Biparental care is more than the sum of its parts: experimental evidence for synergistic effects on offspring fitness

Abstract

Despite an extensive body of theoretical and empirical literature on biparental cooperation, it is still unclear whether offspring fare equally, better or worse when receiving care by two parents versus a single parent. Some models predict that parents should withhold the amount of care they provide due to sexual conflict, thereby shifting as much of the workload as possible to their partner. This conflict should lead to offspring faring worse with two parents. Yet, other models predict that when parents care for their offspring together, their individual contributions can have synergistic (more than additive) effects on offspring fitness. Under this scenario, biparental cooperation should lead to offspring faring better with two parents. We address this fundamental question using a unique experimental design where we compared offspring fitness when the two parents worked together (biparental treatment) and when they worked separately (uniparental treatment), while keeping constant the amount of resources and number of offspring per parent across treatments. This made it possible to directly compare the biparental treatment to the sum of the male and female contributions in the uniparental treatment. Our main finding was that offspring grew larger and were more likely to survive to adulthood when reared by both parents than a single parent. This is the first empirical evidence for a synergistic effect of biparental cooperation on offspring fitness and could provide novel insights into the conditions favouring the evolution of biparental cooperation.

1. Introduction

Biparental care occurs when male and female parents cooperate to provide care for their joint offspring. Although biparental care is relatively rare, it has evolved repeatedly in birds, mammals, fishes, amphibians and insects [1–3]. In general, biparental care is expected to evolve when it increases offspring fitness to such an extent that it outweighs the fitness loss to a caring parent in terms of lost breeding opportunities [4]. Nevertheless, when two parents cooperate to rear their offspring, conflict inevitably arises over how much each should contribute towards care [5–7]. This is because the benefit of care to each parent depends on the combined effort of the two parents, whereas the cost depends on each parent's personal effort [6,8]. Thus, biparental care provides an excellent model system for investigating the balance between cooperation and conflict between two unrelated individuals [6,7].

Due to this sexual conflict, each parent is under selection to reduce its personal cost by shifting as much of the workload as possible over to its partner. Consequently, parents that are working together are expected to withhold the amount of care they provide towards the current brood, compared to when working alone [9]. This may result in offspring faring worse when cared for by two parents than a single parent, as reported in a prior study on zebra finches [10]. On the other hand, offspring may fare better with two parents if there are synergistic (more than additive) effects of the individual contributions of the two parents on offspring fitness [6,11,12]. This is referred to as complementarity and is predicted when male and female parents specialize in different tasks during care [13]. Under this scenario, the total beneficial effect of the combined effort of two parents on offspring performance exceeds the sum of the beneficial effects of each individual parent when they provide care separately. Since sexual conflict and complementarity are not mutually exclusive, it is likely that there is a balance between cooperation and conflict in most systems [14]. Despite extensive theoretical and empirical work on biparental care over the past few decades (e.g. [13–17]), it is still unclear whether this balance between cooperation and conflict leads to offspring faring equally, better or worse when receiving care by a single parent versus both parents.

Here, we address this fundamental gap in our understanding of biparental care, using an experimental design where we compared parental behaviour and offspring performance when the two parents worked together (biparental treatment) and when they worked separately (uniparental treatment). The rationale for this design was that there was scope for sexual conflict and synergy in the biparental treatment, whereas the potential for sexual conflict and synergy was experimentally removed in the uniparental treatment. We first measured parental behaviours to investigate (i) whether parents withhold care when working together, as predicted due to sexual conflict and (ii) whether the type of care provided by male and female parents is more divergent when they work together than separately, indicating task specialization. We then measured offspring fitness-related traits to determine whether (i) offspring perform equally well with two parents, as expected if the beneficial effect of the parents' individual contributions is simply additive, (ii) offspring perform better with two parents, as expected if biparental care has synergistic benefits or (iii) offspring perform worse with two parents, as expected if parents withhold care due to sexual conflict. To allow a direct comparison between the biparental treatment and the sum of the male and female contributions in the uniparental treatment, we kept constant the amount of resources and the number of offspring (i.e. the potential workload) per parent across treatments (figure 1).

Figure 1.

Experimental design illustration (not drawn to scale). Males (blue) and females (orange) in the uniparental treatment were separated after egg laying and were each provided with one mouse of a standardized size and a mixed-parentage brood of 15 larvae. Parents in the biparental treatment were allowed to stay together and were provided with two mice of a standardized size and a mixed-parentage brood of 30 larvae.

