The number of larval instars varies widely across insect species. Although instar number is frequently considered to be invariable within species, intraspecific variability in the number of instars is not an exceptional phenomenon. However, the knowledge has remained fragmentary, and there are no recent attempts to synthesize the results of relevant studies. Based on published case studies, we show that intraspecific variability in the number of larval instars is widespread across insect taxa, occurring in most major orders, in both hemimetabolous and holometabolous insects. We give an overview of various factors that have been observed to affect the number of instars. Temperature, photoperiod, food quality and quantity, humidity, rearing density, physical condition, inheritance, and sex are the most common factors influencing the number of instars. We discuss adaptive scenarios that may provide ultimate explanations for the plasticity in instar number. The data available largely support the compensation scenario, according to which instar number increases in adverse conditions when larvae fail to reach a species-specific threshold size for metamorphosis. However, in Orthoptera and Coleoptera, there are some exceptional species in which instar number is higher in favorable conditions. In more specific cases, the adaptive value of the variability in instar number may be in reaching or maintaining the developmental stage adapted to hibernation, producing additional generations in multivoltine species, or increasing the probability of surviving in long-lasting adverse conditions.

. 1. In arthropods, the evolution of sexual size dimorphism (SSD) may be constrained by a physiological limit on growth within each particular larval instar. A high SSD could, however, be attained if the larvae of the larger sex pass through a higher number of larval instars.2. Based on a survey of published case studies, the present review shows that sexrelated difference in the number of instars is a widespread phenomenon among insects. In the great majority of species with a sexually dimorphic instar number, females develop through a higher number of instars than males.3. Female-biased sexual dimorphism in final sizes in species with sexually dimorphic instar number was found to considerably exceed a previously estimated median value of SSD for insects in general. This suggests a causal connection between high femalebiased SSD, and additional instars in females. Adding an extra instar to larval development allows an insect to increase its adult size at the expense of prolonged larval development.4. As in the case of additional instars, SSD is fully formed late in ontogeny , larval growth schedules and imaginal sizes can be optimised independently. No conflict between selective pressures operating in juvenile and adult stages is therefore expected.5. In most species considered, the number of instars also varied within the sexes. Phenotypic plasticity in instar number may thus be a precondition for a sexual difference in instar number to evolve.

Different levels of sexual size dimorphism (SSD) have usually been explained by selective forces operating in the adult stage. Developmental mechanisms leading to SSD during the juvenile development have received less attention. In particular, it is often not clear if the individuals of the ultimately larger sex are larger already at hatching/birth, do they grow faster, or do they grow for a longer time. In the case of insects, the question about sexually dimorphic growth rates is still open because most previous studies fail to adequately consider the complexity of larval growth curve, the existence of distinct larval instars in particular. Applying an instar-specific approach, we analysed ontogenetic determination of female-biased SSD in a number of distantly related species of Lepidoptera. The species studied showed a remarkable degree of similarity: SSD appeared invariably earlier than in the final instar, and tended to accumulate during development. The higher weight of the females was shown to be primarily a consequence of longer development within several larval instars. There was some evidence of higher instantaneous growth rates of females in the penultimate instar but not in the final instar. Egg size, studied in one species, was found not to be sexually dimorphic. The high across-species similarity may be seen as an indication of constraints on the set of possible mechanisms of size divergence between the two sexes. The results are discussed from the perspective of the evolution of insect body size in general. In particular, this study confirms the idea about limited evolvability of within-instar growth increments. An evolutionary change towards larger adult size appears always to be realised via moderate changes in relative increments of several larval instars, whereas a considerable change in just one instar may not be feasible.

The number of larval instars varies widely across insect species. Although instar number is frequently considered to be invariable within species, intraspecific variability in the number of instars is not an exceptional phenomenon. However, the knowledge has remained fragmentary, and there are no recent attempts to synthesize the results of relevant studies. Based on published case studies, we show that intraspecific variability in the number of larval instars is widespread across insect taxa, occurring in most major orders, in both hemimetabolous and holometabolous insects. We give an overview of various factors that have been observed to affect the number of instars. Temperature, photoperiod, food quality and quantity, humidity, rearing density, physical condition, inheritance, and sex are the most common factors influencing the number of instars. We discuss adaptive scenarios that may provide ultimate explanations for the plasticity in instar number. The data available largely support the compensation scenario, according to which instar number increases in adverse conditions when larvae fail to reach a species-specific threshold size for metamorphosis. However, in Orthoptera and Coleoptera, there are some exceptional species in which instar number is higher in favorable conditions. In more specific cases, the adaptive value of the variability in instar number may be in reaching or maintaining the developmental stage adapted to hibernation, producing additional generations in multivoltine species, or increasing the probability of surviving in long-lasting adverse conditions.

