of group size and auxiliary ecological demand in a cooperative breeder. Proc R Soc B Biol Sci. reproduction. Behav Ecol. 15:907–915. 280:20122772. Zöttl M, Chapuis L, Freiburghaus M, Taborsky M. 2013a. Strategic re- Zöttl M, Heg D, Chervet N, Taborsky M. 2013d. Kinship lowers duction of help before dispersal in a cooperative breeder. Biol Lett. alloparental care where helpers pay-to-stay. Nat Commun. 9:20120878. 4:1341. Zöttl M, Fischer S, Taborsky M. 2013b. Partial brood care compensation by Zöttl M, Schreier T, Taborsky M. 2022. Data for: Coercion promotes female breeders in response to experimental manipulation of alloparental alloparental care in cooperative breeders. Behav Ecol. doi: 10.6084/ care. Anim Behav. 85:1471–1478. m9.figshare.19984580.v1. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology Oxford University Press

Coercion promotes alloparental care in cooperative breeders

Behavioral Ecology , Volume 34 (3): 10 – Feb 28, 2023

Loading next page...
 
/lp/oxford-university-press/coercion-promotes-alloparental-care-in-cooperative-breeders-sE5hMysr5n

References (71)

Publisher
Oxford University Press
Copyright
© The Author(s) 2023. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology.
ISSN
1045-2249
eISSN
1465-7279
DOI
10.1093/beheco/arac125
Publisher site
See Article on Publisher Site

Abstract

The official journal of the Behavioral ISBE Ecology International Society for Behavioral Ecology Behavioral Ecology (2023), 34(3), 363–372. https://doi.org/10.1093/beheco/arac125 Original Article Coercion promotes alloparental care in cooperative breeders a,b, a a,c,d, Markus Zöttl, Tanja Schreier, and Michael Taborsky Division of Behavioural Ecology, Institute of Ecology and Evolution, University of Bern, CH-3032 Hinterkappelen, Switzerland, Department of Biology and Environmental Science, Linnaeus University, c d Kalmar, Sweden, Max Planck Institute of Animal Behavior, D-78467 Konstanz, Germany, and Institute for Advanced Study (Wissenschaftskolleg) Berlin, D-14193 Berlin, Germany Received 3 June 2022; revised 2 December 2022; editorial decision 6 December 2022; accepted 8 December 2022 Members of social groups may negotiate among each other about the exchange of goods and services. If this involves asymmet- ries between interacting partners, for instance in condition, power, or expected payoffs, coercion may be involved in the bargain. Cooperative breeders are excellent models to study such interactions, because asymmetries are inherent in the relationship between dominant breeders and subordinate helpers. Currently it is unclear whether punishment is used to enforce costly cooperation in such systems. Here we investigated experimentally in the cooperatively breeding cichlid Neolamprologus pulcher whether alloparental brood care provided by subordinates is contingent on enforcement by dominant breeders. We manipulated first the brood care beha- vior of a subordinate group member and then the possibility of the dominant breeders to punish idle helpers. When subordinates were prevented from providing brood care, breeders increased their attacks on them, which triggered increased alloparental brood care by helpers as soon as this was again possible. In contrast, when the possibility to punish helpers was prevented, energetically costly alloparental brood care did not increase. Our results confirm predictions of the pay-to-stay mechanism causing alloparental care in this species and they suggest more generally that coercion can play an important role in the control of cooperation. Key words: coercion, punishment, negotiation, cooperative breeding, pay-to-stay, Neolamprologus pulcher. INTRODUCTION direct fitness benefits from their group membership despite being reproductively suppressed (Taborsky 1984; Balshine-Earn et al. Social interactions are typically characterized by a conflict of 1998; Kingma et al. 2011; Leadbeater et al. 2011). They enjoy ac- interest between agents differing in power (Hammerstein 1981; cess to the cooperatively defended territory with resources such as Rubinstein 1982; Binmore 2010; Phillips 2018; Taborsky et al. food and shelters and may benefit from enhanced safety from pred- 2021). This asymmetry affects negotiations between social part- ators (Heg et al. 2004; Groenewoud et al. 2016). Later in life, they ners about access to resources, group membership, or other ad- may inherit the breeding position in their natal group and benefit vantages (McNamara et al. 1999; Melis et al. 2009). Negotiation from the presence of helpers that previously they had helped to processes, the interaction by which individuals adjust their beha- raise by alloparental care (Stiver et al. 2006, Kingma et al. 2011, vior to their partner’s behavior and signaling to achieve a ben- 2014, Leadbeater et al. 2011). However, the presence of subordi- eficial outcome, may encompass combinations of reciprocal nate group members can impose costs to dominants through re- exchanges and elements of coercion, such as punishment for source competition and by enhancing intra-sexual competition for cheating or withheld cooperation (Clutton-Brock and Parker reproduction (Webster et al. 2004; Mitchell et al. 2009a, 2009b; 1995; Raihani et al. 2012; Quinones et al. 2016; Taborsky et al. Hellmann et al. 2015). If these costs outweigh the benefits, the 2021). Currently, it is unclear to which extent coercion and en- dominant individuals will not tolerate subordinates as group mem- forcement affect mutual relationships and reciprocal exchanges in bers and eventually evict them from their territory (Dierkes et al. highly social animals. 1999; Taborsky 1985, 2016; Zöttl et al. 2013c; Thompson et al. In cooperative breeders, dominant individuals benefit from 2016). alloparental brood care performed by subordinate group members Theoretical models suggest that subordinates may provide help (Russell and Rowley 1988; Emlen and Wrege 1988; Brouwer et al. to compensate for the costs imposed to dominants, thereby re- 2005; Zöttl et al. 2013b). In return, the latter often gain substantial moving the incentive for forcible eviction (Gaston 1978; Kokko et al. 2002; Hamilton and Taborsky 2005). Such services provided Address correspondence to M. Zöttl. E-mail: markus.zottl@lnu.se. © The Author(s) 2023. