To maximize long-term average reproductive success, individuals can diversify the phenotypes of offspring produced within a reproductive event by displaying the ‘coin-flipping’ tactic. Wild boar (Sus scrofa scrofa) females have been reported to adopt this tactic. However, whether the magnitude of developmental plasticity within a litter depends on stochasticity in food resources has not been yet investigated. From long-term monitoring, we found that juvenile females produced similar-sized fetuses within a litter independent of food availability. By contrast, adult females adjusted their relative allocation to littermates to the amount of food resources, by providing a similar allocation to all littermates in years of poor food resources but producing highly diversified offspring phenotypes within a litter in years of abundant food resources. By minimizing sibling rivalry, such a plastic reproductive tactic allows adult wild boar females to maximize the number of littermates for a given breeding event.
In unpredictable and variable environments, optimal reproductive tactics of iteroparous organisms should minimize variance in reproductive success among years to maximize long-term average reproductive success [1,2]. To minimize among-year variation in reproductive success, individuals can minimize the variance in the number of offspring produced at each reproductive event (bet-hedging sensu ). Bet-hedging includes two non-exclusive mechanisms, risk-spreading and risk-minimizing. Producing the same limited number of offspring each year corresponds to risk-spreading. To minimize variance in reproductive success, individuals can also diversify the phenotypic quality of offspring produced at a given reproductive event (coin-flipping sensu ). Such developmental plasticity is widespread in invertebrates, fishes, amphibians and reptiles but not common in homeotherms .
Evidence of coin-flipping within a given reproductive event in warm-blooded species has been reported only in wild boar (Sus scrofa scrofa) . Large wild boar females produce offspring with highly diversified phenotypes within a litter, whereas small females display the individual optimization tactic [6,7] by increasing litter size with mass and keeping fetus mass constant . Coin-flipping in wild boar is thus a tactic which depends on female size. By feeding on forest mast, the abundance of which fluctuates greatly among years, wild boar face highly variable and unpredictable environments [8,9]. Therefore, we expect that fluctuating food availability should influence the magnitude of offspring phenotypic variance within a litter. Such a prediction has never been investigated for any species reported to display coin-flipping.
Taking advantage of a long-term detailed monitoring programme for a heavily hunted population, we aim to fill the gap by testing whether female wild boar produce littermates with increasingly variable phenotypes when there is increased masting.
2. Material and methods
(a) Study site and data collection
This study was conducted on a wild boar population in northeastern France in the 11 000 ha forest of Châteauvillain-Arc-en-Barrois. This forest is mainly composed of oak (Quercus petraea), beech (Fagus sylvatica) and hornbeam (Carpinus betulus). Population size of wild boar fluctuated between 1200 and 1500 individuals over the course of the study . Wild boar have no natural predators but are heavily hunted each year between October and February (see  for estimates of mortality owing to hunting). The hunting pressure was mainly oriented towards juveniles . Between 1995 and 2009, we recorded the age class of each female shot (juvenile (less than 1 year of age) versus adult (older than 1 year of age)) based on tooth eruption pattern . We examined uteri for the presence of fetuses. Changes in resource availability did not influence the fetal sex ratio . Litter size was recorded and each fetus was weighed, measured (crown–rump length, in millimetres) and sexed. Measurements of 1743 fetuses from 319 females were collected during 15 hunting seasons (figure 1a).
(b) Mast production
Each year, food availability was measured indirectly through diet composition, using the analysis of stomach contents during the hunting period . We recognized five categories depending on the quantity of beechnuts and acorns found in the stomachs (see  for further details, figure 1b).
(c) Statistical analyses
Not all litters were at the same gestation stage when shot, because mating of wild boar occurs throughout the year . To correct fetus mass among litters by gestation stage, we standardized all fetuses at 110 days of gestation .
To assess whether females produced fetuses which were more diverse in terms of mass with increasing mast production, we first provided a measure of within-litter variation in mass by calculating the coefficient of variation (CV) of fetus mass corrected for gestation stage for each litter. We then fitted a generalized least squares (GLS) framework linking the CV of corrected fetus mass as a response variable to the fixed effects of age (categorical variable with two classes), mast production (treated as a continuous variable to describe the continuum of food resources availability) and their interaction.
