Post-hatching parental care masks the effects of egg size on offspring fitness: a removal experiment on burying beetles

Authors


Correspondence: Per T. Smiseth, Institute of Evolutionary Biology, School of Biological Sciences, University of Edinburgh, West Mains Road, Edinburgh EH9 3JT, UK.

Tel.: +44 131 651 3682; fax: +44 131 650 6564;
e-mail: per.t.smiseth@ed.ac.uk

Abstract

Parents can increase the fitness of their offspring by allocating nutrients to eggs and/or providing care for eggs and offspring. Although we have a good understanding of the adaptive significance of both egg size and parental care, remarkably little is known about the co-evolution of these two mechanisms for increasing offspring fitness. Here, we report a parental removal experiment on the burying beetle Nicrophorus vespilloides in which we test whether post-hatching parental care masks the effect of egg size on offspring fitness. As predicted, we found that the parent's presence or absence had a strong main effect on larval body mass, whereas there was no detectable effect of egg size. Furthermore, egg size had a strong and positive effect on offspring body mass in the parent's absence, whereas it had no effect on offspring body mass in the parent's presence. These results support the suggestion that the stronger effect of post-hatching parental care on offspring growth masks the weaker effect of egg size. We found no correlation between the number and size of eggs. However, there was a negative correlation between larval body mass and brood size in the parent's presence, but not in its absence. These findings suggest that the trade-off between number and size of offspring is shifted from the egg stage towards the end of the parental care period and that post-hatching parental care somehow moderates this trade-off.

Introduction

Parents of many animals engage in various activities that increase the survival and growth of their offspring, often at a cost to the parents' own survival and reproduction (Clutton-Brock, 1991). Firstly, female parents of many species allocate nourishment to the developing embryo either via yolk that is deposited into the egg as in birds and many insects, or via a placenta as in mammals (Smith & Fretwell, 1974). Secondly, parents of either sex may provide parental care after laying or hatching, for example by guarding the eggs or young against predators and/or providing the young with food after hatching (Clutton-Brock, 1991). Although there has been considerable progress over the past four decades in our understanding of the evolution of both egg size (Smith & Fretwell, 1974; Bernardo, 1996; Fox & Czesak, 2000) and parental care (Trivers, 1972; Clutton-Brock, 1991; Royle et al., in press), far less is known about the co-evolution of these two principal mechanisms by which parents can increase offspring fitness. Indeed, the evolution of egg size in species where parents care for eggs or juveniles has been and remains a highly debated subject in evolutionary ecology (Shine, 1978, 1989; Nussbaum, 1985, 1987; Sargent et al., 1987; Nussbaum & Schultz, 1989; Gilbert & Manica, 2010).

Comparative studies on ectothermic vertebrates suggest that species where parents guard their eggs tend to have larger eggs than species where parents provide no care for eggs (Shine, 1978, 1989; Gross & Sargent, 1985; Nussbaum, 1985, 1987; Sargent et al., 1987; Kolm & Ahnesjö, 2005). The main debate has centred on alternative evolutionary explanations for this association between parental care and egg size (Shine, 1978, 1989; Nussbaum, 1985, 1987; Sargent et al., 1987; Nussbaum & Schultz, 1989). For example, the safe harbour hypothesis suggests that selection should favour larger eggs in species where parents guard eggs because this form of care reduces egg mortality relative to juvenile mortality. Thus, larger eggs would allow the offspring to complete more of their development within the relative safety of the egg (Shine, 1978, 1989). Alternatively, selection may favour the evolution of parental care in species that for some reason produce larger eggs if egg mortality otherwise increases with egg size, as is the case in many aquatic species where larger eggs require more oxygenation (Nussbaum, 1987; Nussbaum & Schultz, 1989; Smith, 1997).

