Volume 31, Issue 11 p. 2098-2107
RESEARCH ARTICLE
Free Access

When earwig mothers do not care to share: Parent–offspring competition and the evolution of family life

Jos Kramer

Corresponding Author

Jos Kramer

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Department of Plant and Microbial Biology, University of Zürich, Zürich, Switzerland

Correspondence

Jos Kramer

Email: [email protected]

Search for more papers by this author
Maximilian Körner

Maximilian Körner

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Janina M. C. Diehl

Janina M. C. Diehl

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Christine Scheiner

Christine Scheiner

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Aytül Yüksel-Dadak

Aytül Yüksel-Dadak

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Teresa Christl

Teresa Christl

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Philip Kohlmeier

Philip Kohlmeier

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Search for more papers by this author
Joël Meunier

Joël Meunier

Zoological Institute, Evolutionary Biology, Johannes Gutenberg University, Mainz, Germany

Institut de Recherche sur la Biologie de l'Insecte, UMR 7261, CNRS/Université François-Rabelais, Tours, France

Search for more papers by this author
First published: 07 June 2017
Citations: 19

Paper previously published as Standard Paper.

Abstract

  1. Kin competition often reduces – and sometimes entirely negates – the benefits of cooperation among family members. Surprisingly, the impact of kin competition on the fitness effects of family life only received close scrutiny in studies on sibling rivalry, whereas the possibility of parent–offspring competition has attracted much less attention. As a consequence, it remains unclear whether and how parent–offspring competition could have affected the early evolution of parental care and family life.
  2. Here, we examined the occurrence and consequences of parent–offspring competition over food access in the European earwig Forficula auricularia, an insect with facultative family life reminiscent of an ancestral state. Specifically, we (1) raised earwig offspring under food limitation either together with or without their mother, and (2) tested whether and how the — potentially competitive — weight gains of mothers and offspring during family life affected the offsprings' survival rate and morphology, and the future reproductive investment of their mother.
  3. In line with the occurrence of parent–offspring competition, we showed that high maternal weight gains during family life reduced the survival prospects of maternally tended offspring, while they increased the mothers' investment into the production of a second clutch (but not the body size of the surviving offspring). Conversely, high offspring weight gains generally increased the offsprings' survival, but did so to a larger extent when they were together with their mother. Intriguingly, mothers that had exhibited a low initial weight showed especially high weight gains.
  4. Overall, our results demonstrate that maternal presence under food restriction triggered a local competition between mothers and their offspring. This competition limited offspring survival, but allowed mothers to increase their investment into future reproduction and/or to maintain their current body condition. On a general level, our findings reveal that parent–offspring competition can counteract the benefits of (facultative) parental care, and may thus impede the evolution of family life in resource-poor environments.

A plain language summary is available for this article.

1 INTRODUCTION

The association of offspring with their parents after birth or hatching (defined as family life; Mock & Parker, 1997) is a rare, but taxonomically widespread phenomenon (Royle, Smiseth, & Kölliker, 2012) that can only evolve if the fitness benefits of family interactions outweigh the costs of a prolonged association of the family members (Alonso-Alvarez & Velando, 2012). The fitness benefits of family life predominantly derive from the expression of parental care (Costa, 2006; Wong, Meunier, & Kölliker, 2013), which ultimately increases the survival and/or quality of tended juveniles (Klug, Alonso, & Bonsall, 2012). By contrast, the costs of family life typically result from kin competition among family members for limited resources and reproduction (Krause & Ruxton, 2002; West, Pen, & Griffin, 2002). This competition is known to affect crucial life-history traits such as sex allocation (Frank, 1985) and aggressive behaviour (West, Murray, Machado, Griffin, & Herre, 2001), and commonly arises during family life where juveniles strive to monopolize limited parental resources (sibling rivalry; Mock & Parker, 1997; Roulin & Dreiss, 2012). However, kin competition during family life can also arise between offspring and their parents. Such parent–offspring competition can, for instance, promote the dispersal of offspring that reached nutritional independence (Cockburn, 1998; Cote, Clobert, & Fitze, 2007), and therefore hampers the evolution of complex family systems in cooperative breeders (Sorato, Griffith, & Russell, 2016). Shedding light upon the parties engaged in competition is hence crucial to advance our understanding of the cost-benefit ratio of family life, and more generally its emergence and maintenance in nature (Klug & Bonsall, 2010; Klug et al., 2012).