We checked the expectation that fetuses are not heavier when food resources are abundant. We used linear mixed models with individual fetus mass corrected for gestation stage as the response variable, and mother identity as a random effect to account for the non-independence of fetuses within a litter. We then tested the fixed effects of age and mast production, and their interaction. We assessed whether larger litters are produced in years with abundant food. We fitted a linear regression including litter size as the response variable and mast production, age and their interaction as explanatory variables.
When necessary, we applied a correction for heteroscedasticity by using the ‘varIdent’ variance function  to account for different standard deviations among mast productions (performed with R v. 2.12.2 ). Data deposited in the Dryad repository: doi:10.5061/dryad.8hf1c.
The CV of fetus mass was affected by a positive interaction between mast production and age (slope = 0.003 (s.e.: 0.0009); ). In juvenile females, the CV of fetus mass increased with mast production (; figure 2), but the slope was weak (0.001 (s.e.: 0.0004)). A much stronger positive relationship between within-litter variation in fetus mass and mast production occurred in adult females (slope: 0.005 (s.e.: 0.0005), ; figure 2).
We did not find any influence of mast production and age on fetus mass (table 1). A weak positive effect of mast production occurred on litter size for both juvenile and adult females (table 1). Both female categories increased their litter size only slightly in years of abundant food compared with non-masting years (from 3.65 to 4.32 in juveniles and from 5.35 to 6.02 in adults).
Nutrition often shapes observed variation in growth of warm-blooded species . Using a direct measure of nutrition for wild boar, we found that fetus mass and litter size are not markedly dependent on mast production. Such a noteworthy lack of effect of food resources on offspring mass and litter size does not mean, however, that wild boar reproductive tactics are independent of food. Mast production strongly influenced the within-litter variation in fetus mass in adult females (figure 2).
By contrast, juvenile females produced fetuses of similar mass independent of food availability. Juveniles usually do not reproduce in wild boar and did here in response to the high hunting pressure [14,18]. Having reached only 33–41% of their full body mass , juvenile females have to allocate a large amount of energy to both growth and reproduction. This constraint might explain why juvenile females did not adjust the phenotype of their offspring to available resources.
Adult females diversified the phenotype of their offspring in masting years. Competition between siblings might be involved in this diversification of offspring phenotype. Sibling rivalry is common among species where young share the same litter . In wild boar, variation in milk availability at different teats (teat order effect ) leads littermates to compete for the most productive teats [20,21]. Large offspring have an advantage over small ones in this competition . By producing highly diversified offspring phenotypes, adult females match the mass variation of their offspring with variation in productivity among teats, leading to decreased sibling rivalry and thereby increasing the chance of rearing many offspring at a given breeding event. Adult females displayed a coin-flipping tactic involving the diversification of phenotypes within a litter only in masting years, while maintaining a constant mean fetus mass independent of resource availability. In masting years, females thus produce large fetuses requiring more energy allocation and also small ones requiring less allocation but likely to survive under good food conditions. On the contrary, in non-masting years, small fetuses are no longer viable , thus leading adult females to equi-allocate to offspring by producing fetuses of similar mass.
This study provides further support that wild boar females exhibit a unique life-history strategy among ungulates  by displaying different reproductive tactics to maximize the number of viable offspring in variable environments.
M.D. is supported by a PhD scholarship from the French Ministry of Higher Education and Research. M.G. is supported by a grant of the French National Agency for Wildlife (ONCFS).
We are grateful to all those who helped collecting harvested wild boars, particularly P. Van den Bulck and G. Corbeau. We are grateful to the Office National des Forêts and to F. Jehlé, who allowed us to work on the study area. We thank two anonymous referees for constructive comments on a previous draft.
- Received May 4, 2013.
- Accepted July 11, 2013.
- © 2013 The Author(s) Published by the Royal Society. All rights reserved.