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 364 Behavioral Ecology in exchange of tolerance in the group have been referred to as METHODS “pay-to-stay” (PTS), which can lead to evolutionarily stable group Experimental subjects and group formation living and cooperation (Quinones et al. 2016). This mechanism Experimental subjects were taken from the laboratory stock popu- may involve a negotiation process and is particularly important lation originating from wild N. pulcher caught in Lake Tanganyika when relatedness between dominants and subordinates is low, the near Mpulungu, Zambia, in the years 1999, 2006, and 2009. The presence of subordinates inflicts costs on dominants, dispersal experimental groups were assembled following a standard proce- and independent breeding by subordinates is constrained, and dure and consisted of a male, a female, and a subordinate helper individuals are generally long-lived. The crucial difference from of unknown sex (Taborsky 1984). Resembling the natural group other mechanisms selecting for cooperative brood care, such as composition, males were the largest individuals in the group kin selection and benefits from group augmentation (García-Ruiz (standard length (SL), mean ± SD: 63.8 mm ± 8.3) followed by fe- et al. 2022), is that subordinates do not provide help voluntarily males (55.8 mm ± 5.5) and helpers (26.6 mm ± 3.0). The minimal but are forced by dominants to do so (Fischer et al. 2014; Naef size difference between the male and female pair members was and Taborsky 2020a, 2020b). PTS predicts that when subordin- 0.5  mm, whereas the female was at least 10  mm larger than the ates fail to produce sufficient benefits to the dominants they are respective helper. Each group was kept in a 100  L compartment punished and eventually evicted from the territory (Gaston 1978; of a subdivided 200 L aquarium (N = 17). All compartments con- Mulder and Langmore 1993). However, the predicted causal ef- tained a semi-transparent tube suspended below the water surface fect of punishment on increased helping behavior has hitherto serving as a retreat shelter, an air-driven biological filter suspended not been demonstrated in non-human animals (Raihani et al. in the water, and 4 flowerpot-halves serving as breeding shelters 2012). (see Figure 1). The bottom of the tank was covered with a mix- In the African cichlid fish Neolamplologus pulcher several key aspects ture of gravel and sand. The water temperature was held constant of the mechanism regulating the asymmetric negotiation between at 26–28°C and the light regime was set at a 13 h-light:11 h-dark dominants and subordinates that characterizes the pay-to-stay re- cycle, simulating natural conditions in Lake Tanganyika. The fish lationship have been experimentally demonstrated (Taborsky were fed six times a week with dried food and 2–4 times weekly 1985; Balshine-Earn et al. 1998; Bergmüller and Taborsky 2005; with additional krill. The experiments were conducted at the Bergmuller et al. 2005; Heg and Taborsky 2010; Zöttl et al. 2013a, Ethologische Station Hasli of the University of Bern, Switzerland 2013d; Fischer et al. 2014; Naef and Taborsky 2020a, 2020b). For between June and October 2012 under the ethical approval license instance, individuals prevented from defending or subordinates that of Veterinaeramt Bern 16/09. are temporarily removed from the group are subsequently attacked In the first days after group formation, the fish were habituated by more dominant helpers and breeders (Balshine-Earn et al. to a mesh cage (mesh size 10  ×  10  mm) that later in the experi- 1998; Fischer et al. 2014; Naef and Taborsky 2020a). Helpers pre- mental procedure was used to prevent breeders from entering the vented from defending the territory against intruders subsequently breeding shelter. Helpers learned that they could enter a shelter de- show increased levels of defense (Bergmuller et al. 2005; Naef and spite the mesh net put over it, whereas the breeders were unable to Taborsky 2020b), and a similar pattern emerged when temporarily pass the mesh. After this initial habituation, families were checked preventing helpers from providing alloparental care (Zöttl et al. for clutches every day. If eggs were found, the experiments took 2013b). Subsequent to the experimental manipulation of defense, place on the first and second day after spawning. helpers increased either cooperative defense or submissive beha- The helpers habituated quickly to the introduced mesh cage. All vior, suggesting that cooperative defense serves an appeasement helpers passed the mesh cage within a few days and our observa- function (Bergmuller and Taborsky, 2005). Similar results were tions confirmed that all helpers had learnt the task before the ex- obtained when helpers were experimentally prevented from dig- periment started. During the time between training and breeding ging out the territory shelter (Naef and Taborsky 2020a). In ad- the groups were exposed to the mesh net cage intermittently to en- dition, subordinates are only accepted in the territory when help sure that they remained habituated to the experimental procedures. is needed (Taborsky 1985; Zöttl et al. 2013c). However, all these experiments manipulated the behavior or presence of subordinate Experimental design and procedure helpers rather than the behavior of dominant breeders, so that the potentially underlying negotiation process between both parties is We started the experiment when the group produced their first still poorly understood. clutch (Figure 1). At that point the fish were habituated to the ex- Here we tested experimentally whether breeders force non- perimental setup and the subordinates had learned that they could breeders to provide alloparental care by direct physical attacks and pass freely through the mesh net temporarily covering the breeding very close proximity, which