Our study species, Nicrophorus vespilloides, has a number of important attributes that make it particularly well-suited for this experiment. Firstly, biparental care, female uniparental care and male uniparental care all occur in natural populations of this species, with a relative frequency of 52%, 39% and 3%, respectively [18,19]. Our biparental and uniparental treatments therefore match conditions in the wild. Secondly, burying beetles breed on carcasses of small vertebrates, which provide the sole source of food for the parents and developing larvae. This made it possible to keep the amount of resources per parent and offspring constant across treatments by providing the two parents in the biparental treatment with two carcasses of a standardized size and each parent in the uniparental treatment with one carcass (figure 1). Thirdly, parents do not distinguish between their own larvae and unrelated larvae [20]. This allowed us to standardize the number of offspring per parent by providing parents with mixed-parentage broods of 30 larvae in the biparental treatment and 15 larvae in the uniparental treatment (figure 1). Lastly, prior work in burying beetles has found good evidence for sexual conflict over care, as well as sex differences in care indicating a potential for task specialization [21–26]. Thus, the potential for both conflict and complementarity makes this species an excellent study system for investigating whether offspring fare equally, better, or worse with two parents working together or with a single parent working alone.

2. Material and methods

(a). Study species

Biparental care in N. vespilloides is elaborate and complex. Once a carcass is found, parents bury it into the soil, remove any fur or feathers, deposit antimicrobial secretions to its surface, and lay eggs around it 24–48 h after mating [18,27]. When the eggs hatch approximately 60 h later [28], the larvae crawl to the carcass and start feeding in a crater created by the parents on the top of the carcass. The larvae can self-feed, but the parents also provision larvae with predigested carrion [29]. There is some evidence for sex differences in parental care with females spending more time provisioning food for the larvae and staying on the carcass for longer than males [21–23,26]. Larvae disperse from the carcass about 5 days after hatching, which corresponds to the end of the parental care period. They pupate about 10–12 days after dispersal and eclose as adults about 10–12 days after pupation.

(b). Animal husbandry

We used virgin beetles from an outbred laboratory population maintained at the University of Edinburgh. The beetles used in this study comprised fifth- and sixth-generation beetles from lines originally collected in Edinburgh, UK. They were housed individually in transparent plastic containers (12 × 8 × 2 cm) filled with moist soil and kept at 22°C and a 16 h : 8 h light : dark cycle. All non-breeding adults were fed small pieces of raw organic beef twice a week.

(c). Experimental design and procedures

Our experimental design included a biparental treatment, where the two parents worked together and a uniparental treatment, where the two parents worked separately (figure 1). We kept constant the number of offspring (i.e. the potential workload) and the amount of resources per parent across treatments (figure 1) to allow a direct comparison of parental effort and offspring performance between the biparental treatment and the sum of the uniparental male and female contributions.

The parents used in this experiment were mated within two weeks after reaching sexual maturity (i.e. 10–24 days after eclosion). We only mated unrelated males and females that did not share any common ancestors for at least two generations. Just before mating, we recorded the pre-breeding mass of each parent and measured their pronotum width using digital calipers. Each experimental pair (n = 130) was then placed in a transparent plastic container (17 × 12 × 6 cm) filled with 1 cm of moist soil and two freshly thawed mouse carcasses of a standardized size (10–12 g). Half of these pairs were randomly assigned to the uniparental treatment (n = 65) and the other half to the biparental treatment (n = 65). There was no difference (two-sample t-test: t₁₂₉ = −0.07, p = 0.95) between total mass of the two carcasses assigned to the biparental treatment (mean ± s.d. = 22.03 ± 0.97 g) and the uniparental treatment (22.05 ± 1.29 g).

Immediately after egg laying, we separated the parents from the eggs so that the larvae would hatch in isolation from the parents. Parents in the biparental treatment were moved jointly, along with the two carcasses, to a new container with moist soil. Parents in the uniparental treatment were transferred to separate containers with moist soil, and each parent was provided with one of the two carcasses given to them initially.

When the eggs started hatching, we used the newly hatched larvae to generate broods of 15 or 30 larvae for the uniparental or biparental treatments, respectively (figure 1). All experimental broods included larvae of mixed parentage to eliminate any effects of parent–offspring coadaptation [23,30,31]. In this species, parents do not distinguish between unrelated foster broods and their own broods, as long as the larvae are at the same developmental stage [20]. As parents kill any larvae that arrive on the carcass before their eggs are expected to hatch [32], we only provided parents with a brood once their own eggs had hatched. We were not able to provide experimental broods to all parents, because we were limited by the number of larvae that hatched at the same time. Thus, our final sample sizes were n = 40 for the biparental treatment where both parents cared for a brood of 30 larvae and n = 49 for the uniparental treatment where the male and the female cared separately for two broods of 15 larvae (figure 1).

