SR: Counter Intuitive Behavior Flashcards
Suppressing subordinate reproduction provides benefits to dominants in cooperative societies of meerkats
M.B.V. Bell, M.A. Cant, C. Borgeaud N. Thavarajah J. Samson T.H. Clutton-Brock 2014
In many animal societies, a small proportion of dominant females monopolize reproduction by actively suppressing subordinates. Theory assumes that this is because subordinate reproduction depresses the fitness of dominants, yet the effect of subordinate reproduction on dominant behaviour and reproductive success has never been directly assessed.
Here, we describe the consequences of experimentally preventing subordinate breeding in 12 groups of wild meerkats (Suricata suricatta) for three breeding attempts, using contraceptive injections. When subordinates are prevented from breeding, dominants are less aggressive towards subordinates and evict them less often, leading to a higher ratio of helpers to dependent pups, and increased provisioning of the dominantʼs pups by subordinate females. When sub- ordinate breeding is suppressed, dominants also show improved foraging efficiency, gain more weight during pregnancy and produce heavier pups, which grow faster. These results confirm the benefits of suppression to dominants, and help explain the evolution of singular breeding in vertebrate societies.
Here, subordinates themselves may represent a resource, which is depleted when they attempt to breed and start investing in their own offspring, rather than contributing to cooperative offspring care. This is thought to explain why dominant females in cooperative breeders commonly suppress subordinate reproduction by targeted aggression, temporary eviction or infanticide11–13.
A secondary assumption in an important subset of models:
Selection to minimize the cost of fighting may explain why dominant efforts at suppression are not inevitable, instead appearing to be sensitive to variation in the payoffs of interfering with subordinate breeding: attacks are targeted at subordinates who are most likely to breed15; are restricted to periods when resource competition peaks and the offspring of dominants may be at a competitive disadvantage15,16; or are avoided entirely when the subordinate retaliation is likely to be effective17.
Evidence for negative effects of subordinate reproduction on dominants is limited to observations that subordinate breeding is sometimes associated with
reductions in aspects of dominant female reproductive success, including care received by offspring18, offspring weight at independence15 and offspring survival13,16,19.
It is unclear that this is causal, and may be due to age or condition related declines in dominant condition or capacity to prevent subordinate breeding. Evidence for the cost of attempting to alter the distribution of reproduction by interfering with the reproduction of other individuals is restricted to a few observations of declines in the quality of offspring subsequently produced by dominants, the quality of care they receive or their probability of survival20,21. Again, it is difficult to rule out correlations with other social or ecological variables (food availability in particular).
We experimentally tested the assumptions of reproductive skew theory using injections of the contraceptive hormone Depo-provera (medroxyprogesterone acetate, at 7.5 mg kg 1) to prevent subordinate female reproduction in our long term study population of wild, habituated meerkats at the Kuruman River Reserve, South Africa
Every week, we conducted at least two 30-min focal watches on each dominant female (to give a total of 1,067 h of observations on 12 females), recording every instance of aggres- sion directed towards subordinate females. Dominant females attacked treated subordinates at lower rates
We also recorded the total amount of time when at least one subordinate female was within 2m of the dominant, finding that dominants were more tolerant of the presence of treated subordinates
we investigated the probability of eviction during the dominant femaleʼs gestation for 128 subordinate females, present at 59 breeding attempts (33 control and 26 treated).
Treated females were less likely to be evicted by the dominant female during her gestation
It is unclear how dominants detected the suppressed reproductive state of subordinates: it is likely that olfactory cues played a role22, and there were some behavioural changes in subordinates (for instance subordinates were less submissive
Helper:pup ratio. Under natural conditions, evicted females frequently fail to return, either because they die, or because they establish new groups with unrelated males23. Eviction may therefore reduce the number of helpers present in a group, with negative effects on pup development24.
Reducing the eviction rates may therefore increase the number of helpers present during pup provisioning, which we assessed by calculating the average number of adult females (4360 days old) present between 20 and 40 days after birth (the period during which pups are primarily dependent
on provisioning by adults). We restricted the analysis to females because (i) they contribute more than males to pup provisioning25; and (ii) the number of
males fluctuates due to temporary absences while prospecting for mating opportunities in other groups26. The ratio of female helpers to pups was greater in treated groups (LMM F1,58 1⁄4 5.89, P 1⁄4 0.019; Fig. 1c; Supplementary Table 6), which is likely to have substantial positive effects on subsequent pup development.
This increase in foraging efficiency, coupled with the reduced effort invested in evicting subordinates, meant that dominant females in treated groups gained more weight during pregnancy
Pup growth rate. Pups in treated groups started life heavier, but were also in groups with more helpers, many of whom were feeding at higher rates, so we expected them to show elevated growth rates after emergence. Analysis of pup morning weight revealed that pups in treated groups grew faster between emer- gence and 95 days (LMM interaction between treatment and age F1,7619 1⁄4 4.35, P 1⁄4 0.03; interaction between treatment and age2 F1,7619 1⁄4 30.66, Po0.001; Fig. 3; Supplementary Table 14).
