|Home | About | Journals | Submit | Contact Us | Français|
Maternal effects can induce adjustments in offspring phenotype to the environment experienced by the mother. Of particular interest is if mothers can programme their offspring to cope best under matching environmental conditions, but the evidence for such anticipatory maternal effects (AME) is limited. In this study, we manipulated experimentally the food availability experienced by mothers and their offspring in the European earwig (Forficula auricularia). Offspring produced by females that had access to high or low food quantities were cross-fostered to foster mothers experiencing matched or mismatched environments. Offspring experiencing food availability matching the one of their mothers had an increased survival to adulthood compared with offspring experiencing mismatched conditions. Females experiencing high food laid larger clutches. This clutch-size adjustment statistically explained the matching effect when offspring experienced high food, but not when experiencing low food conditions. There were no effects of matching on offspring growth and developmental rate. Overall, our study demonstrates that AME occurs in relation to food availability enhancing offspring survival to adulthood under matching food conditions.
The environment experienced by the mother and her nutritional condition often influences the nature and magnitude of environmentally induced maternal effects [1,2]. Such transgenerational plasticity can be adaptive when mother and/or offspring fitness is enhanced by the environmentally induced phenotypic adjustments in offspring, or non-adaptive when experiences by the mother, such as food limitation, carry over to the next generation limiting offspring growth or survival (i.e. silver-spoon hypothesis) .
A particular type of adaptive transgenerational plasticity is when mothers programme their offspring to cope better under the environmental conditions that the mothers themselves experienced [3–6]. Under such anticipatory maternal effects (AME) , mothers transmit information about their environment to their young, resulting in a plastic adjustment in the offspring's phenotype enhancing offspring performance in a matching environment, at the risk of reducing offspring performance in a mismatching environment [3,5,7]. Mechanisms for AMEs are expected to evolve when the environment varies temporally and/or spatially, on a scale in which the mother can predict at least to some extent the environment that the offspring will probably encounter [1,3,8,9].
Evidence for AMEs are the offspring immune priming by mothers exposed to particular pathogens (e.g. [10,11]), the maternal induction of antipredator or herbivore structures or behaviours in animal and plant progeny [12,13], and the priming of offspring to food quality in butterflies . It is less known if mothers can programme their offspring in response to variation in food availability, possibly owing to the added difficulty of separating non-adaptive carry-over effects of food quantity (resource-based or ‘silver-spoon’ maternal effects) from AMEs. Still, food availability is a likely candidate for a broadly occurring agent of selection on AMEs, because it plays a central role as a limiting factor for growth, survival and reproductive success . Former studies on transgenerational effects of food availability show mixed results (e.g. [16–20]). To test AME in relation to food availabilities, experiments are needed to disentangle the food availability experienced by the mother from the one experienced by the offspring [3,5,9].
In this study, we experimentally manipulated food availabilities in two generations to investigate AME in the European earwig, Forficula auricularia, a social insect with maternal care for eggs and hatched nymphs [21,22]. Although the predictability of food availability experienced by mothers and their offspring remains to be quantified in this species, AMEs are expected because: (i) food availability is a main determinant of offspring fitness affecting offspring development, growth and survival [23,24], and (ii) dispersal is limited [25,26]. Thus, offspring tend to live in the same habitat or spatial location as their mothers and a correlation in food availability across generations seems at least plausible. We manipulated the food availability experienced by females and their offspring using a full-factorial cross-fostering design. Clutches from females reared under high or low food availabilities were cross-fostered within and between food treatments resulting in combinations, where the food treatment of the foster environment matched the food treatment of the biological mother and in combinations in which they mismatched. As maternal programming of offspring may affect offspring fitness directly by enhancing survival independently of growth and development, or indirectly by allowing offspring to grow better or develop faster, we took several measures of offspring performance across the early and late juvenile stages, including their developmental time, growth and survival . We tested whether different food availabilities in the maternal and/or foster condition affect offspring development time, growth and survival. If AME determine offspring performance traits, we predicted that offspring cope better with matching maternal and foster food conditions than mismatching ones.