The emergence of family life marks the appearance of facultative forms of (post-hatching) parental care (Kölliker, 2007; Smiseth, Darwell, & Moore, 2003). Somewhat surprisingly, the possibility of local parent–offspring competition in species featuring such facultative care has been poorly explored, and its impact on the early evolution of parental care and family life therefore remains unknown. This oversight conceivably results from a historical bias towards studying parental care in altricial, i.e. obligatory-caring, species (Clutton-Brock, 1991), in which the scope for parent–offspring competition during early family life is limited by the low foraging capabilities of juveniles. During earlier stages of the evolution of family life (and in contemporary precocial species), however, juveniles are not fully dependent on parental resources (Kölliker, 2007; Smiseth et al., 2003), and the early onset of offspring foraging might put them into competition with their caring parents early on. In such situations, the costs of parent–offspring competition could diminish the benefits of parental care, and might ultimately render parental presence (Boncoraglio & Kilner, 2012; Scott & Gladstein, 1993) and family life maladaptive altogether (Meunier & Kölliker, 2012).

The intensity of competition typically depends on the environmentally determined availability of limited resources. Harsh environments could therefore have a profound impact on parent–offspring competition and the evolution of family life (Klug et al., 2012; Wilson, 1975). The nature of this impact, however, remains controversial. On the one hand, harsh conditions should favour the evolution of family life, because the benefits of parental care are then likely substantial (Clutton-Brock, 1991; Wilson, 1975). On the other hand, harsh environments are expected to exacerbate the costs of care (such as an increased energy loss; Alonso-Alvarez & Velando, 2012), and the limited resource availability may not only favour the deferral of parental investment (Clutton-Brock, 1991; Klug & Bonsall, 2007), but also increase parent–offspring competition and thus hamper the evolution of family life. In line with this hypothesis, the prolonged presence of fathers has been shown to reduce offspring survival under food limitation in Nicrophorus vespilloides Herbst, a burying beetle with biparental care in which both parents feed on the carcass employed for breeding (Boncoraglio & Kilner, 2012; Scott & Gladstein, 1993). Moreover, food restriction has been shown to offset the benefits of maternal care in the European earwig Forficula auricularia L. (Meunier & Kölliker, 2012), suggesting that kin competition between offspring and their tending mother might have rendered maternal presence and hence family life detrimental to offspring survival.

In this study, we investigated the occurrence and consequences of mother–offspring competition under harsh conditions (emulated by food restriction) in the European earwig F. auricularia. In this precocial insect, mothers provide extensive forms of care to their mobile offspring (called nymphs) for several weeks after hatching (Lamb, 1976a). Maternal care includes the protection and grooming of nymphs, as well as their provisioning with food (e.g. through regurgitation; Lamb, 1976b; Staerkle & Kölliker, 2008), and is even expressed by mothers in a bad nutritional condition (Kramer & Meunier, 2016). However, post-hatching maternal presence and care are not obligatory for offspring survival (Kölliker, 2007; Kölliker & Vancassel, 2007), as nymphs can forage independently soon after hatching (relying on the same food resources than their mother; Lamb, 1976b) and may even obtain food from their siblings (Falk, Wong, Kölliker, & Meunier, 2014; Kramer, Thesing, & Meunier, 2015). Consequently, the cost-benefit ratio of maternal presence in F. auricularia generally depends on the environmental conditions experienced by mothers and nymphs during family life (Kölliker, 2007; Meunier & Kölliker, 2012; Thesing, Kramer, Koch, & Meunier, 2015).

To assess the occurrence of mother–offspring competition, we raised nymphs under food restriction either together with or without their mother, and investigated whether offspring survival was reduced when mothers consumed food and thus restricted offspring feeding (or vice versa). We then examined whether the condition of mothers and/or their nymphs at hatching determined the intensity of the putative competition, as well as whether this intensity affected the mothers' future reproduction and the morphology of their (surviving) first-brood offspring. If mother–offspring competition occurred, we would expect that maternal weight gains during family life negatively affect offspring survival under maternal presence (but not absence). Moreover, we predicted that mothers exhibiting a low initial weight would compete more intensely — and thus curtail their offspring's survival more extensively — than mothers exhibiting a high initial weight. Similarly, we expected that groups of nymphs with low initial weight would compete more intensely with their mother to ensure their own survival. Finally, we predicted that intense competition might benefit mothers and the surviving offspring, for instance by enabling them to increase their investment into future reproduction and their final body size respectively.