We conducted behavioural observations to compare the amount of care that male and female parents provided to their offspring in the biparental versus the two uniparental treatments. These observations were done 24 h after parents were given a brood, as this stage in larval development corresponds to a peak in post-hatching care in this species [29]. We used instantaneous sampling every 1 min for 30 min [21–24,33]. We recorded the number of scans that a female spent providing (i) direct care, defined as provisioning food or interacting with the larvae and (ii) indirect care, defined as guarding or carcass maintenance (i.e. deposition of secretions to the surface of the carcass or excavation of the crypt).

We checked the containers daily in the morning and in the afternoon to determine whether the parents were present on the carcass or were away from the brood in the soil. Parents that were away for more than two consecutive checks were deemed to have abandoned the brood and were removed from the boxes to prevent infanticide. Based on the last observation when the parent was present on the carcass, we were able to estimate the duration of care by each parent. At the dispersal stage, we recorded the number of surviving larvae and measured the total brood mass to calculate average larval mass in each brood. After being weighed, all larvae from each brood were placed into large transparent containers (17 × 12 × 6 cm) filled with moist soil. At eclosion, we recorded the number of individuals that eclosed successfully. These data were used to calculate the survival rate for each brood from the dispersal stage to the eclosion stage (survival to adulthood). We also measured the parents’ post-breeding mass at the dispersal stage. By subtracting each parent's pre-breeding mass from its post-breeding mass, we calculated its overall mass change over the breeding attempt, which is a measure of somatic investment [33,34]. The parents were transferred to individual containers (12 × 8 × 2 cm) filled with moist soil. They were checked twice a week until death to determine their post-breeding lifespan, which served as a proxy for residual reproductive value [34].

(d). Data analysis

We used linear models for continuous traits with normally distributed random errors (average larval mass at dispersal, parent mass change and parent post-breeding lifespan). For discrete traits, we used generalized linear models fitted with a negative binomial error distribution (brood size at dispersal) or a quasi-poisson error distribution (amount of direct care, amount of indirect care and duration of care). For proportion data (probability of brood abandonment, offspring survival to adulthood), we used generalized linear models fitted with a binomial and quasi-binomial distribution, respectively. Quasi-Poisson and quasi-binomial distributions were used to account for overdispersion by including a dispersion parameter that describes additional variance in the data.

We first compared parental behaviour and offspring performance between the biparental treatment and the uniparental treatment. To this end, we calculated the sum of the male and female contributions for the following variables: amount of direct and indirect care, duration of care, mass change of each parent, brood size at dispersal and number of offspring surviving to adulthood. For average larval mass, we calculated total brood mass by adding up the brood mass across the two uniparental treatments and divided that by the total number of larvae across the two broods. Each of these variables was then used as a response variable with treatment (uniparental or biparental) as a factor.

Carcass size was added as a covariate to the models for average larval mass and brood size at dispersal, because the amount of resources available may influence offspring growth and survival, respectively. Based on prior evidence that the parents' body size can influence offspring fitness in this species [35,36], we also added male and female pronotum width as covariates to the models for average larval mass, brood size at dispersal and survival to adulthood.

We next compared the behaviours of male and female parents across treatments. The response variables were amount of direct and indirect care, duration of care, parental mass change over the breeding attempt and parent post-breeding lifespan. The explanatory variables were parent sex (male or female), treatment (uniparental or biparental) and the interaction between the two. Decisions about which variables to include in the final models were based on AIC values following criteria from Burnham and Anderson [37]. All analyses were performed using R v. 3.4.2 [38]. The ggplot2 package was used for generating figures [39].

3. Results

(a). Parental behaviour

Broods received a similar total amount of care (direct care: t₈₆ = 0.83, p = 0.50; indirect care: t₈₆ = −1.55, p = 0.13; duration of care: t₈₆ = 0.13, p = 0.89) regardless of whether the two parents worked together (i.e. biparental treatment) or separately (i.e. uniparental treatment). Nevertheless, examining each sex separately revealed that males provided less care and females provided more care when working with a partner than when working alone (figure 2a–c). This pattern was true for amount of direct care (treatment × parent sex: LR , p < 0.0001), amount of indirect care (treatment × parent sex: LR , p < 0.001), as well as duration of care (treatment × parent sex: LR , p = 0.004). Similarly, females were less likely to abandon the brood in the biparental (5%) than the uniparental treatment (18%), whereas males were more likely to abandon the brood in the biparental (35%) than the uniparental treatment (30%) (treatment × parent sex: LR , p = 0.038). The average probability of abandonment was the same across treatments (LR , p = 0.48).