There is mounting evidence suggesting that the relationship between egg size and parental care may be less clear-cut than traditionally thought (Gilbert & Manica, 2010). For example, a recent comparative study on insects found no difference in egg size between species where parents provide some care for eggs or offspring and species where parents provide no care (Gilbert & Manica, 2010). Furthermore, evidence from birds suggests that altricial species, where parents provide more elaborate forms of post-hatching care that include provisioning of food to hatched offspring, appear to have smaller eggs than precocial species, where parents provide simpler forms of post-hatching care (Wesolowski, 1994; Williams, 1994). At present, little is known about why different animal taxa show such contrasting patterns, but potential explanations include phylogenetic constraints on egg size and/or parental care, and differences between taxonomic groups with respect to environmental conditions or the predominant forms of care.

Although comparative studies provide valuable insights into the relationship between egg size and parental care, such studies are observational in nature and of limited value in terms of understanding the causal effects that parental care may have on the evolution of egg size and vice versa. Thus, to elucidate the effects of post-hatching parental care on the evolution of egg size, it is now essential to conduct experiments that compare the effects of egg size on offspring fitness in the presence and absence of parental care. Such parental removal designs require the use of suitable study species where parents provide elaborate post-hatching care that includes parental food provisioning and where offspring can survive in the absence of post-hatching parental care, such as burying beetles of the genus Nicrophorus (Eggert et al., 1998; Smiseth et al., 2003).

Here, we investigate whether post-hatching parental care masks or otherwise modifies the effects of egg size on larval fitness in Nicrophorus vespilloides. To this end, we compare the effects of egg size on larval fitness in the absence and presence of parents. Previously, similar designs have been used to investigate the effects of parental care on offspring growth and survival (Eggert et al., 1998), the outcome of sibling competition (Smiseth et al., 2007a,b) and the transition to nutritional independence (Smiseth et al., 2003). To examine whether post-hatching parental care may mask or otherwise modify the initial effects of egg size on offspring fitness, we test for an effect of the interaction between parental presence or absence and egg size on offspring size and/or survival. We also tested for a trade-off between the size and number of offspring at both the egg stage and the end of parental care. Given that parental allocation of resources through the egg is small relative that through food provisioning after hatching, we expect the trade-off between the size and number of offspring to be stronger at the end of parental care than at the egg stage and that the trade-off at the end of parental care would be more pronounced in the parent's presence when offspring obtain resources from the parents. To examine how post-hatching parental care influences the trade-off between the size and number of offspring, we test for an effect of the interaction between parental presence or absence and offspring number on offspring body size.

Materials and methods

Study species

Like all members of the genus Nicrophorus, Nvespilloides breed on carcasses of small vertebrates, which provide the sole source of food for the larvae (Eggert & Müller, 1997; Scott, 1998). Females lay up to 58 eggs in the soil near the carcass (Smiseth et al., 2006). Parents may later adjust the brood size to match the size of the carcass through filial cannibalism (Bartlett, 1987). Although the larvae self-feed immediately from hatching, they also obtain food by begging for predigested carrion from their parents (Smiseth et al., 2003). Parental food provisioning increases offspring growth and survival (Eggert et al., 1998; Smiseth et al., 2003).

Experimental procedures

The beetles were from two out-bred laboratory popu-lations derived from wild-caught beetles collected at Kennel Vale, Cornwall, and Corstorphine Hill, Edinburgh. Prior to the experiments, all beetles were housed individually in clear plastic containers (12 × 8 cm and 2 cm high) filled with moist soil. All beetles were kept at a temperature of 20 °C. Nonbreeding beetles were fed small pieces of organic beef twice a week. For use in the experiments, we randomly selected 76 pairs of virgin, nonsibling males and females for use in the experiment. Each pair was placed in a clear plastic container (17 cm × 12 cm and 6 cm high) filled with about 1 cm of moist soil and provided with a defrosted mouse carcass (supplied by Livefoods Direct, Sheffield, UK). All carcasses used were of a standardized weight, ranging from 16.5 to 21.9 g.