2 MATERIALS AND METHODS

2.1 Laboratory rearing and experimental design

We investigated the occurrence and potential repercussions of mother–offspring competition under harsh conditions in 128 female earwigs and their first brood. The females had been collected in July and August 2013 in Mainz, Germany and were reared under standard laboratory conditions (adapted from Meunier et al., 2012; see Appendix S1) until they produced their first clutch. On the first day after egg hatching, we haphazardly distributed a random subset of 32 nymphs (original brood size: 49.6 ± 1.1 nymphs; mean ± SE) among two equally sized groups. We manipulated the potential for mother–offspring competition over food access by raising one of these groups together with their mother (maternal presence-, or MP-group) and the other group without their mother (maternal absence-, or MA-group). After 16 days, mothers were removed from the MP-groups to mimic natural family disruption (Meunier et al., 2012) and to determine their investment into the production of a second (and final) clutch. In parallel, the MP- and MA-nymphs were maintained in their groups, and their survival and morphology were measured upon adult emergence. Note that we excluded 16 of the 128 initially employed families from our analyses because the mother died during family life.

Harsh environmental conditions were emulated according to an established protocol (Meunier & Kölliker, 2012) by providing MP- and MA-groups with a restricted amount of an artificial diet (composition detailed in Kramer et al., 2015) every 6 days for the duration of 3 days respectively. This 6-day cycle was initiated on the first day after hatching and repeated until the juveniles reached adulthood. The amount of food the groups received at a time was increased stepwise from 60 mg (until the end of family life) to 120 mg (until day 31) and finally to 240 mg (until adulthood; cf. Meunier & Kölliker, 2012). This feeding regime successfully established scope for mother–offspring competition, as the food provided during family life was more often fully consumed before removal in MP- than in MA-groups (in 81% vs. 25% of cases; paired Wilcoxon signed rank test: V = 4656, p > .0001). We also manipulated the conditions experienced by mothers after family life to investigate the effect of food availability during that period on their investment into a second clutch. To this end, isolated mothers either received 60 mg of food in a continuation of the above detailed 6-day cycle (low-food treatment; n = 54 mothers), or an ad libitum amount that was renewed twice per week (high-food-treatment; n = 58 mothers). Isolated mothers and all groups of young nymphs were kept in medium Petri dishes (Ø = 9 cm) throughout the experiment. On day 31, groups of older nymphs were transferred into large Petri dishes (Ø = 14 cm). All Petri dishes contained humid sand as a substrate and a plastic tube as a shelter.

2.2 Measurements

We determined the occurrence and intensity of mother–offspring competition by testing for the occurrence and strength of a negative association between maternal food consumption and offspring fitness. To this end, we measured the relative weight gains of mothers and nymphs during family life (as a proxy of food intake), and gathered the survival rate of nymphs at the end of family life and upon adult emergence in all remaining 112 families (see above). The weights of mothers were determined by weighing each mother on the first day after egg hatching and at the end of family life. Similarly, the average weights of nymphs were determined by weighing all nymphs of a family on the first day after hatching, as well as all surviving nymphs of a group at the end of family life, and then dividing these weights by the corresponding number of weighed nymphs. The relative weight changes were calculated by subtracting the weight at the beginning of family life from the corresponding weight at its end, and then dividing this difference by the initial weight. All mothers and nymphs were weighed to the nearest 0.01 mg using a microscale (MYA5, PESCALE, Bisingen, Germany). Offspring survival rates at the end of family life were determined by counting the nymphs that survived until day 16, and then dividing this number by the number of nymphs initially distributed to that group. Likewise, the survival rates upon adult emergence were determined by counting the nymphs that survived until adulthood, and then dividing this number by the number of nymphs alive at the end of family life. Note that two nymphs per group were removed 6 days after hatching to conduct an independent experiment (data not shown). These nymphs were not considered in the calculation of survival rates.

We also determined the consequences of the suspected mother–offspring competition for the future (and final) reproductive effort of mothers, as well as for the morphology of the surviving, adult offspring. To this end, we checked all 112 isolated mothers daily for oviposition over a period of 100 days, and assessed (1) the occurrence of egg deposition and — where applicable, (2) the length of the inter-clutch interval and (3) clutch size (the number of eggs produced within 3 days after the onset of egg deposition; Meunier & Kölliker, 2013). The inter-clutch interval was defined as the number of days between isolation from the first brood and the deposition of the second clutch (Kölliker, 2007). The morphology of adult offspring was assessed by measuring two fitness-relevant morphological traits — eye distance (a proxy of body size) and forceps length (Radesäter & Halldórsdóttir, 1993) — in the first male and female adult that emerged in each group. Overall, we measured 192 males and 198 females [at least one male emerged in 92 (100), and at least one female in 94 (104) of the 112 MP- (MA-) groups]. All morphological measurements were taken under CO2-anesthetization to the nearest 0.001 cm using a camera coupled to a stereo microscope (DFC425, Leica Microsystems Ltd, Heerbrugg, Switzerland) and operated with the software Leica Application Suite 4.5.0.