Figure 2.

Boxplots showing the amount of direct care (a), amount of indirect care (b), duration of care (c) and mass change (d) by males (blue) and females (orange) in the uniparental (n = 49) and biparental (n = 40) treatments. Direct care refers to food provisioning and interactions with larvae (e.g. grooming). Indirect care refers to carcass maintenance and guarding. The amount of direct and indirect care provided by parents was measured using scan sampling during 30-min behavioural observations. Duration of care refers to the number of days each parent was present on the carcass before abandoning the brood. Filled circles indicate individual data points with the size of the circle representing the frequency of observations.

Females gained more mass and males gained less mass when the two parents worked together compared to when they worked separately (treatment × parent sex: F_1,143 = 8.59, p = 0.004; figure 2d). However, total mass change did not differ between the uniparental and biparental treatments (t₈₆ = 1.33, p = 0.19). Post-breeding lifespan did not depend on sex (F_1,170 = 1.56, p = 0.21), treatment (F_1,170 = 0.15, p = 0.70) or the interaction between sex and treatment (F_1,170 = 0.46, p = 0.50).

(b). Offspring performance

Larvae reared by parents who worked together were larger at the end of the parental care period than larvae reared by parents who worked separately (LR , p < 0.001; figure 3a). This difference in average larval mass was not associated with a trade-off between offspring size and number, because there was no evidence for a difference in brood size at the dispersal stage between the biparental and uniparental treatments (LR , p = 0.92). In addition to having a higher larval mass, offspring reared by both parents had a higher survival rate to adulthood than offspring reared by a single parent (LR , p = 0.03; figure 3b).

Figure 3.

Boxplots showing average offspring mass (a) and number of surviving offspring (b) at the dispersal stage in the uniparental (green) and biparental (purple) treatments (n = 49 and n = 40, respectively). Filled circles indicate individual data points with the size of the circle representing the frequency of observations.

In terms of the covariates included in the above models, average larval mass was higher on larger carcasses (LR , p = 0.028) and when the female was larger (LR , p = 0.042) but not when the male was larger (LR , p = 0.65). Brood size at dispersal was not influenced by the male's body size (LR , p = 0.78), the female's body size (LR , p = 0.76), or the size of the carcass (LR , p = 0.28). Lastly, offspring of larger males (LR , p = 0.04) and larger females (LR , p = 0.02) were more likely to survive to adulthood.

4. Discussion

In this study, we first tested (i) whether parents withhold care when working together, as predicted due to sexual conflict and (ii) whether the type of care provided by male and female parents is more divergent when they work together than separately, indicating task specialization. We found that males, but not females, provided less care when working with a partner, and there was no evidence for task specialization. We then tested whether (i) offspring perform equally well with two parents, as expected if the beneficial effect of the parents' individual contributions is simply additive, (ii) offspring perform better with two parents, as expected if biparental care has synergistic benefits or (iii) offspring perform worse with two parents, as expected if parents withhold care due to sexual conflict. In accordance with the second scenario, we found that larvae reared by both parents were larger at the end of the parental care period and more likely to survive to adulthood than offspring reared by a single parent. To our knowledge, this is the first empirical evidence for a synergistic effect of biparental care on offspring fitness. Below, we offer potential explanations for our results and discuss their implications for our understanding of biparental care.

Our first main finding was that parents adjusted their effort depending on whether they were caring alone or together. Males provided less care, whereas females provided more care, when working with a partner. This pattern may be a consequence of sexual conflict over care where males, but not females, withheld the amount of care they provided to shift some of the workload over to their partner. This would suggest that females were exploited by males, because they were forced to increase their effort to compensate for their partner's reduced effort. However, an alternative explanation is that females had the upper hand in sexual conflict over carcass consumption. Previous work on this species suggests that sexual conflict over carcass consumption is closely linked to sexual conflict over parental care and may be equally important [24,25]. In our study, females consumed more of the carcass and gained more mass in the biparental than in the uniparental treatment, whereas the opposite was true for males. Mouse carcasses are a highly nutritional resource for parents who feed on the carcass before and during a breeding attempt to replenish their energy reserves. Thus, if females controlled access to the carcass in the biparental treatment, the lower level of care by males might reflect that males were prevented from feeding on the carcass and were thus unable or unwilling to provide an equal amount of care [24,25].