After 48 h, we checked each container for eggs. We recorded clutch size as the number of eggs visible through the container's bottom. We recorded clutch size in this way instead of counting all eggs, as the latter would involve handling of eggs that, in turn, may cause damage to eggs. To verify our methodology, we conducted a pilot study on a separate sample of beetles. We compared the estimated clutch size, obtained as described above, with the actual clutch size based on counting all eggs that had been laid in the container. There was a very strong correlation between the estimated and actual clutch sizes (Pearson's correlation: r = 0.98, n = 21, P < 0.001), although the estimated clutch size was slightly below (15.5 ± 2.7 eggs) the actual clutch size (17.6 ± 2.9 eggs).

We selected ten eggs at random from each container that were carefully removed using a damp paintbrush. The eggs were photographed using an Olympus DP20 digital camera mounted on an Olympus SZX10 microscope at 6.3 × magnification. The eggs were then returned to the container and covered with soil. All damaged eggs (n = 7) were excluded from further analyses. We measured egg length and width in pixels from the photographs using imagej (Abramoff et al., 2004). The measurements were then converted into metric length (mm) and used to calculate a prolate spheroid volume V for each egg using the equation V = (1/6)πw2L, where w is the width and L the length of the egg (Berrigan, 1991).

We excluded six pairs from further analysis because they produced fewer than ten eggs. The remaining 70 pairs were allocated randomly to the two treatment groups. We allocated 40 pairs to the parent's absence treatment by removing both parents 72 h after pairing. Furthermore, we allocated 30 pairs to the of parent's presence treatment (control treatment) by leaving the female to care for the larvae. From these broods, we removed the male 72 h after pairing because male involvement in care is highly variable and male removal has no detectable effect on larval fitness (Smiseth et al., 2005). Previous work has shown that parents open the carcass around the time of hatching to provide larvae with access to the interior of the carcass and that this form of care is essential to larval survival (Eggert et al., 1998). Thus, to eliminate potential effects due to parental opening of the carcass, we used a razor blade to cut an opening through the skin of the carcass immediately after we had removed the parents. The female was observed providing food to the larvae on at least one occasion, thus verifying that parental food provisioning took place in broods allocated to presence of parents treatment. We allocated more pairs to the absence of parent treatment because these broods were less likely to produce larvae that survived until dispersal (χ1 = 5.45, P < 0.025).

We left the containers for 5–7 days to allow the larvae time to develop and reach dispersal. Of the 70 pairs, 53 produced broods that survived until larval dispersal from the carcass, which marks the end of parental care. The final analyses of the effects of egg size on offspring fitness and the trade-off between the size and number of offspring at the end of parental care were based on these 53 broods. At the time of dispersal, we recorded the brood size and weighed each brood to the nearest 0.1 mg to calculate an average larval body mass per brood.

Given that female characteristics may confound the relationships between egg size, parental care, and offspring body size and survival, we recorded female age and size. We recorded female age as the number of days from the date of her dispersal from the carcass to the date she was paired up for use in the experiment. We recorded female size as pronotum length, which was measured using electronic callipers. Two females were not measured as they were lost or damaged during the experiment. We measured each female three times. We used a one-way anova with individual as factor to calculate repeatability for size measurements (Lessells & Boag, 1987). This analysis confirmed that the measurements of female size had a very high repeatability (r = 0.96). We used the mean value for each female in the statistical analysis.