2.3 Statistical analyses

Establishing the occurrence of mother–offspring competition over food access requires to demonstrate that offspring fitness under maternal presence is reduced (and maternal fitness increased) when mothers restrict offspring feeding through their own food consumption (or vice versa). Accordingly, we predicted that high maternal weight gains would (1) only impair the survival and/or morphology of offspring raised under maternal presence, and (2) increase maternal investment into the production of a second clutch. To verify our first prediction, we used a generalized linear mixed model (GLMM) with binomial error distribution to test whether offspring survival rate (entered as odds ratio) was affected by offspring and maternal weight gains, maternal presence (MP or MA), and the observation period (from the beginning until the end of family life or from the end of family life until adult emergence). An observation-level variable and family-ID were entered into the model as random effects to account for overdispersion (Harrison, 2015) and the common origin of nymphs in the MP- and MA-groups respectively. Because a four-way-interaction between all explanatory variables shaped offspring survival (Table S1; Wald urn:x-wiley:02698463:media:fec12915:fec12915-math-0001 = 7.80, p = .0052), we split the dataset by the observation period and fitted two GLMMs with the same set of variables (except the observation period) within each subset. Similarly, we fitted two separate linear mixed models (LMMs) using family-ID as a random effect to test whether either the eye distance or the forceps length (corrected for eye distance; details in Appendix S2) of adult offspring was affected by maternal and offspring weight gains, maternal presence and adult sex.

To examine our second prediction, we tested in two generalized linear models (GLMs) fitted with a binomial error distribution (corrected for overdispersion) and one linear model (LM) whether the weight gains of mothers and/or offspring affected the mothers' future reproduction. Each of these models included maternal and offspring weight gains, as well as the availability of food after family life (HF or LF) as explanatory variables. The GLMs were fitted using either the occurrence of second clutch production or the relative investment of mothers into the production of their second clutch (entered as odds ratio of second to first clutch eggs) as a response variable. Conversely, the LM was fitted using the length of the inter-clutch interval as response variable. Note that this LM and the GLM analysing the relative investment of mothers into their second clutch were fitted on the subset of those mothers that eventually produced a second clutch. In the final step of our analysis, we investigated the determinants of maternal and offspring weight gains. To this end, we fitted one LM and one LMM that each used the weight of the mother and of the nymphs at hatching (corrected for the initial weight of the mother; details in Appendix S3) as explanatory variables. The LM and LMM were fitted using maternal weight gain and offspring weight gain as response variable respectively. Because this latter model was fitted on both MP- and MA-groups, we included family-ID as a random effect.

All statistical analyses were performed with the statistics software r version 3.0.3 (http://www.r-project.org/). Mixed model analyses were implemented using the packages lme4, car and lmerTest. Note that we scaled the ‘relative weight gain of mothers’ and the ‘average relative weight gain of the offspring’ to unit variance to avoid model bias due to collinearity between these variables (variance inflation factor <4.0 after scaling in all models). All statistical models initially included all possible interactions between the tested variables and were then simplified via the stepwise deletion of non-significant interactions (all p < .05). Significant interactions between continuous variables were plotted using the package ‘effects’ to display the predicted relationship between the response variable and one explanatory variable for different values of the interacting variable(s) (details in Fox, 2003).

3 RESULTS

Two independent interactions between maternal presence and, respectively, the relative weight gains of mothers and the relative weight gains of the offspring determined offspring survival until adulthood (Table 1a). In line with the occurrence of mother–offspring competition, the first interaction revealed that high maternal weight gains reduced the long-term survival of offspring when the mother had been present (Figure 1a; estimate ± SE = −0.175 ± 0.070, z = −2.518, p = .0118), but not when she had been absent during family life (Figure 1a; estimate ± SE = 0.047 ± 0.067, z = 0.698, p = .4854). Conversely, the second interaction showed that the offsprings' long-term survival increased more steeply with their weight gains under maternal presence (Figure 1b; estimate ± SE = 0.359 ± 0.072, z = 4.983, p < .0001) than under maternal absence (Figure 1b; estimate ± SE = 0.161 ± 0.067, z = 2.396, p = .0166).