The study animals were F1 and F2 descendants of earwigs collected in an olive grove in Dolcedo, Italy (7° 560550 E, 43° 540140 N, altitude 75 m above sea level) in July 2012, and bred in the laboratory under standard conditions . Females produce either a single clutch or two clutches in their lifetime, and perform facultative brood care for eggs and hatched nymphs [21,22]. Before reaching adulthood (approx. 60 days) the nymphs undergo four instar stages . The basic experimental design consisted of the manipulation of food availability across two generations, and cross-fostering eggs within and between food treatments after the first half of the egg stage to experimentally create all four maternal–offspring food treatment combinations (electronic supplementary material, figure S1).
The hatched F1 nymphs of each of 33 broods were kept as family groups with their mothers until day 16 after hatching in Petri dishes (10 × 2 cm) provided with humid sand as substrate, and plastic shelters under standard laboratory conditions . On day 16 after hatching, the mothers were removed  and the nymphs from each brood were transferred to larger Petri dishes (14 × 2 cm). They were randomly assigned to a high food treatment approximating ad libitum conditions (HF, n = 16 broods) or low food treatment guaranteeing a loss in body mass (LF, n = 17 broods) (electronic supplementary material). The latter were determined based on a previous study using a similar food regime that showed reduced body weight and reproductive success . The animals were kept as family groups until adult emergence at 70% humidity 20°C and a 14 L : 10 D photoperiod schedule (summer conditions). When reaching adulthood F1 females were set up for mating (males used for mating were from the HF treatment throughout; electronic supplementary material). After the first female was observed to oviposit, all the females were isolated for egg laying .
On day 11–14 after oviposition of F1 females, the F2 clutches were cross-fostered among females within and between food treatments (HF and LF) resulting in the four possible combinations of maternal and foster environments (HF–HF: n = 28 broods, HF–LF: n = 29 broods, LF–HF: n = 26 broods, LF–LF: n = 25 broods; electronic supplementary material, figure S1). Cross-fostering took place among a set of females ovipositing within 4 days, and clutches were randomly assigned to foster mothers, but excluding assignments to sisters.
One day after hatching of a clutch the nymphs were counted to quantify ‘hatching success’ (total nymphs at day 1 divided by clutch size) and the female and 10 nymphs were weighed to the nearest 0.001 mg (using a Mettler-Toledo MT5 Micro-balance; Mettler, Roche, Basel) to obtain the female and average nymph body mass at day 1, respectively. A measure of embryonic developmental time was calculated as the number of days between oviposition and hatching.
The nymphs and the foster female were set up in a new Petri dish, kept under summer conditions (see above) and fed according to the food treatment of the foster female. On day 16 after hatching, all surviving nymphs were counted to compute ‘early survival’ (as the number of surviving nymphs on day 16 divided by the number of hatchlings at day 1) and 10 nymphs per clutch were weighed. Moreover, ‘early developmental time’ was recorded as the number of days between hatching and the first moult to 2nd instar in a clutch (which is a good approximation for the average moult time of a clutch ). Nymphs were transferred to larger Petri dishes (14 × 2 cm), fed according to the treatment of their foster mother and reared until adulthood (electronic supplementary material, table S1).
The number of emerging adults was counted to quantify ‘late survival’ as the number of emerging adults divided by the number of nymphs on day 16, and the weight of one haphazardly chosen adult per clutch was taken. ‘Late developmental time’ was calculated as the number of days between moult to 2nd instar and the first adult emergence in a clutch.