Statistical analysis

Because we have no means of marking individual eggs and larvae, we cannot track the performance of indivi-dual larvae hatching from particular eggs. We therefore base all statistical analyses on mean values of egg size, larval body mass and survival for each brood. We first tested for a trade-off between the size and number of offspring at the egg stage using a general linear model (GLM) with egg size as dependent variable and clutch size as a covariate. We included female size and female age as covariates as these female characteristics may influence egg size and/or parental care (Bolton, 1991; Lock et al., 2007). We next used GLM to test for effects on larval body mass at the time of larval dispersal. We added egg size as a covariate and parental presence (or absence) treatment as a factor. We included the interaction between parental presence and egg size to test whether post-hatching parental care modifies the effect of egg size. We included brood size as a covariate to test for a trade-off between larval body mass and brood size, and the interaction between parental presence and brood size to test whether post-hatching parental care modifies this trade-off. We added female size and age as covariates. Finally, we used GLM to test for effects on larval survival measured as the proportion and number of larvae that survived until the time of dispersal from the carcass. The data on the proportion of surviving larvae were logit-transformed before being used in the statistical analyses (Warton & Hui, 2011). In both models, we added egg size as a covariate, parental presence as a factor, and the interaction between parental presence and egg size, and female size and age as covariates. In the model on the number of surviving larvae, we included clutch size as covariate as clutch size could influence the number of surviving larvae.

Results

Variation in egg size

The overall mean egg size in the population was 1.89 ± 0.03 mm3 (n = 700 eggs from 70 clutches). Variation in the mean egg size per clutch ranged from 1.44 to 2.43 mm3. Sixty-two percentage of the variation in egg size was due to variation between clutches, and the remaining 38% was due to variation within clutches. There was no significant relationship between egg size and clutch size (F1,63 = 0.21, P = 0.65), suggesting that there was no trade-off between the size and number of offspring at the egg stage (Fig. 1). Furthermore, female characteristics had no significant effects on egg size (female size: F1,65 = 0.13, P = 0.72; female age: F1,64 = 0.15, P = 0.70).

Figure 1.

Relationship between egg size (measured as egg volume) and clutch size in Nicrophorus vespilloides.

Larval body mass

Post-hatching parental care had a statistically significant main effect on larval body mass at dispersal from the carcass (Table 1); larvae grew to a larger size in the parent's presence than in its absence (Fig. 2a). Egg size had no significant main effect on larval body mass at dispersal (Table 1). However, there was a significant effect of the interaction between egg size and the parent's presence (Table 1), suggesting that post-hatching parental care modifies the effects of egg size on offspring fitness. Egg size had a stronger effect on larval body mass at dispersal in the parent's absence than in its presence (Fig. 2a). Brood size had a significant main effect on larval body mass at dispersal (Table 1); larvae grew to a larger size in smaller broods than in larger ones (Fig. 2b). There was a significant effect of the interaction between brood size and the parent's presence (Table 1), suggesting that post-hatching parental care influences the trade-off between the size and number of offspring. Brood size had a stronger effect on larval body mass at dispersal in the parent's presence than in its absence (Fig. 2b). Female age had a significant main effect on larval body mass (Table 1), and larvae grew to a larger size when cared for by older females (Fig. 2c). Female body size had no significant effect on larval body mass (Table 1).

Figure 2.

Effects of egg size (a), brood size (b) and female age (c) on larval body mass at dispersal in Nicrophorus vespilloides under two conditions: the presence (open symbols and broken line) and absence (filled symbols and solid line) of the female parent.

Table 1.  GLM model testing for the effects of egg size, the presence or absence of parental care by the female, brood size, female size and female age on larval body mass at dispersal in Nicrophorus vespilloides
Larval body mass at dispersalF1,43P
  1. GLM: general linear model.

  2. Adjusted R2 = 0.73.

Egg size2.200.16
Parental care13.150.001
Egg size*parental care4.350.043
Brood size27.51< 0.001
Brood size*parental care9. 910.003
Female size0.270.61
Female age5.300.026

To explore further details on the interaction between egg size and post-hatching parental care, we conducted separate post hoc analyses for the effect of egg size in the parent's presence and absence, respectively. In these analyses, we included the same covariates as in the full model above. Egg size had a highly significant and positive effect on larval body mass at dispersal in parent's absence (F1,23 = 7.13, P = 0.015), whereas egg size had no significant effect on larval body mass at dispersal in parent's presence (F1,23 = 0.12, P = 0.73).