Table 1. Offspring survival
Offspring survival
(a) Upon adult emergence (b) At the end of family life
Wald urn:x-wiley:02698463:media:fec12915:fec12915-math-0002 p Wald urn:x-wiley:02698463:media:fec12915:fec12915-math-0003 p
Maternal presence (MP) 51.92 <.0001 4.67 .0306
Offspring weight gain (OWG) 23.61 <.0001 1.49 .2228
Maternal weight gain (MWG) 1.21 .2711 <0.01 .9735
MP: MWG 4.56 .0327 0.14 .7085
MP: OWG 6.84 .0089 0.92 .3383
MWG: OWG 1.56a .2114a 0.02 .9029
MP: MWG: OWG 2.35a .1255a 7.00 .0082
  • Influence of maternal presence, and the relative weight gain of mothers and their offspring on the offsprings' survival (a) upon adult emergence and (b) at the end of family life. Significant p-values are given in bold print. (aValues as indicated before removal of the interaction from the model).
Details are in the caption following the image
Offspring survival until adulthood. Influence of (a) the relative weight gain of mothers and (b) the relative weight gain of their offspring on offspring survival until adulthood under maternal presence (filled squares) and absence (empty squares)

Offspring survival until the end of family life was shaped by a triple interaction between maternal presence and the weight gains of mothers and their offspring (Table 1b). In particular, high maternal weight gains reduced the positive impact of high offspring weight gains on offspring survival under maternal presence (Figure 2; maternal weight gain: urn:x-wiley:02698463:media:fec12915:fec12915-math-0004 = 1.083, p = .2981; offspring weight gain: urn:x-wiley:02698463:media:fec12915:fec12915-math-0005 = 0.083, p = .7736; interaction: urn:x-wiley:02698463:media:fec12915:fec12915-math-0006 = 8.278, p = .0040). By contrast, offspring survival during family life was independent of both maternal and offspring weight gains under maternal absence (interaction: urn:x-wiley:02698463:media:fec12915:fec12915-math-0007 = 2.588, p = .1077; maternal weight gain: urn:x-wiley:02698463:media:fec12915:fec12915-math-0008 = 0.276, p = .5995; offspring weight gain: urn:x-wiley:02698463:media:fec12915:fec12915-math-0009 = 0.592, p = .4416).

Details are in the caption following the image
Interactive effect of maternal and offspring weight gains on the survival of maternally tended offspring during family life. To illustrate the interaction, regression lines are given for an average value of maternal weight gains (median = 0.13; dashed line), as well as for a comparatively low (1st quantile = 0.07; solid line) and high (3rd quantile = 0.20; dotted line) weight gains respectively

The weight gains of mothers and their offspring during family life overall decreased with their own initial weight (weight gain of mother: Figure 3a, estimate ± SE = −7.202 ± 1.081, t = −6.663, p < .0001; weight gain of offspring: Figure 3b, estimate ± SE = −0.091 ± 0.022, t112.00 = −4.164, p < .0001), but were independent of each other's initial weight (weight gain of mother: estimate ± SE: −62.058 ± 48.345, t = −1.284, p = .2020; weight gain of offspring: estimate ± SE: 1.685 ± 3.157, t112.00 = 0.534, p = .5940) respectively. Notably, offspring collectively gained less weight under maternal presence (mean ± SE: 0.86 ± 0.03) than under maternal absence (0.99 ± 0.03; urn:x-wiley:02698463:media:fec12915:fec12915-math-0010 = 23.478, p < .0001).

Details are in the caption following the image
Weight gains and initial weight. Influence of (a) the weight of mothers and (b) the weight of nymphs at egg hatching (corrected for the weight of their mother; see above) on their respective weight gains during family life

Overall, 61 of the 112 mothers (54.5%) produced a second clutch. Both the likelihood of second clutch production (Table 2a) and the length of the inter-clutch interval (Table 2b) were shaped by an interaction between maternal and offspring weight gains during first brood family life. Intriguingly, these interactions indicated that high maternal weight gains increased the likelihood of second clutch production (Figure 4a) and decreased the length of the inter-clutch interval (Figure 4b) only if the first brood offspring had featured high weight gains. By contrast, high maternal weight gains actually reduced the likelihood of second clutch production (Figure 4a), and increased the inter-clutch interval (Figure 4b), if they occurred in combination with low offspring weight gains. Note that the interactive effect of maternal and offspring weight gains on the length of the inter-clutch interval was contingent upon the mothers' access to food after family life, as the interaction only emerged when food was restricted (interaction: F1 = 6.03, p = .0234; maternal weight gain: F1 = 0.11, p = .7458; offspring weight gain: F1 = 0.01, p = .9364), but not when it was provided ad libitum (interaction: F1 = 2.68, p = .1112; maternal weight gain: F1 = 0.18, p = .6742; offspring weight gain: F1 = 0.12, p = .7325). Finally, the investment into second clutch production increased with maternal weight gains during first brood family life (Table 2c; estimate ± SE: 0.228 ± 0.098, t = 2.322, p = .0238), and was overall higher when mothers had received food ad libitum after family life (Table 2c). By contrast, the weight gains of 1st brood offspring did not affect the relative investment of mothers into their second clutch (Table 2c; estimate ± SE: −0.150 ± 0.104, t = −1.453, p = .1516).