Pup weight at emergence and independence is likely to have profound long term effects on pup fitness: size at emergence determines competitive ability in early life18; experimental feeding to increase pup weight increases survival24; and size at adulthood affects probability of attaining dominance, dominance tenure and reproductive success29.
Rather than direct attack, dominants may simply reduce opportunities for subordinate reproduction, for instance by denying access to unrelated members of the opposite sex
We should also expect dominants to deploy tactics which reduce the cost of conflict, for instance by exaggerating power asymmetries, via strategic resource allocation in early life.
Owing to trade-offs between investment in current and future reproduction, factors that diminish a parent’s survival prospects, such as predation threat, are expected to increase investment in existing young. Nevertheless, effects of predation risk on parental investment have only rarely been examined, and not at all within the context of filial cannibalism (parental consumption of their own offspring). We examined filial cannibalism and nest characteristics in a small fish with paternal egg care, the sand goby, Pomatoschistus minutus, both when exposed to a common piscivore, the perch, Perca fluviatilis, and in the absence of predators.
We found that when males consumed only some of their eggs (partial filial cannibalism), the number of eaten eggs did not depend on predation threat, possibly indicating that partial clutch consumption is largely motivated by benefits to existing young. Total filial cannibalism (whole clutch consumption) was marginally less common under predator exposure, while its strongest predictor was small clutch size. This suggests that the return on parental investment has a greater influence on total filial cannibalism than the likelihood of future breeding. Regarding nest architecture, males that consumed their entire brood after exposure to a predator built larger nest entrances, possibly to facilitate predator evasion. Males that cared for at least part of their brood, however, maintained small nest entrances regardless of predation threat. Furthermore, more elaborate nests were not associated with greater egg consumption, suggesting that filial cannibalism is not employed to sustain nest building.
In this study, we set out to examine the effect of perceived predation risk on filial cannibalism in the sand goby by comparing the behaviour of egg-tending males exposed to a perch, Perca flu- viatilis, a common predator of sand gobies (Koli et al., 1985; Lappalainen et al., 2001), with that of males guarding eggs in a comparatively safe environment. We also examined the effect of predation threat on nest construction. This could elucidate the motives behind any adjustment of the level of filial cannibalism and test whether filial cannibalism is employed to acquire energy for nest maintenance and construction as suggested by earlier findings showing that good body condition and supplemental feeding in sand gobies promote higher quality or more extensive nest building (Lehtonen & Wong, 2009; Lindstro€m, 1998; Olsson, Kvarnemo, & Svensson, 2009).
Parental care confers important fitness benefits to parents by improving the survival of their offspring (Alonso-Alvarez & Velando, 2012; Clutton-Brock, 1991). However, looking after young can be costly (Alonso-Alvarez & Velando, 2012; Clutton- Brock, 1991). It can be time consuming (e.g. Thomson et al., 2014), energetically demanding (e.g. Gravel & Cooke, 2013), and expose parents to predation (e.g. Li & Jackson, 2003) or disease (e.g. Nordling, Andersson, Zohari, & Lars, 1998). As a result, parents may have to trade off investment in existing young against investment in future reproduction (Clutton-Brock, 1991; Trivers, 1972). In this regard, a range of factors can alter the optimal balance of invest- ment in these two fitness components (Klug, Alonzo, & Bonsall, 2012).
Experimentation took place during the sand goby breeding season (MayeJuly 2014) at the Tva€rminne Zoological Station (59 50.70N, 23 15.00E) on the Baltic Sea’s coast. Gobies were collected within the nearby nature reserve using a hand trawl (Evans & Tallmark, 1979; Lehtonen & Kvarnemo, 2015a) and dip- nets, while a gillnet was used to capture perch. After capture, all fish were brought to the station and placed in single-species stock aquaria within a semi-outdoor laboratory facility where experi- mentation occurred. Within this facility, all aquaria received sea water flow-through and were exposed to natural light and tem- perature conditions. Sand gobies housed in stock aquaria were segregated by sex and fed daily on frozen chironomid larvae and live Neomysis shrimp. Perch remained unfed for the duration of the experiment.