Larval survival

We first investigated potential factors influencing the proportion of the eggs laid that survived until larval dispersal from the carcass. There were no significant effects of the parent's presence, egg size or the interaction between the two on the proportion of surviving larvae (Table 2). Thus, there was no evidence that post-hatching parental care modified the effects of egg size on larval survival. Furthermore, there were no significant effects of either female size or age on larval survival (Table 2). We next investigated potential factors influencing the number of larvae surviving to dispersal. As in the previous model, there were no significant main effects of the parent's presence, egg size or the interaction between the two on the number of surviving larvae (Table 2). There was a highly significant main effect of clutch size (Table 2), suggesting that the number of larvae surviving increased as a function of egg number. As in the previous model, there were no significant effects of female size and age (Table 2).

Table 2. Two GLM models testing for the effects of egg size, the presence or absence of parental care by the female, clutch size, female size and female age on larval survival in Nicrophorus vespilloides. In the first model, larval survival was measured in terms of the proportion of larvae surviving until dispersal, whereas the second model measured it in terms of number of larvae surviving until dispersal
Proportion of larvae surviving until dispersalF1,45PNumber of larvae surviving until dispersalF1,44P
  1. GLM: general linear model.

  2. Adjusted R2 = 0.01; Adjusted R2 = 0.46.

Egg size3.770.058Egg size1.390.25
Parental care0.170.69Parental care0.300.86
Egg size* parental care0.140.71Egg size* parental care0.050.83
Female size0.280.60Clutch size41.12< 0.001
Female age0.090.77Female size0.040.84
   Female age0.010.1

Discussion

Effects of egg size on larval fitness

The main aim of this study was to investigate whether elaborate post-hatching parental care may mask or otherwise modify the effect of egg size on offspring fitness. To this end, we used a novel parental removal design that allowed us to compare the effect of egg size on offspring fitness in the absence and presence of post-hatching parental care. Firstly, we found that the parent's presence or absence had a strong main effect on larval body mass, whereas there was no detectable main effect of egg size. This finding supports the suggestion that post-hatching parental care has a much stronger effect on offspring growth than egg size (Ricklefs, 1984). Secondly, we found a significant effect of the interaction between the parents' presence and egg size on larval body mass. Egg size had a positive effect on larval body mass in the parent's absence, whereas it had no effect on larval body mass in the parent's presence. These results provide support for the suggestion that the strong effect of post-hatching parental care on offspring fitness will mask any initial effect of egg size (Ricklefs, 1984). Previous removal experiments suggest that post-hatching parental care may increase offspring growth and survival through various mechanisms, including parental food provisioning, opening of the interior of the carcass and deposition of antimicrobials (Eggert et al., 1998; Rozen et al., 2008). Our results cannot identify the specific mechanism by which post-hatching care is masking the effects of egg size, but we suspect that parental food provisioning would be a prime candidate because this form of care has a particularly strong effect on offspring growth (Ricklefs, 1984). Further work including observations of parental behaviour is now needed to establish whether this is indeed the case.

Although our results support the suggestion that the strong effect of post-hatching parental care masks the much weaker effect of egg size on offspring fitness, it is important to consider alternative explanations. Avian ecologists are often concerned that variation in territorial and/or parental quality may generate a positive correlation between the effects of egg size and parental care. Cross-fostering designs provide a powerful tool for eliminating such confounding effects, and studies based on such designs show that egg size can have an independent effect on offspring fitness (Bolton, 1991; Magrath, 1992). If there were such a positive correlation between the effects of egg and parental care in N. vespilloides, we should expect a stronger association between egg size and offspring fitness in the presence of parental care. The reason is that our parental removal design eliminates confounding effects due to a positive correlation between the effects of egg size and parental care when parents were removed, but not when parents were left to provide care. Thus, the finding that egg size had no effect on offspring body mass in the presence of parental care, whereas it had a positive effect in its absence, suggests that there is not a positive correlation between the effects of post-hatching parental care and egg size in N. vespilloides.