Table 2. Maternal investment into 2nd clutch production
(a) Likelihood of production (b) Inter-clutch interval (c) Relative investment
LR urn:x-wiley:02698463:media:fec12915:fec12915-math-0011 p F 1 p LR urn:x-wiley:02698463:media:fec12915:fec12915-math-0012 p
Maternal weight gain (MWG) 0.01 .9177 0.28 .5962 5.39 .0203
Offspring weight gain (OWG) 2.03 .1546 0.05 .8285 2.18 .1399
Food availability (FA) 3.70 .0544 4.49 .0388 4.10 .0429
MWG: OWG 10.19 .0014 0.13 .7172 0.80a .3712a
MWG: FA 2.77a .0961a 0.02 .8967 0.03a .8607a
OWG: FA 1.11a .2921a 0.12 .7302 1.00a .3185a
MWG: OWG: FA 0.48a .4904a 7.32 .0092 0.02a .8950a
Type of model | sample size GLM | 112 LM | 61 GLM | 61
  • Influence of maternal weight gains, offspring weight gains and the food availability after family life (high or low food) on (a) the likelihood of second clutch production, (b) the length of the inter-clutch interval, and (c) the relative investment into second clutch production. Significant p-values are given in bold print (aValues as indicated before removal of the interaction from the model).
Details are in the caption following the image
Interactive effects of maternal and offspring weight gains on second clutch traits. Depicted are the interactive effects of maternal and offspring weight gains on (a) the likelihood of second clutch production among all mothers, as well as on (b) the length of the inter-clutch interval among those mothers with restricted food access after family life. To illustrate the interactions, regression lines are given for an average value of offspring weight gains (median = 0.82; dashed line), as well as for a low (1st quantile = 0.64; solid line) and a high value (3rd quantile = 1.03; dotted line) respectively. Note that the axis depicting the inter-clutch interval is reversed as shorter intervals likely reflect higher maternal fitness

Maternal presence affected the eye distance, but not the corrected forceps length, of those offspring that survived until adulthood. In particular, the eye distance of adult offspring was overall smaller when they grew up together with (mean ± SE = 1.340 ± 0.005 mm) rather than without their mother (1.354 ± 0.005 mm; Table 3a), and overall larger in females (1.370 ± 0.005 mm) than in males (1.324 ± 0.004 mm; Table 3a). Moreover, eye distance decreased with increasing maternal weight gains irrespective of maternal presence (Table 3a, estimate ± SE: −0.011 ± 0.004, t102.20 = −3.137, p = .0022), but was independent of offspring weight gains (Table 3a). By contrast, the corrected forceps length was overall larger in males (0.240 ± 0.017 mm) than in females (−0.233 ± 0.012 mm), but independent of maternal presence and the weight gains of the offspring and their mother during family life (Table 3b).

Table 3. Morphology of the surviving adult offspring
(a) Eye distance (b) Forceps length
Wald urn:x-wiley:02698463:media:fec12915:fec12915-math-0013 p Wald urn:x-wiley:02698463:media:fec12915:fec12915-math-0014 p
Maternal presence 6.16 .0131 0.69 .4075
Sex 59.51 <.0001 659.51 <.0001
Maternal weight gain 9.84 .0017 0.69 .4075
Offspring weight gain 0.37 .5423 1.97 .1608
  • Influence of maternal presence, offspring sex, and the weight gains of nymphs and their mother during family life on (a) the eye distance and (b) the (corrected) forceps length of the surviving adult offspring. Significant p-values are given in bold print.

4 DISCUSSION

Kin competition often reduces and sometimes even entirely negates the benefits of cooperation among relatives (West et al., 2002). Surprisingly, however, the role of kin competition between parents and their offspring in the early evolution of parental care has been largely overlooked. Here, we demonstrated that maternal presence under food restriction triggered a local mother–offspring competition in the European earwig F. auricularia. This competition manifested itself as a negative effect of high maternal weight gains on the survival of maternally tended (but not maternally deprived) offspring, and positively affected the future reproductive investment of tending mothers. Our results also reveal that the extent of maternal weight gains — and thus the intensity of mother–offspring competition — was highest when mothers had featured a low weight at egg hatching. Finally, we found that maternal presence curtailed the adult body size of those offspring that overcame the lethal consequences of the competition with their mother.