may signal the threat of imminent ag- shelter (Supplementary Figure 1 and Supplementary Table 1). After gression in N. pulcher. First, we created a situation in which helpers a baseline behavior recording (“initial phase”), the group was ex- either showed normal levels of alloparental brood care or were pre- posed to one of two treatments in the experience phase. In the vented to care for the dominant breeders’ clutch. Then we either experience treatment “help prevented” the whole group was pre- prevented physical social interaction between breeders and helpers vented from caring for the brood for 30  min by placing a close- to inhibit the possibility of punishment, or we allowed unrestricted meshed cage over the breeding shelter so that no fish could pass. physical contact. We predicted that if physical social interaction Previous experiments have shown that individuals react to time between breeders and helpers is required to enforce help, more periods where they were unable to conduct brood care by increasing help should be provided in the treatment where social interaction their brood care investment subsequently (Zottl et al 2013d). In the is possible. In contrast, if the mere presence of dominant breeders experience treatment “help allowed” a broad-meshed cage was put suffices to induce help, increased levels of help should be observed over the breeding shelter for 30 min that allowed the helper to pass, after experimentally imposed idleness, regardless of whether phys- but the breeders were excluded from entering the breeding shelter ical interactions were possible of not. to provide egg care. Zöttl et al. · Coercion promotes alloparental care 365 Initial phase Experience phase Test phase Manipulation of possibility to Behavioural observations of brood care conduct brood care Physical contact Help allowed: No need prevented for compensation Breeder male Physical contact Subordinate allowed Breeder female helper Helper can conduct brood care. Physical contact prevented Help prevented: Need for compensation from helper Breeding shelter Flower pot halves as containing eggs shelters Physical contact allowed All individuals are excluded from conducting brood care. Figure 1 Sketch of the experimental set-up. The experiment started one or two days after the dominant female had spawned for the first time. In the “Initial phase”, fish behavior was recorded without experimental manipulation. In the experience phase, for 30 min either all group members were prevented from entering the breeding shelter (impassable mesh cage put over the breeding shelter; i.e., experience treatment “help prevented”), or the helper could access the breeding shelter to care for the eggs (mesh cage was passable for the helper; i.e., experience treatment “help allowed”). In the test phase, for 20  min the treatments either allowed physical contact between the breeders and helpers enabling punishment by physical attacks (ramming or biting; i.e., experimental treatment “physical contact allowed”), or the breeders were confined in clear perspex cylinders preventing bodily contact (i.e., experimental treatment “physical contact prevented”). All groups underwent all four possible combinations of experience and test phases in random order yielding a within subject, full factorial experimental design. In the test phase immediately following the experience phase, contact allowed” condition the cylinders were inserted in the same each group was assigned to a treatment either preventing social way, but the breeders remained outside and were unconstrained interaction between breeders and helpers (“physical contact pre- in their ability to physically attack helpers (Figure 1). All experi- vented”) or allowing social interaction (“physical contact allowed”, mental groups underwent all treatments in the experience and test see Figure 1). In both situations, the broad-meshed cage was put phases in random order, so that each group was tested four times. over the breeding shelter which allowed helpers to care for the eggs The minimum time between two consecutive trials was 1 hour. In whereas the breeders were excluded from entering the shelter. In total, 17 groups were used in the experiment and a total of 68 trials the experimental treatment “physical contact prevented” physical of the test phase were conducted. In 11 of these 17 groups, the interactions were prevented by restricting the breeders to an area experimental trials were conducted on two consecutive days, with in the tank using transparent perspex cylinders. Hence, they re- two trials per day separated by at least an interval of one hour. In mained visually present and could interact with the helper, but they 6 groups we conducted all 4 trials on the same day, again with one were unable to exert any overt physical aggression. In the “physical hour minimum time between consecutive trials. 366 Behavioral Ecology To answer the question whether breeders increased aggression Brood care and behavioral observations towards the subordinate and whether the subordinate increased In the test phase, we conducted continuous focal observation of submission to the breeders in response to preventing the possibility 15  min on the helper individual (Altmann 1974) focusing on two to provide brood care we used the behavioral data collected in the alloparental care behaviors that have contrasting energetic im- experience phase and specified the frequency of overt aggression, plications. We recorded the time individuals engaged in fanning. This behavior serves to create a water current around the clutch and increases the supply of oxygen for the clutch. It is energet- 25 (a) ically costly because it involves a forward swimming movement Help treatment * contact treatment: p = 0.001 that is compensated by intense beats with the pectoral fins to keep the fish on the spot (Grantner and Taborsky 1998; Taborsky and Grantner 1998). Second, we quantified egg cleaning by counting the frequency of micro nips towards the eggs, which serves a hygi- enic function by removing fungi and other microorganisms. This