Although we found an effect of the interaction between the parent's presence and egg size on larval body mass, this was not the case for larval survival. The difference in the results for larval survival and body mass may simply reflect that larval survival was more variable than larval body mass, leading to lower statistical power in the analysis on survival. Larval survival is obviously more directly linked to offspring fitness than larval body mass, but it is important to note that larval body mass has important fitness consequences in N. vespilloides. Because the larvae do not feed after dispersal from the carcass, adult body size is determined by larval body mass at dispersal (Bartlett & Ashworth, 1988; Lock et al., 2004). Adult body size, in turn, is a major predictor of reproductive success because it influences the outcome of competition for carcasses used for breeding (Otronen, 1988).

The results reported in this study raise the intriguing question as to why there is so much variation in egg size. Given that female parents normally remain to provide post-hatching parental care for the larvae in N. vespilloides (Smiseth & Moore, 2002) and that our results show that egg size had no effect on larval fitness when larvae receive such parental care, it would be logical to expect females to be under selection to produce the smallest possible egg size. By doing so, females would save resources that otherwise could be allocated to other functions, such as production of additional eggs, increased parental care or future survival, without jeopardising the offspring's fitness (Brockelman, 1975). One potential explanation for the maintenance of variation in egg size is that the optimal egg size depends on the quality or type of resource used for breeding. Burying beetles breed on a wide range of carrion including small mammals, birds and amphibians (Scott, 1998), and fresh and rotten carrion (Rozen et al., 2008). Studies on other species suggest that the relative advantages of larger eggs are greater when offspring face harsh environmental conditions (Fox & Czesak, 2000). Thus, if some of the resources used by the beetles constitute a harsher environment than others, environmental heterogeneity may maintain variation in egg size.

Finally, our results suggest that female age had a positive effect on larval body size at dispersal. Potentially, this effect might reflect that older females allocate more time to post-hatching parental care that promote larval growth, including provisioning of food to larvae (Lock et al., 2007). If this were the case, we would expect that the effect of female age would be stronger when the female parent was present than when she was absent. However, our results show that this effect had a similar magnitude in the presence and absence of the parent (Fig. 2c), suggesting that it is mediated through some mechanism that occurs prior to hatching, such as quality or composition of eggs or various forms of prehatching parental care including carcass preparation. Furthermore, our finding that female age had a positive effect on larval body size at dispersal contrasts with previous work on N. vespilloides, which has reported that female age has no effect on larval body size (Lock et al., 2007) or that female breeding performance measured as the total brood mass declines with female age (Cotter et al., 2010). Thus, our results suggest that further work is needed to elucidate the effects of female age on larval growth and survival.

Trade-off between size and number of offspring

An additional aim of our study was to test for a trade-off between the number and size of offspring in N. vespilloides and whether post-hatching parental care might somehow modify this trade-off. We found no evidence of a negative correlation between egg size and clutch size, suggesting that the females that allocated more resources to each egg did not have fewer resources available for the production of other eggs. We found a negative correlation between larval body mass and brood size at larval dispersal in the parent's presence, whereas there was no such correlation in the parent's absence. Our results provide two interesting insights into the trade-off between the number and size of offspring in species where parents provide extensive post-hatching parental care. Firstly, the lack of a correlation between egg size and clutch size, which may reflect that the costs of egg production are small in this species as females feed off the carcass prior to laying, supports the suggestion that the trade-off between number and size of offspring in species with elaborate post-hatching parental care is shifted from the egg stage towards the end of the period when larvae receive care from their parents (Smith & Fretwell, 1974). Secondly, the finding that there was a negative correlation between larval body mass and brood size in the parent's presence, but not in its absence, suggests that post-hatching parental care somehow moderates the trade-off between the number and size of offspring.