We showed that high maternal weight gains during family life reduced offspring survival until adulthood. Notably, this negative effect (1) was only present in offspring that grew up with their mother, (2) arose after the end of (i.e. was not detectable during) family life, and (3) was paralleled by a generally lower weight gain of maternally tended (as compared to maternally deprived) offspring during family life. Together, these findings demonstrate that mothers competed with their offspring by consuming portions of the limited amount of food available, and thereby indirectly limited the long-term survival prospects of their progeny, e.g. by increasing the likelihood of offspring starvation or by triggering high levels of siblicide (Dobler & Kölliker, 2010). These findings are unlikely to result from the prevention of offspring dispersal in our experimental setup, since nymph dispersal in F. auricularia is not accelerated by food limitation (Wong & Kölliker, 2012). Hence, our findings emphasize that local competition between parents and their offspring might counteract the benefits of facultative care (Boncoraglio & Kilner, 2012; Scott & Gladstein, 1993; Ward, Cotter, & Kilner, 2009) and thus impede the transition to — and the maintenance of — social life in resource-poor environments.

Interestingly, maternal and offspring weight gains interacted in determining the survival of maternally tended offspring during family life: the association between offspring weight gains and survival rates shifted from positive to negative when maternal weight gains were low and high respectively. This suggest that high offspring weight gains are beneficial when mother–offspring competition is moderate (i.e. when maternal weight gains are low), but costly when it is intense (i.e. when maternal weight gains are high). The first pattern would then indicate that the offsprings' survival mostly relies on their own quality (as reflected by their weight gains) when their mother does not monopolize the food resources. Conversely, the second pattern would indicate that high levels of mother–offspring competition result in a lower amount of left-over food per offspring, which could in turn trigger higher levels of sibling rivalry and siblicide (Dobler & Kölliker, 2010; Gardner & Smiseth, 2011; Wong, Lucas, & Kölliker, 2014). Alternatively, this second pattern could reflect filial cannibalism, for instance aimed at increasing the survival prospects of the remaining juveniles (Forbes & Mock, 1998; Klug & Bonsall, 2007). In both cases, the resulting lower overall offspring survival would increase the per-capita amount of food available to — and thus the maximum weight gain achievable by — the remaining (high-quality) juveniles and their mother. The exact mechanism notwithstanding, our results show that maternal presence under food limitation can reduce offspring fitness even in the short run (see also Thesing et al., 2015).

We showed that the initial weight of the mothers, but not the initial weight of their offspring, affected the intensity of mother–offspring competition. Mothers with a low weight at egg hatching gained more weight at the expense of their offspring — and thus reduced their offspring's survival more substantially — than initially heavier mothers. Importantly, maternal weight gains did not differ between clutches of particularly heavy and light nymphs, suggesting that mothers did not adjust the extent of competition in relation to the perceived reproductive value of their offspring (Kilner & Hinde, 2012). Together with the observation that offspring overall gained less weight under maternal presence, these findings indicate that the extent of mother–offspring competition is subject to maternal control (Hinde, Johnstone, & Kilner, 2010; Smiseth, Wright, & Kölliker, 2008). Note that this holds true although initially heavy nymphs showed lower relative weight gains than initially light nymphs. This is because this latter relationship arose independent of maternal presence, indicating that it reflected intrinsic factors such as the genetic quality of family members (Wilson & Nussey, 2010) rather than the extrinsic influence of competition. Overall, our findings thus suggest that the (genetic) quality and/or the condition of mothers (see also Koch & Meunier, 2014) determine how mothers react to food limitation (e.g. by affecting their responsiveness to offspring solicitation behaviours; Grodzinski & Johnstone, 2012) and thus whether they compete with their offspring. Here, mothers of low quality or bad condition might have favoured somatic maintenance over self-restraint (Alonso-Alvarez & Velando, 2012; McNamara & Houston, 1996), either to safeguard their ability to perform crucial parenting behaviours such as predator defence (Bateson, 1994), or in an attempt to shift their investment from current to future offspring (Kramer & Meunier, 2016; Thorogood, Ewen, & Kilner, 2011).