We used linear mixed models (LMMs) with measures of survival, growth and developmental time as the dependent variables and included the maternal environment, the foster environment and their interaction as the main fixed factors. We further included clutch size and the interaction between clutch size and the foster environment as fixed terms to correct for potential differential effects of clutch size on offspring performance in the HF and LF foster environment e.g. owing to clutch-size-related sibling competition. We used the clutch-size measures at the beginning of each developmental stage of the corresponding performance measure (number of eggs for performance during egg stage, number of hatched nymphs for performance during early nymph stage (until day 16), number of nymphs on day 16 for performance during late nymph stage (until adulthood)). The mating groups were included as a random effect in the models to account for potential dependencies among females that were mated to the same set of four males. All variables were standardized to n (0/1) by subtracting from each value the variable's mean and dividing by its standard deviation. Proportional-dependent variables (hatching success, early and late survival) were logit transformed before standardization . All models were checked for normal distributions of residuals by visual inspection. Data of early developmental time were inversely transformed (‘1/x’) to achieve normality of residuals. All p-values are two tailed and estimates for the regression coefficients b are provided. All analyses where conducted using IBM SPSS® v. 21.0 (SPSS Inc., Chicago, IL, USA) software.
HF females produced significantly larger clutches than LF females (F1,106 = 19.91, p < 0.001), and they were significantly heavier at the time of hatching (F1,106 = 140.9, p < 0.001). There were no significant effects of the maternal environment, foster environment or clutch size on embryonic developmental time, nymph body mass at day 1 or hatching success (table 1).
The early developmental time of nymphs was significantly affected by the foster environment, with nymphs reaching the 2nd instar faster in the HF compared with the LF foster environment (table 1). Likewise, offspring body mass at day 16 was significantly higher in the HF foster than in the LF foster environment and decreased with clutch size (table 1). ‘Early survival’ was not significantly affected by the maternal environment. However, survival was significantly higher in the HF than the LF foster environment (table 1 and figure 1a). The interaction between the maternal and foster environment was not significant for any of the measured early nymph traits.
Late developmental time was significantly affected by the maternal and foster environment and depended on clutch size. Nymphs reached adulthood faster when originating from HF mothers compared with individuals from LF mothers. Furthermore, there was a significant interaction between the foster environment and clutch size (table 1; electronic supplementary material, figure S2). Developmental time decreased with clutch size in the HF foster treatment (Pearson's correlation: r = −0.32, p = 0.04), and increased with clutch size in the LF foster treatment (r = 0.55, p < 0.001).
The average offspring body mass at the adult stage was significantly heavier in the HF compared with the LF foster environment (table 1), but no other effects on the adult body mass of offspring were significant.
‘Late survival’ of offspring was overall significantly higher in the HF than the LF foster environment. Furthermore, there was a significant interaction between maternal and foster environment (table 1), which was owing to an effect of environmental matching on late nymph survival in the LF foster environment (figure 1b). To test effects of the maternal environment on late nymph survival separately in the two foster environments, we used t-tests to compare estimated marginal means among maternal treatments within foster treatments. In the LF foster environment, nymphs originating from LF mothers (matching environment) survived significantly better to adulthood compared with those originating from HF mothers (mismatching environment; t = 3.28, p = 0.002). However, the effect of environmental matching was non-significant in the HF foster environment (t = 1.44, p = 0.156; figure 1b).
‘Late survival’ was related to clutch size (no. nymphs at day 16) through an interaction with the foster environment (table 1). To address whether this interaction accounted for the differential matching effect in the LF versus HF foster treatment, we reran the model removing clutch size from the model. Without correcting statistically for variation in clutch size, the enhanced survival effect for matching conditions was significant in both the LF and HF foster environments (electronic supplementary material, figure S3).
In this study, we experimentally tested the potential for AME in relation to food availability. The maternal and offspring environments were disentangled using a cross-fostering experiment which allowed us to test the separate and combined effects of the maternal and foster environment on offspring development, growth and survival. The foster environment influenced all measured offspring performance traits after hatching. This was expected because it reflected the actual amount of food available for growth and development of the offspring and confirms former evidence in this species that food availability is a key limiting factor for offspring growth, development and survival . This result is also in line with a meta-analysis across studies and species that showed strong effects of the offspring food environment on performance . Conversely, the maternal food environment only affected late development: offspring reached adulthood faster when originating from HF mothers irrespective of the foster food environment.