Our results provide no information about the mechanism by which post-hatching parental care may modify the trade-off between the number and size of offspring. Here, we briefly discuss two potential, nonexclusive explanations. Firstly, post-hatching parental care may modify the trade-off between the number and size of offspring because it exacerbates the effects of sibling competition. In a previous brood-size manipulation experiment on N. vespilloides, we found that larvae assigned to larger broods had lower effectiveness of begging (i.e. were less likely to be fed at a given begging level) and spent less time begging than larvae assigned to smaller broods (Smiseth et al., 2007a). These results suggest that parental food provisioning leads to increased interference among competing siblings because whenever a given larva is successful at obtaining food from the parent, its success will necessarily come at the expense of other larvae in the brood. By contrast, there is no evidence of interference when larvae are forced to compete by self-feeding in the absence of the parent (Smiseth et al., 2007a).

Secondly, post-hatching parental care may modify the trade-off between the number and size of offspring because it enhances the larvae's ability to exploit the carcass. If so, the lack of a trade-off in the parent's absence may reflect that larval growth in this situation is limited by the larvae's ability to self-feed rather than amount of resources they have available (at least for the range of carcass sizes used in this experiment). Indeed, the observation that larvae reared in the absence of a parent dispersed from the carcass at a similar body mass regardless of brood size (Fig. 2b) suggests that larvae may disperse once they reach a certain threshold size. In contrast, larvae reared in the presence of a parent showed increased growth (Smiseth et al., 2003), suggesting that post-hatching parental care facilitates larval feeding to the extent that the larvae consume more of the carcass. Thus, the finding that larval body mass decreases as a function of brood size in the parent's presence may reflect that larvae from larger broods consumed more of the carcass and ended up smaller simply because they had shared it with a larger number of competitors. Interestingly, the body mass of larvae from large broods reared in parent's presence converge with the putative threshold size for larvae reared in parent's absence (Fig. 2b). This finding suggests that the benefit of post-hatching parental care is dependent on brood size and that this benefit (at least under laboratory conditions) may be relatively small for larger brood sizes.

Conclusions

Here, we have shown that egg size has a strong and positive effect on offspring body mass in the parent's absence, while it has no effect on offspring body mass in the parent's presence. Based on these findings, we conclude that post-hatching parental care masks the effect of egg size on offspring growth. Our results have implications for our general understanding of the co-evolution between egg size and parental care in species where parents provide elaborate post-hatching care that includes food provisioning to offspring. For example, our results are consistent with the suggestion that altricial birds have a reduced egg size compared to precocial birds (Wesolowski, 1994; Williams, 1994), because the evolution of parental food provisioning in the former had the effect of masking any initial effects of egg size (Ricklefs, 1984). There is now a need for further experimental work on a wide range of species, including insects and altricial birds, that specially address how different forms of post-hatching parental care may mask or otherwise modify the effects of egg size. Such experiments could be based on the use of parental removal experiments on suitable species, but could also be based on other designs such as handicapping of parents or food supplementation. These latter designs would allow researchers to test whether the effects of egg size on offspring fitness decreases as a function of increasing levels of a specific form of post-hatching parental care.

Acknowledgments

We thank the Countryside Rangers service, Edinburgh, for permission to collect beetles at Corstorphine Hill and Allen Moore and Chloe Bird at the University of Exeter for providing beetles caught at Kennel Vale. We thank Ronnie Mooney for help with animal husbandry, Tom Little and Phil Wilson for access to their digital camera and microscope, and Matt Bell, James Gilbert, Tamas Székely, Steve Trumbo and two anonymous reviewers for valuable comments on earlier drafts of the manuscript. PTS is funded by a grant from NERC (NE/G004293/1).

Ancillary