brood care behavior probably involves less energy expenditure (Taborsky and Grantner 1998). Additionally, in each phase of the experiment, we recorded social interactions between breeders and helpers during the 15 min focal observations. We present frequen- cies of overt physical aggression from both breeders towards the helper (ramming and biting), frequencies of restrained aggres- sion (threat signals without physical contact), and the frequencies of submissive behavior (tail quivering) of the helper towards the (b) breeders. We did not record other submissive behaviors (submis- sive postures and hook swimming) because they are usually shown less frequently and it is unclear whether they involve energetic costs comparable to tail quivering. Statistical analyses To model fanning duration, we fitted the log transformed duration of fanning behavior as response variable and added the experi- mental condition of the experience phase (factor with two levels: help prevented/allowed) and the current experimental treatment (factor with two levels: physical interactions prevented/allowed) as fixed factors. We included the interaction between these fixed factors and specified a random effect that identified the experi- Prevented contact Contact Prevented contact Contact mental group. We modeled the data specifying a Gaussian distri- Help allowed Help prevented bution using a Linear Mixed Effect Model (LMM). To model egg cleaning behavior we fitted the number of egg cleaning events as Figure 2 the response variable and again added the experimental condition Fanning duration (A) and egg cleaning frequency (B) of helpers in the test of the experience phase and the current experimental treatment as phase, as measures of alloparental care in response to the experimental fixed factors. We included the interaction between the fixed factors treatments. The boxes depict the median and the interquartile ranges, and and the identity of the test group as random effect. Because the the whiskers show the range of the raw data from N = 17 observations egg cleaning count data showed signs of overdispersion we mod- in each condition. The dots represent data points of each replicate. Stars eled the data specifying a negative binomial distribution using a indicate significance at the level of p < 0.05. Full model details are shown in Table 1. Generalized Linear Mixed Effect Model (GLMM). Table 1 Brood care behavior of the helper in the test phase. Shown are the estimates and incidence rate ratios, 95% confidence intervals, and p-values of an LMM and a GLMM modeling the fanning duration and the frequency of egg cleaning during 15 min observations, respectively. As predictors the model included the treatments of the experience and test phases, and their interaction. Both models included the group identity as random effect (N = 17 groups, N = 68 observations). Significant p -values are printed in bold. Fanning Egg cleaning Predictors Estimates CI p Incidence Rate Ratios CI p (Intercept) 1.02 0.52–1.53 <0.001 44.89 30.36–66.37 <0.001 Treatment experience phase 0.25 −0.30 to 0.80 0.367 1.7 1.18–2.46 0.005 Treatment test phase -0.39 -0.94–0.16 0.166 1.52 1.05–2.20 0.027 Experience * test phase 1.33 0.55–2.11 0.001 1 0.63–1.61 0.991 2 2 Marginal R /Conditional R 0.234/0.548 0.188/0.625 Frequency of egg cleaning Fanning duration (s) Zöttl et al. · Coercion promotes alloparental care 367 Breeder male to helper Breeder female to helper Initial phase Experience phaseTest phase Initial phase Experience phase Test phase 15 (a) 15 (b) ** 10 10 5 5 0 0 (c) (d) 30 30 * * 20 20 10 10 0 0 Figure 3 Frequency of overt aggression (A & B) and restrained aggression (C, D) against the helper from the male (A & C) and the female (B & D) breeder during the different phases and treatments. The boxes depict the median and the interquartile ranges, and the whiskers show the range of the raw data from N = 17 observations in each condition. Stars indicate significance at the level of p < 0.05. Full model details are shown in Tables 2 and 3. restrained aggression and of submission as response variable in sep- they had been prevented from conducting alloparental care in arate models. We fitted the treatment as factor with 2 levels (help the experience phase and physical contact between breeders and allowed/ help prevented) and the group identity as random factor. helpers was allowed in the test phase (Table 1, Figure 2A). We also modeled the distribution of aggressive and submissive be- Helper egg cleaning in the test phase varied in dependence of havior longitudinally throughout the experiment when the groups both the help prevented in the experience phase and the manipula- moved from the initial phase to the experience phase (help pre- tion of physical contact in the test phase (Table 1, Figure 2B). Help vented) and then to the test phase (contact allowed) by fitting the prevented during the experience phase triggered increased levels of phase as factor with 3 levels and group identity as random effect. egg cleaning behavior of helpers during the test phase. The same To conduct all pairwise comparisons between the 3 phases we sub- effect was produced by the treatment in which dominants could sequently used a Tukey post-hoc analysis with adjusted significance exert overt aggression on the subordinates during the test phase level for multiple comparisons. Here too, we used a negative bino- (Figure 2B, Table 1B). In contrast to fanning behavior, the inter- mial distribution to account for overdispersion of the data. action between these two factors was not significant suggesting that All analyses were conducted in R Version 4.2 (R Core Team this form of energetically cheap alloparental brood care did not 2022) using the packages lme4 and glmmTMB (Bates et al. 2011; only depend on overt aggression from breeders (Table 1B). Brooks et al. 2017). All figures were produced from raw, untrans- Breeder aggression against the helper formed data. Overt attacks (including physical contact) depended on the treat- ment and the phase of the experiment. During the experience