Our second main finding was that offspring fared better when receiving care by both parents than a single parent. These synergistic fitness benefits of care were evident before offspring independence (i.e. larval mass) and persisted after independence (i.e. survival to adulthood). The mass of a larva at the dispersal stage is a crucially important fitness component in this species. Because larvae do not feed after dispersal and before eclosion, larval mass determines adult size [40]. In turn, adult size influences lifespan, fecundity and the likelihood of acquiring a carcass for breeding [34,35,41,42]. Thus, the higher larval mass of offspring reared by two parents may have downstream fitness benefits with respect to the offspring's reproductive success as adults. Interestingly, the only other study to directly test how offspring fare with one versus two parents found the opposite pattern. Using zebra finches, Royle et al. [10] compared a biparental treatment to a female uniparental treatment (but not a male uniparental treatment) and showed that nestlings reared by a single female received more care per offspring than those reared by both parents. This difference in parental investment had consequences for offspring fitness later in life, with male offspring from uniparental broods being more sexually attractive than male offspring from biparental broods [10].

Synergistic effects are predicted when there is task specialization between male and female parents [13]. In burying beetles, parents provide care by provisioning food to the larvae, grooming larvae, maintaining the carcass and guarding the carcass from predators and competitors. Under task specialization, we would expect a greater divergence in parental care behaviours when parents work together than when they work separately, but this was not the case here. We found that female parents provided more direct care (food provisioning) and more indirect care (carcass maintenance) in the biparental treatment than the uniparental treatment, whereas male parents provided less direct care and less indirect care in the biparental than the uniparental treatment. Thus, even though there were sex differences in care (with females providing more care than males), there was no evidence for task specialization.

Our study provides no evidence for the mechanistic basis of the synergistic effects we observed, but one possibility is that they were driven by differences in carrion consumption between treatments. Given that carrion consumption by the parents can negatively affect offspring fitness in this species [43], the lower offspring performance in the uniparental treatment could be due to higher carcass consumption by the parents. However, this explanation is not supported by our results, because there was no difference in the parents' total mass change between treatments.

A more likely explanation is that these synergistic effects are related to a component of parental care that we did not measure directly. For example, N. vespilloides parents deposit oral and anal fluids onto the carcass during larval development. These secretions contain a wide range of compounds, which have been shown to increase larval survival by facilitating the vertical transmission of symbiotic microbiota [44] and by preventing bacterial and fungal growth on the decomposing carcass [27]. In the absence of these secretions, microbes can compromise larval health and degrade the quality of the carcass as a food resource to the offspring [27]. The lysozymes in these secretions can vary between individuals [45], so secretions from two parents are likely to be more diverse than secretions from a single parent. We encourage future research to investigate the mechanistic basis of the synergistic effects of biparental cooperation.

Overall, our study shows that offspring fare better in broods with two parents despite one of the parents providing less care (in this case, the male). These findings contradict the assumption that sexual conflict between parents reduces offspring fitness by causing parents to withhold parental investment [10,46]. We suggest that synergistic effects may be more common than currently appreciated and that the lack of empirical evidence for complementarity may be due to the lack of studies explicitly testing for it. It is important to note that it is not possible to detect synergistic effects by only studying parents who are providing care together (as is typically done in most studies of biparental care). Instead, it is essential to use an experimental design where some parents are allowed to work together and some work separately, while the number of offspring per parent is kept constant.

In conclusion, this work provides evidence for a synergistic effect of biparental care on offspring fitness by showing that offspring grow better and are more likely to survive to adulthood when reared by two parents working together than by a single parent working alone. Evolutionary theory has long considered the role of synergistic effects [11,13,17], but until now, we have lacked empirical support for their existence. Our results can therefore provide valuable insights into the conditions that might favour the evolution of biparental care.

Data accessibility

Data have been deposited in the Dryad Digital Repository: http://dx.doi.org/10.5061/dryad.s7n04j8 [47].

Authors' contributions

N.P. and P.T.S. conceived and designed the experiment. N.P. and E.J.H.H. performed the experiment. N.P. performed the data analysis and wrote the first draft of the manuscript. P.T.S. edited the manuscript. All authors read and approved the final manuscript.

Competing interests

We declare we have no competing interests.

Funding

The study was funded by the Institute of Evolutionary Biology and the School of Biological Sciences at The University of Edinburgh.