Strong correlation between female body size and potential fecundity is often observed in insects. Directional selection favouring increased body sizes is thus predicted in the absence of opposing selection pressure. The evolutionary forces capable of counterbalancing such a 'fecundity advantage' are poorly documented. This study focuses on revealing the costs of large body size in the wingless females of Orgyia antiqua and O. leucostigma, two related species of lymantriid moths. Extreme behavioural simplicity of these animals allows systematic assessment of various fitness components in conditions that are close to natural. A linear relationship between pupal weight and potential fecundity was observed. This association was found to be independent of particular rearing conditions. There was no evidence that the relationship between fecundity and body mass becomes asymptotic when body sizes increases. No component of fitness showed a negative phenotypic correlation with body size; some displayed a weakly positive one. In particular, pupal mortality, adult longevity, mating and oviposition success, as well as egg hatching rate and egg size, were established as independent of body size in a series of field and laboratory experiments. There was a very high overall efficiency of converting resources accumulated during the larval stage to egg masses. With no costs of large adult size, selective forces balancing the fecundity advantage should operate in the course of immature development. The strong dependence of realized fecundity on body size is considered characteristic of the capital breeding strategy.

Summary 1.The allometric relationship between growth rate and body mass has received considerable attention but different taxa have not been equally studied. In particular, a limited amount of information is available on growth allometry of insect larvae. 2. In life-history studies, it is often assumed that insect larvae grow exponentially. This leads one to expect that potential rewards of extending growth periods are high in terms of increased adult body masses and fecundities. Therefore, it has been a challenge to find costs of large size that counterbalance the fecundity advantage of attaining larger sizes. 3. This study examines the intraspecific growth allometry of lepidopteran larvae. Original methodology is proposed to address problems arising from the complexity of the insect growth curve, and the high sensitivity of growth rates to environmental conditions. To facilitate generalizations, larvae of 11 unrelated lepidopteran species were subjected to an identical study design. 4. Instantaneous absolute growth rates of larvae were related to body size by an intraspecific allometric exponent in the range between 0·41 and 0·88. There were significant differences between the species but values of the exponent as high as 1 (exponential growth), and as low as 0 (linear growth) could safely be excluded. 5. Instantaneous relative growth rates of larvae were typically 35% lower in their last instar when compared to the penultimate one. Using the exponential growth curve (i.e. assuming the constancy of relative growth rates) in modelling insect life-histories may therefore lead to substantially biased conclusions.

Within a season, successive generations of short-lived organisms experience different combinations of environmental parameters, such as temperature, food quality and mortality risk. Adult body size of e.g. insects is therefore expected to vary both as a consequence of proximate environmental effects as well as adaptive responses to seasonal cues. In this study, we examined intraspecific differences in body size between successive generations in 12 temperate bivoltine moths (Lepidoptera), with the ultimate goal to critically compare the role of proximate and adaptive mechanisms in determining seasonal size differences. In nearly all species, individuals developing late in the season (diapausing generation) attained a larger adult size than their conspecifics with the larval period early in the season (directly developing generation) despite the typically lower food quality in late summer. Rearing experiments conducted on one of the studied species, Selenia tetralunaria also largely exclude the possibility that the proximate effects of food quality and temperature are decisive in determining size differences between successive generations. Adaptive explanations appear likely instead: the larger body size in the diapausing generation may be adaptively associated with the lower bird predation pressure late in the season, and/or the likely advantage of large pupal size during overwintering.

Sexual size dimorphism (SSD) can vary drastically across environments, demonstrating pronounced sex‐specific plasticity. In insects, females are usually the larger and more plastic sex. However, the shortage of taxa with male‐biased SSD hampers the assessment of whether the greater plasticity in females is driven by selection on size or represents an effect of the female reproductive role. Here, we specifically address the role of sex‐specific plasticity of body size in the evolution of SSD reversals to disentangle sex and size effects. We first investigate sex‐specific body size plasticity in Sepsis punctum and Sepsis neocynipsea as two independent cases of intraspecific SSD reversals in sepsid flies. In both species, directional variation in SSD between populations is driven by stronger sexual selection on male size. Using controlled laboratory breeding, we find evidence for sex‐specific plasticity and increased condition dependence of male size in populations with male‐biased SSD, but not of female size in populations with female‐biased SSD. To extend the comparative scope, we next estimate sex‐specific body size plasticity in eight additional fly species that differ in the direction of SSD under laboratory conditions. In all species with male‐biased SSD we find males to be the more plastic sex, while this was only rarely the case in species with female‐biased SSD, thus suggesting a more general trend in Diptera. To examine the generality of this pattern in holometabolous insects, we combine our data with data from the literature in a meta‐analysis. Again, male body size tends to be more plastic than female size when males are the larger sex, though female size is now also generally more plastic when females are larger. Our findings indicate that primarily selection on size, rather than the reproductive role per se, drives the evolution of sex‐specific body size plasticity. However, sepsid flies, and possibly Diptera in general, show a clear sexual asymmetry with greater male than female plasticity related to SSD, likely driven by strong sexual selection on males. Although further research controlling for phylogenetic and ecological confounding effects is needed, our findings are congruent with theory in suggesting that condition dependence plays a pivotal role in the evolution of sexual size dimorphism. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.13004/suppinfo is available for this article.