To initiate a replicate, a male and female sand goby were selected and their wet mass and standard length were measured. Females were chosen based on the presence of a distended abdomen, indicating gravidity (Kvarnemo, 1997). Males were selected haphazardly but those under 30 mm standard length were avoided, as larger males dominate nesting sites in this species (Lindstro€m, 1988; Lindstro€m & Pampoulie, 2005; Magnhagen & Kvarnemo, 1989), with smaller males often prevented from spawning or resorting to sneak spawning tactics. After selection, each maleefemale pair of gobies was added to an experimental aquarium (Fig. 1). Each of these aquaria contained a pair of plastic barriers, one opaque and one clear, which divided them into a ‘nesting compartment’ and an ‘exposure compartment’ (Fig. 1). The maleefemale pair was intro- duced to the nesting compartment, which had been provided with a sand substrate and an artificial nesting site. The nesting site comprised a flowerpot half (diameter of the mouth: 8 cm) with its interior lined with a thin acetate sheet, upon which females attached their eggs during spawning. Photographed after spawning, photographed after 7 days brooding. Counted eggs. After the eggs were returned to the nest-holding male, he was randomly assigned to either the ‘predator’ (N 1⁄4 40) or ‘control’ (N 1⁄4 39) treatment. In the predator treatment, we then added a perch (standard length: 176 ± 31 mm [mean ± SD], N 1⁄4 40) and a plastic plant to the aquarium’s exposure compartment, with the
plastic plant providing refuge for the perch. In the control treat- ment, only a plastic plant was added to the exposure compartment. This ensured that all fish were subject to novel stimuli. In both treatments, we then removed the opaque barrier dividing the nesting and exposure compartments, leaving only the transparent barrier in the tanks. To prevent other cues disturbing the fish after this, we wrapped the exterior vertical walls of the aquaria in black plastic and left the males to brood their eggs. We did not offer any food to males during this brooding period, as opportunities to forage are restricted during brood care (Lindstro€m & Hellstro€m, 1993; Salgado, Cabral, & Costa, 2004).
Exposure to a perch also tended to decrease the likelihood of a male engaging in complete clutch cannibalism, occurring in seven of 33 (21%) perch-exposed males compared to 11 of 31 (35%) unexposed males, although this effect was marginally nonsignificant
Males with small initial clutches were significantly more likely to engage in total filial cannibalism.
For partial filial cannibalism, none of the potential predictors we examined (predation treatment, prespawning male body condition and dissolved oxygen levels) had significant effects on the number of eggs eaten (Table 1).
We predicted that exposure to a predator would result in increased investment in the existing young, since the perceived likelihood of surviving to engage in future reproduction should be diminished by a seemingly risky environment. However, we found that partial filial cannibalism was not influenced by the presence of a predatory perch.
For total filial cannibalism, males tended to consume their entire brood less often in the presence of a perch, although this result was marginally nonsignificant.
Therefore, there is currently insufficient evidence to show that parental predation risk influences filial cannibalism. Future research addressing total filial cannibalism should thus consider predation risk. First, our results suggest that threat of predation may have a more prominent effect on this form of filial cannibalism. Second, total filial canni- balism is only thought to be adaptive if parents can survive to reproduce again (Manica, 2002), which is contingent on avoiding predation. By contrast, partial filial cannibalism may be less (or not at all) influenced by predation risk because it can benefit parents in ways that do not depend on further reproduction, such as through improved survival of existing young (Klug et al., 2006; Lehtonen & Kvarnemo, 2015a, 2015b; see also Klug & Lindstro€m, 2008).
For example, in the side-blotched lizard, Uta stansburiana, and shaded broad-bar moth, Scotopteryx chenopodiata, females that have been injured, and are thus less able to evade predators, appear to elevate their level of parental expenditure on current offspring (Fox & McCoy, 2000; Javois & Tammaru, 2004). Thus, injury may have a more direct bearing on parents’ perception of predation risk and, as a result, their current reproductive decisions. Similarly, the timing of predator cues may be important especially for total filial cannibalism, which typically occurs soon after spawning (Forsgren et al., 1996; reviewed in Manica, 2002)
The factor that most strongly influenced the occurrence of total filial cannibalism was clutch size. In line with findings from pre- vious work on both sand gobies (see Methods section), as well as other species (see Manica, 2002 for a review) such as the fantail darter, Etheostoma flabellare (Lindstro€m & Sargent, 1997) and blue- gilled sunfish, Lepomis macrochirus (Neff, 2003), males were more likely to engage in total filial cannibalism when their initial clutch size was small. This suggests that the benefits of total filial canni- balism outweigh the costs of offspring consumption when brood size is small. Moreover, our finding that consumption of a large number of eggs among total filial cannibals was associated with piling less sand onto the nest suggests that those males that totally consume larger clutches may be less motivated to maintain the nest site or are inherently less capable of doing so.
In conclusion, our results indicate that partial filial cannibalism was not affected by the perceived threat of predation to the parent, possibly because this type of cannibalism is performed to benefit the current brood. Our results with regard to total filial cannibalism were less clear-cut, with a nonsignificant tendency for males to avoid total filial cannibalism when predation threat was high, suggesting that further research into this area could be informative. With regard to nest construction, we showed that consumption of a larger clutch was not linked to more elaborate nest construction, which may indicate that males do not consume eggs for the pur- pose of nest maintenance. Moreover, we found that aspects of nest architecture were influenced by the presence of a predatory perch only in total filial cannibals, suggesting that care-giving males do not compromise nest maintenance when confronted by a predator. Predator presence was linked to large nest entrances among total filial cannibals, possibly due to reluctance to renest in the vicinity of a predator or because large-entrance nests may be safer to build or occupy. Overall, the findings of this study highlight the importance of considering the effect of adult predation risk on parental effort decisions.