In line with the hypothesis of a condition-dependent shift of maternal investment, we found that high maternal weight gains — and thus intense mother–offspring competition — increased the relative investment into second clutch production among mothers that produced another clutch. Conversely, high maternal weight gains promoted or inhibited the likelihood of second clutch production depending on whether these gains had occurred in concert with high or low offspring weight gains during first brood family life respectively. Finally, we found that the same pattern shaped the length of the inter-clutch interval, but only among mothers that had received a restricted (rather than an ad libitum) amount of food after the isolation from their first brood offspring. Overall, these findings illustrate that mother–offspring competition might not always trigger a shift in maternal investment from current to future reproduction (Klug et al., 2012). This supports the conjecture of an additional motivation for this competition (e.g. the maintenance of maternal condition to perform crucial parenting behaviours; Bateson, 1994), which in turn highlights that parent–offspring competition should not uncritically be taken as evidence for parent–offspring conflict (sensu Trivers, 1974). In particular, some mothers might have heavily invested into their current (low-quality) offspring, thus giving rise to the negative effect of high maternal weight gains on the likelihood of second clutch production in broods featuring low offspring weight gains. This suggests that the scope for conflict might be limited if mothers re-invest competitively acquired resources into their current brood and thereby offset the costs of competition for first brood juveniles.

In contrast to its influence on offspring survival and maternal investment into future reproduction, mother–offspring competition did not affect the size and forceps length of offspring that survived until adulthood. This result suggests that this competition likely does not reflect an indirect mechanism of adaptive brood reduction by mothers. Specifically, mother–offspring competition does not seem to have handicapped low-quality offspring to an extent that favoured the survival and development of higher-quality offspring, as the average size of maternally tended offspring should have been larger than that of non-tended juveniles in this scenario (Mock & Parker, 1997; Simmons, 1988). Nevertheless, our results showed that adults were overall smaller when they had been raised with their mother, a negative effect of maternal presence that has already been found under favourable conditions (featuring unrestricted food access) and that might result from maternal behaviours that are maladaptive under laboratory conditions (such as the burying of food, presumably to prevent microbial growth; see Thesing et al., 2015). Our results also revealed that males were overall smaller but had longer forceps than females, confirming the sexual dimorphism of these morphological traits (Radesäter & Halldórsdóttir, 1993; Thesing et al., 2015). Finally, we showed that body size generally decreased with increasing weight gains of mothers during first brood family life. This effect arose independent of maternal presence, indicating that it likely reflected the overall lower initial weight (and thus size) of offspring produced by those mothers that gained a lot of weight during family life rather than the extrinsic influence of competition.

In conclusion, we demonstrated that local competition between parents and their offspring can drastically reduce offspring fitness in species with facultative family life. Parent–offspring competition might thus not only diminish the benefits of a prolonged association of parents with their offspring (Scott & Gladstein, 1993), but even impede the evolution of family life altogether. This finding illustrates that parental presence can be associated with costs for the tended offspring (Meunier & Kölliker, 2012; Scott & Gladstein, 1993; Thesing et al., 2015) that are usually masked by the benefits of parenting behaviours in the wild (Costa, 2006; Wong et al., 2013), but that emerge whenever the (laboratory) conditions prevent these benefits from taking effect. While parents might be able to reduce the costs of local competition with their offspring under natural conditions, the behavioural changes necessary to do so are likely themselves costly. For example, parents might increase their foraging range (West et al., 2002), but will then likely suffer from a higher predation risk (Alonso-Alvarez & Velando, 2012). Our results thus generally stress the crucial role of environmental factors such as resource availability and predation pressure in the early evolution of social life. Importantly, our findings also provide a diachronic perspective on social evolution: they suggest that competition between parents and their offspring should decline with an increasing reliance of offspring on parentally provided resources, a process that in turn is known to increase the scope for competition among offspring (Gardner & Smiseth, 2011). The transition from facultative to obligatory forms of family life might hence be accompanied by a shift in the type of competition that most profoundly affects family interactions.

ACKNOWLEDGEMENT

We thank two anonymous reviewers and the editorial board for their helpful comments on an earlier version of this manuscript. This research was supported by the German Science Foundation (DFG; ME4179/1-1 to J.M.).

    AUTHORS' CONTRIBUTIONS

    J.K. and J.M. conceived the ideas and designed the experiment; J.K., M.K., J.M.C.D., C.S., A.Y.D., T.C. and P.K. collected the data; J.K. analysed the data; J.K. led the writing of the manuscript. All authors contributed critically to the drafts and gave final approval for publication.

    DATA ACCESSIBILITY

    Data available from the Dryad Digital Repository https://doi.org/10.5061/dryad.651ss (Kramer et al., 2017).

    COMPETING INTERESTS

    The authors declare to have no competing interests.