RESULTS phase treatment “help prevented”, breeders attacked the helper Alloparental care more often than during the treatment “help allowed” (Figure 3A and B, Table 2A). The longitudinal comparison between the three The subordinates provided different amounts of alloparental brood different phases also suggested an effect of the experimental manip- care depending on the behavioral manipulation treatment. The ulation “help prevented” during the experience phase on breeder helpers showed more fanning than in all other treatments when Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Frequency overt aggression Frequency restrained aggression 368 Behavioral Ecology Helper submission to breeders aggression: overt attacks increased from the initial phase to the ex- Initial phase Experience phase Test phase perience phase (when helping was prevented) and decreased there- * * after in the test phase (when helpers resumed brood care, Figure 3A and B, Table 2B). This pattern was the same for both the male and female breeders, with the exception that female breeders did not show a significant decrease of aggression between experience and test phases (Figure 3A and B; Table 2B). Restrained aggression (threat signals) was not equally affected by the treatment, but also changed throughout the phases of the experiment. During the experience phase treatment “help pre- vented”, breeders showed restrained aggression toward the helper at similar rates as when the helper had the opportunity to con- duct brood care (Figure 3C and D, Table 3A). The longitudinal comparison between the three different phases again showed an effect of the manipulation “help prevented” during the experi- ence phase on breeder restrained aggression. Restrained aggres- sion increased from the initial phase to the experience phase in males when helping was prevented (Figure 3C, Table 3B). In fe- males the increase in restrained aggression was significant when the initial phase was compared with the test phase (Figure 3D; Figure 4 Table 3B). Frequency of tail quivering exhibited by the helper toward the male and female breeders during the different phases and treatments of the Helper submission towards breeders experiment. The boxes depict the median and the interquartile ranges, and the whiskers show the range of the raw data from N = 17 replicates in each Helpers showed more tail quivering towards the breeders in the ex- condition. Stars indicate significance at the level of p < 0.05. Full model perience phase treatment “help prevented” than when they were details are shown in Table 4. Table 2 Overt aggressive behavior of the male and female breeders in A) the experience phase and B) longitudinally throughout the experimental phases when help was prevented during the experience phase and contact was allowed during the test phase. Shown are the incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression during 15 min observations. As predictors the model in A) included the treatment of the experience phase and B) the three different stages of the experiment. Both models included the group identity as random effect (N = 17 groups and N = 68 or 51 observations, respectively). Significant p -values are printed in bold A) Female overt aggression Predictors Male overt aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI (Intercept) 0.66 0.31 – 1.38 0.271 1.59 0.85 – 2.98 0.147 Help prevented 2.88 1.51 – 5.50 0.001 1.93 1.16 – 3.20 0.011 2 2 Marginal R /Conditional R 0.154/ 0.433 0.068/ 0.438 B) Male overt aggression Female overt aggression Predictors Incidence Rate Ratios CI p Incidence Rate Ratios CI p (Intercept) 0.14 0.04 – 0.50 0.002 0.77 0.28 – 2.10 0.611 Experience Phase 7.17 3.05 – 16.84 <0.001 4.1 1.55 – 10.90 0.005 Test Phase 0.83 0.25 – 2.73 0.763 2.37 0.83 – 6.72 0.106 Tukey pairwise comparisons (adjusted for multiple comparisons) Estimate Standard Error p Estimate Standard Error p Initial vs Experience -1.969 0.436 <0.001 -1.412 0.498 0.018 Initial vs Test 0.182 0.606 0.951 -0.861 0.533 0.249 Experience vs Test 2.152 0.472 <0.001 0.551 0.385 0.335 2 2 Marginal R /Conditional R 0.220/0.728 0.191/0.369 Control Help allowed Help prevented Prevented contact Contact Prevented contact Contact Frequency of tail quiver Zöttl et al. · Coercion promotes alloparental care 369 Table 3 Restrained aggressive behavior of the male and female breeders in A) the experience phase and B) longitudinally throughout the experimental phases when help was prevented during the experience phase and contact was allowed during the test phase. Shown are the incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression during 15 min observations. As predictors the model in A) included the treatment of the experience phase and B) the three different stages of the experiment. Both models included the group identity as random effect (N = 17 groups and N = 68 or 51 observations, respectively). Significant p -values are printed in bold A) Predictors Male restrained aggression Female restrained aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI p Intercept 6.63 4.13–10.65 <0.001 13.61 9.08–20.40 <0.001 Help prevented 0.83 0.51–1.37 0.476 0.93 0.54–1.62 0.808 2 2 Marginal R /Conditional R 0.006/0.323 0.001/0.092 B) Predictors Male restrained aggression Female restrained aggression Incidence Rate Ratios CI p Incidence Rate Ratios CI p (Intercept) 2.3 1.08–4.89 0.031 5.71 3.02–10.81 <0.001 Experience Phase 2.27 1.18–4.39 0.015 1.64 0.86–3.12 0.135 Test Phase 1.94 0.92–4.10 0.083 2.39 1.19–4.79 0.015 Tukey pairwise comparisons (adjusted for multiple comparisons) Estimate Standard Error p Estimate Standard Error p Initial vs Experience −0.82 0.34 0.048 −0.5 0.33 0.3 Initial vs Test −0.66 0.38 0.21 −0.87 0.35 0.048 Experience vs Test 0.16 0.32 0.87 −0.38 0.38 0.5 2 2 Marginal R /Conditional R 0.090/0.426 0.124/0.219 allowed to care for the brood (Figure 4, Table 3A). The longitu- that dominant breeders punish idle subordinates, which should dinal comparison revealed that the submission rate of the helpers in turn enhance their helping effort. We manipulated both the