Effects of environmental matching were observed at the level of late juvenile survival. Survival was higher when nymphs experienced the food conditions matching the one of their mothers, as predicted by AMEs. In the LF foster treatment, this effect occurred independently of variation in clutch size, which was not the case in the HF foster treatment. Overall (i.e. without correcting for clutch size), in the HF foster treatment, offspring of HF mothers had an 8% higher survival than offspring of LF mothers, and in the LF foster environment, offspring of LF mothers had an 11% higher survival than those of HF mothers (electronic supplementary material, figure S3). When incorporating clutch size in the model, only the effect in the LF foster environment remained significant. We aimed at testing the significance of environmental matching irrespective of the underlying mechanism and, thus, allowed for females to adjust clutch size in response to different food availabilities (i.e. HF females lay larger clutches, this study, ). Our results indicate that maternal clutch-size adjustment indeed plays a part mediating effects of environmental matching on late offspring survival, but also that AMEs occur independent of this effect, specifically in a poor environment. Future studies to unravel the underlying mechanisms may experimentally control for clutch size to better isolate clutch-size independent AMEs on the survival of individual offspring.
The question of whether our results were owing to a transfer of resources or information cannot be answered conclusively. If purely owing to resource transfer, we would have expected to find strong influences by the maternal food treatment on hatchling body mass. But this was not the case. Note that although we did not quantify egg size, hatchling body mass is known to be positively correlated with egg size . More generally, maternal effects owing to variation in the amount of maternally transferred resources to the eggs are usually strongest and most detectable during early juvenile development, and they tend to dissipate with time (e.g. ). By contrast, AMEs have the function to enhance offspring fitness by programming the offspring's physiological or metabolic machinery to particular food availability (e.g. through hormones and/or epigenetic modifications ). If adaptive, they should be the strongest during the most sensitive phases of offspring development, irrespective of age. Our finding that environmental matching specifically affected survival during the late juvenile stage is more consistent with this latter expectation because in F. auricularia the later juvenile stages are most sensitive to food limitation in terms of later reproductive success .
The mechanisms underlying the found AMEs on late juvenile survival in earwigs are currently unknown. Potential mechanisms include resources, hormones, transcription factors and epigenetic marks that are directly transferred from mothers to the eggs, but also physiological effects in the offspring triggered by maternal behaviour (e.g. [34–36]). In our study, the cross-fostering took place 11–14 days after oviposition, which is about halfway through the egg stage. Thus, the mechanisms underlying the observed transgenerational effect must have operated before this point in time. The foster environment had no significant effects on embryonic development and hatching success suggesting that an effect through maternal behaviour is unlikely. The found effects are more likely owing to maternal information transfer through the eggs (e.g. hormones) . Given the central role of ectysteroids and juvenile hormones in insect development and growth (e.g. [38,39]), they are candidates for the found effects requiring further study.
To conclude, our study provides experimental support for maternal clutch-size adjustments and AMEs to enhance offspring fitness under conditions of matching food availability in the European earwig (F. auricularia). It further highlights that studies on AMEs should focus on the juvenile stage that is most sensitive to the environmental factor of interest and on a range of offspring performance traits.
We thank J. W. Y. Wong, S. Boos, and L. Röllin for their help during the experiment, K. E. Thonhauser and L. Engqvist on statistical advice and B. Taborsky, J. Meunier and three anonymous reviewers for helpful comments on an earlier version of the manuscript.
All research described here adhered to local guidelines and all appropriate ethical approval and licences were obtained.
All data from this study will be publicly available in the Dryad Digital Repository, as soon as accepted.
S.R. and M.K. conceived and designed the experiments; S.R. and D.V. performed the experiments; S.R. analysed the data and S.R. and M.K. wrote the paper. All authors commented on drafts of the manuscript and approved the final version.
All authors declare they have no competing interests.
This work was funded by the Swiss National Science Foundation (grant no. PP00P3-139188 to M.K.).