increased from the initial phase to the experience phase (when helpful behavior of subordinates and the respective response of helping was prevented) and decreased thereafter significantly in the dominants, and we measured the behaviors resulting from the test phase (when helpers reassumed brood care; Figure 4, Table these manipulations. Our data reveal that indeed, idle subordi- 2B). nate group members were punished by enhanced overt attacks from dominant breeders, which in turn caused the affected sub- ordinate to increase its helping effort. In contrast, when pun- DISCUSSION ishment was experimentally prevented, helpers did not increase There are many indications in a wide range of taxa that energetically costly fanning behavior. This confirms the predic- interacting agents negotiate among each other about the ex- tions of punishment and its functionality, which is the basis of change of goods and services (reviewed in Taborsky et al. 2021). negotiations in a pay-to-stay relationship, thereby confirming Pertinent theoretical studies have outlined the conditions and the causality between social enforcement and helping behavior. functionality of such trade and suggest that when social partners Such feedback control and social regulation of cooperation has negotiate among each other, power asymmetry among them may been suggested to operate also in some other social species, for lead to punishment and enforcement of cooperative behavior example in paper wasps (Polistes fuscatus; [Reeve and Gamboa (Rubinstein 1982; Clutton-Brock and Parker 1995; McNamara et 1987]), and we suggest that this may be more widespread than al. 1999; Gardner and West 2004; McNamara and Leimar 2010; currently assumed. Press and Dyson 2012; Hellmann and Hamilton; 2018). However, In asymmetric relationships, the answer to the question how the prediction that aggressive behavior can induce cooperation much agents that are more powerful than their social partners in idle subordinates has rarely been tested. A critical test of this can demand from them depends on the alternatives, or “out- hypothesis must involve experimental manipulations of the re- side options”, of the latter (Reeve 2000; Cant et al. 2012; Cant sponses of both partners to each other’s actions. This was the aim and Johnstone 2009). In N. pulcher, the outside options for sub- of our study. ordinate group members are typically poor because of the high Based on the hypothesis that social partners with greater mortality risk when leaving a group (Taborsky and Limberger power can enforce service by less powerful partners we predicted 1980; Taborsky 1984; Heg et al. 2004; Jungwirth et al. 2015). 370 Behavioral Ecology Table 4 Zöttl et al. 2013b). Hence, dominants should not raise the price too Tail quivering of helper displayed toward the male and female high for the tolerance of subordinate group members, but this should breeders as measure of submission in A) the experience phase depend on the local conditions determining the outside options of and B) longitudinally throughout the experimental phases when subordinates. The remarkable variation of group composition and help was prevented and contact was allowed. Shown are the group size between populations exposed to different levels of preda- incidence rate ratios, 95% confidence intervals, and p -values of GLMMs modeling the frequency of male and female aggression tion risk supports this conjecture: in populations with enhanced pre- during 15 min observations. As predictors, the model in A) dation risk the groups contained comparatively more large helpers, included the treatment of the experience phase and in B) the which may suggest that the breeders can demand more help from three different stages of the experiment. Both models included their subordinates when predation pressure is high (Groenewoud the group identity as random effect (N = 17 groups and N = et al. 2016). However, our study worked with standardized groups 68 or 51 observations, respectively). Significant p -values are printed in bold of three individuals and it is currently unclear how group size and composition would affect the results of this experiment. Helper submission In contrast to the evidence for PTS, kin selection is apparently of minor importance concerning the relationship among group Incidence p members in this species, as large helpers are often unrelated to the Predictors Rate Ratios CI dominant breeders (Dierkes et al. 2005; Stiver et al.; 2005). In a A) laboratory experiment, unrelated helpers exhibited higher levels of (Intercept) 15.72 11.50–21.49 <0.001 alloparental brood care than related helpers, confirming an impor - Help prevented 1.67 1.31–2.11 <0.001 2 tant prediction of PTS and refuting kin selection as an alternative Marginal R  / 0.119/0.572 explanation (Zöttl et al. 2013d). Furthermore, only unrelated indi- Conditional R B) viduals increased alloparental care when the costs of their presence (Intercept) 14.85 10.46–21.08 <0.001 in the territory were experimentally increased (Zöttl et al. 2013d). Experience Phase 1.84 1.32–2.58 <0.001 In conclusion, our study demonstrates the causal relationship Test Phase 0.8 0.53–1.20 0.277 between punishment for withheld help and subsequent compen- sation by enhanced cooperation, which reflects an important part Tukey pairwise comparisons (adjusted for multiple of the negotiation process among individuals in asymmetric social comparisons) relationships. Estimate Standard Error p SUPPLEMENTARY MATERIAL Initial vs Experience −0.61 0.171 0.0024 Initial vs Test 0.224 0.206 0.5266 Supplementary material can be found at http://www.beheco. Experience vs Test 0.834 0.186 <0.001 oxfordjournals.org/ Observations 51 Marginal R / 0.220/0.543 2 We thank all members of the Ethologische Station Hasli for help during Conditional R the data collection and Stefan Fischer and Arne Jungwirth for helpful discussions. Hence, subordinates should invest as much cooperative effort as FUNDING is required to be tolerated in a safe territory. A previous experi- This work was supported by grants from the Swiss National Science ment investigating the shelter digging effort of helpers, which is Foundation to Michael Taborsky (SNSF grants 310030B_138660, an energetically very expensive behavior (Grantner and Taborsky 31003A_156152, and 31003A 176174). 1998; Taborsky and Grantner 1998), suggested that this is in- deed the case. When large, reproductively mature helpers were DATA AVAILABILITY given the chance to disperse and reproduce elsewhere in a safe Analyses reported in this article can be reproduced using the data provided environment, they reduced the shelter digging effort in their by Zöttl et al 2022. group (Bergmuller et al. 2005) as predicted by pay-to-stay theory (Kokko et al. 2002). In the field, helpers reduced their cooper - Handling Editor: Amanda Ridley ative antipredator defense effort shortly before dispersing (Zöttl et al. 2013a). Similarly, in paper wasps, Polistes dominula, subordi- REFERENCES nate group members reduced their cooperative foraging effort in Altmann J. 1974. Observational study of behavior: sampling methods. their group when outside options were experimentally enhanced Behavior. 49:227–265. (Grinsted and Field 2017). The extent to which outside options Balshine-Earn S, Neat FC, Reid H, Taborsky M. 1998. Paying to stay or affect the negotiation process between helpers and breeders would paying to breed? Field evidence for direct benefits of helping behavior in be a worthwhile subject for future studies. a cooperatively breeding fish. Behav Ecol. 9:432–438. In cooperatively breeding animals dominants typically also rely Balshine S, Leach B, Neat F, Reid H, Taborsky M, Werner N. 2001. Correlates of group size in a cooperatively breeding cichlid fish on the presence and cooperation of subordinates. In N. pulcher, (Neolamprologus pulcher). Behav Ecol Sociobiol. 50:134–140. depending on the local predation pressure breeders may rarely be Bates D, Maechler M, Bollinge B. 2011. lme4: linear mixed-effects models able to raise young on their own (Brouwer et al. 2005; Groenewoud using S4 classes. R package version 0.999375-42 ed. et al. 2016). In addition, they benefit from the presence of subordi- Bergmuller R, Heg D, Taborsky M. 2005. Helpers in a cooperatively breeding cichlid stay and pay or disperse and breed, depending on eco- nate brood care helpers by load lightening and enhanced produc- logical constraints. Proc R Soc B Biol Sci. 272:325–331. tivity (Taborsky 1984; Balshine et al. 2001; Taborsky et al. 2007; Zöttl et al. · Coercion promotes alloparental care 371 Bergmuller R, Taborsky M. 2005. Experimental manipulation of helping Mcnamara JM, Gasson CE, Houston AI. 1999. Incorporating rules for re- in a cooperative breeder: helpers “pay to stay” by pre-emptive appease- sponding into evolutionary games. Nature. 401:368–371. ment. Anim Behav. 69:19–28. Mcnamara JM, Leimar O. 2010. Variation and the response to variation Binmore K. 2010. Bargaining in biology? J Evol Biol. 23:1351–1363. as a basis for successful cooperation. Philos Trans R Soc B Biol Sci. Brooks ME, Kristensen K, Van Benthem KJ, Magnusson A, Berg CW, 365:2627–2633. Nielsen A, Skaug HJ, Machler M, Bolker BM. 2017. glmmTMB balances Melis AP, Hare B, Tomasello M. 2009. Chimpanzees coordinate in a nego- speed and flexibility among packages for zero-inflated generalized linear tiation game. Evol Hum Behav. 30:381–392. mixed modeling. R J. 9:378–400. Mitchell JS, Jutzeler E, Heg D, Taborsky M. 2009a. Dominant members of Brouwer L, Heg D, Taborsky M. 2005. Experimental evidence for helper cooperatively-breeding groups adjust their behaviour in response to the effects in a cooperatively breeding cichlid. Behav Ecol. 16:667–673. sexes of their subordinates. Behaviour. 146:1665–1686. Cant MA. 2012. Suppression of social conflict and evolutionary transitions Mitchell JS, Jutzeler E, Heg D, Taborsky M. 2009b. Gender differences in to cooperation. Am Nat. 179:293–301. the costs that subordinate group members impose on dominant males in Cant MA, Johnstone RA. 2009. How threats influence the evolutionary res- a cooperative breeder. Ethology. 115:1162–1174. olution of within-group conflict. Am Nat. 173:759–771. Mulder RA, Langmore NE. 1993. Dominant males punish helpers Clutton-Brock TH, Parker GA. 1995. Punishment in animal societies. for temporary defection in superb fairy-wrens. Anim Behav. Nature. 373:209–216. 45:830–833. Dierkes P, Heg D, Taborsky M, Skubic E, Achmann R. 2005. Genetic re- Naef J, Taborsky M. 2020a. Commodity-specific punishment for experi- latedness in groups is sex-specific and declines with age of helpers in a mentally induced defection in cooperatively breeding fish. R Soc Open cooperatively breeding cichlid. Ecol Lett. 8:968–975. Sci. 7:191808. Dierkes P, Taborsky M, Kohler U. 1999. Reproductive parasitism of Naef J, Taborsky M. 2020b. Punishment controls helper defence against broodcare helpers in a cooperatively breeding fish. Behav Ecol. egg predators but not fish predators in cooperatively breeding cichlids. 10:510–515. Anim Behav. 168:137–147. Emlen ST, Wrege PH. 1988. The role of kinship in helping decisions Phillips T. 2018. The concepts of asymmetric and symmetric power can among white-fronted bee-eaters. Behav Ecol Sociobiol. 23:305–315. help resolve the puzzle of altruistic and cooperative behaviour. Biol Rev. Fischer S, Zöttl M, Groenewoud F, Taborsky B. 2014. Group-size- 93:457–468. dependent punishment of idle subordinates in a cooperative breeder Press WH, Dyson FJ. 2012. Iterated Prisoner’s Dilemma contains strategies where helpers pay to stay. Proc R Soc B Biol Sci. 281:20140184. that dominate any evolutionary opponent. Proc Natl Acad Sci USA. García-Ruiz I, Quiñones A, Taborsky M. 2022. The evolution of cooper- 109:10409–10413. ative breeding by direct and indirect fitness effects. Sci Adv. 8:eabl7853. Quinones AE, Van Doorn GS, Pen I, Weissing FJ, Taborsky M. Gardner A, West SA. 2004. Cooperation and punishment, especially in hu- 2016. Negotiation and appeasement can be more effective drivers mans. Am Nat. 164:753–764. of sociality than kin selection. Philos Trans R Soc B Biol Sci. Gaston AJ. 1978. The evolution of group territorial behavior and coopera- 371:20150089. tive breeding. Am Nat. 112:1091–1100. R Core Team. 2022. R: a language and environment for statistical com- Grantner A, Taborsky M. 1998. The metabolic rates associated with resting, puting. Vienna: R Foundation for Statistical Computing. and with the performance of agonistic, submissive and digging behav- Raihani NJ, Thornton A, Bshary R. 2012. Punishment and cooperation in iours in the cichlid fish Neolamprologus pulcher (Pisces: Cichlidae). J Comp nature. Trends Ecol Evol. 27:288–295. Physiol B Biochem Syst Environ Physiol. 168:427–433. Reeve HK. 2000. A transactional theory of within-group conflict. Am Nat. Grinsted L, Field J. 2017. Market forces influence helping behaviour in co- 155:365–382. operatively breeding paper wasps. Nat Commun. 8:1–8. Reeve HK, Gamboa GJ. 1987. Queen regulation of worker foraging Groenewoud F, Frommen JG, Josi D, Tanaka H, Jungwirth A, Taborsky in paper wasps: a social feedback control system (Polistes fuscatus, M. 2016. Predation risk drives social complexity in cooperative breeders. Hymenoptera: Vespidae). Behaviour. 102:147–167. Proc Natl Acad Sci USA. 113:4104–4109. Rubinstein A. 1982. Perfect equilibrium in a bargaining model. Econom J Hamilton IM, Taborsky M. 2005. Unrelated helpers will not fully compen- Econom Soc. 50:97–109. sate for costs imposed on breeders when they pay to stay. Proc R Soc B Russell E, Rowley I. 1988. Helper contributions to reproductive success Biol Sci. 272:445–454. in the splendid fairy-wren (Malurus splendens). Behav Ecol Sociobiol. Hammerstein P. 1981. The role of asymmetries in animal contests. Anim 22:131–140. Behav. 29:193–205. Stiver KA, Dierkes P, Taborsky M, Gibbs HL, Balshine S. 2005. Relatedness Heg D, Bachar Z, Brouwer L, Taborsky M. 2004. Predation risk is an eco- and helping in fish: examining the theoretical predictions. Proc R Soc B logical constraint for helper dispersal in a cooperatively breeding cichlid. Biol Sci. 272:1593–1599. Proc R Soc Lond Ser B Biol Sci. 271:2367–2374. Stiver KA, Fitzpatrick J, Desjardins JK, Balshine S. 2006. Sex differences Heg D, Taborsky M. 2010. Helper Response to Experimentally Manipulated in rates of territory joining and inheritance in a cooperatively breeding Predation Risk in the Cooperatively Breeding Cichlid Neolamprologus cichlid fish. Anim Behav. 71:449–456. pulcher. PloS One 5:e10784. Taborsky B, Skubic E, Bruintjes R. 2007. Mothers adjust egg size to Hellmann JK, Hamilton IM. 2018. Dominant and subordinate outside op- helper number in a cooperatively breeding cichlid. Behav Ecol. tions alter help and eviction in a pay-to-stay negotiation model. Behav 18:652–657. Ecol. 29:553–562. Taborsky M. 1984. Broodcare helpers in the cichlid fish Lamprologus brichardi Hellmann JK, Ligocki IY, O’connor CM, Reddon AR, Garvy KA, Marsh- - their costs and benefits. Anim Behav. 32:1236–1252. Rollo SE, Gibbs HL, Balshine S, Hamilton IM. 2015. Reproductive Taborsky M. 1985. Breeder-helper conflict in a cichlid fish with broodcare sharing in relation to group and colony-level attributes in a cooperative helpers - an experimental analysis. Behaviour. 95:45–75. breeding fish. Proc R Soc B Biol Sci. 282:20150954. Taborsky M. 2016. Cichlid fishes: a model for the integrative study of social Jungwirth A, Walker J, Taborsky M. 2015. Prospecting precedes dispersal behavior. Cooperative breeding in vertebrates. Cambridge: Cambridge and increases survival chances in cooperatively breeding cichlids. Anim University Press, p 272–293. Behav. 106:107–114. Taborsky M, Cant MA, Komdeur J. 2021. The evolution of social behav- Kingma SA, Hall ML, Peters A. 2011. Multiple benefits drive helping be- iour. Cambridge: Cambridge University Press. havior in a cooperatively breeding bird: an integrated analysis. Am Nat. Taborsky M, Grantner A. 1998. Behavioural time-energy budgets of co- 177:486–495. operatively breeding Neolamprologus pulcher (Pisces: Cichlidae). Anim Kingma SA, Santema P, Taborsky M, Komdeur J. 2014. Group augmen- Behav. 56:1375–1382. tation and the evolution of cooperation. Trends Ecol Evol. 29:476–484. Taborsky M, Limberger D. 1980. Lamprologus brichardi, a cichlid fish with fac- Kokko H, Johnstone RA, Wright J. 2002. The evolution of parental and ultative polygyny and helpers at brood care and territory maintenance. alloparental effort in cooperatively breeding groups: when should helpers Newsl IAFE. 3:63–64. pay to stay? Behav Ecol. 13:291–300. Thompson FJ, Marshall HH, Sanderson JL, Vitikainen EI, Nichols HJ, Leadbeater E, Carruthers JM, Green JP, Rosser NS, Field J. 2011. Nest in- Gilchrist JS, Young AJ, Hodge SJ, Cant MA. 2016. Reproductive com- heritance is the missing source of direct fitness in a primitively eusocial petition triggers mass eviction in cooperative banded mongooses. Proc R insect. Science. 333:874–876. Soc B Biol Sci. 283:20152607. 372 Behavioral Ecology Webster MS, Tarvin KA, Tuttle EM, Pruett-Jones S. 2004. Reproductive Zöttl M, Frommen JG, Taborsky M. 2013c. Group size adjustment to promiscuity in the splendid fairy-wren: effects of group size and auxiliary ecological demand in a cooperative breeder. Proc R Soc B Biol Sci. reproduction. Behav Ecol. 15:907–915. 280:20122772. Zöttl M, Chapuis L, Freiburghaus M, Taborsky M. 2013a. Strategic re- Zöttl M, Heg D, Chervet N, Taborsky M. 2013d. Kinship lowers duction of help before dispersal in a cooperative breeder. Biol Lett. alloparental care where helpers pay-to-stay. Nat Commun. 9:20120878. 4:1341. Zöttl M, Fischer S, Taborsky M. 2013b. Partial brood care compensation by Zöttl M, Schreier T, Taborsky M. 2022. Data for: Coercion promotes female breeders in response to experimental manipulation of alloparental alloparental care in cooperative breeders. Behav Ecol. doi: 10.6084/ care. Anim Behav. 85:1471–1478. m9.figshare.19984580.v1.

Journal

Behavioral EcologyOxford University Press

Published: Feb 28, 2023

Keywords: coercion; punishment; negotiation; cooperative breeding; pay-to-stay; Neolamprologus pulcher

There are no references for this article.