Protandry, here defined as the earlier emergence of males, is a common feature in life histories and could be the result of sexual selection on males to maximize matings, or alternatively an incidental by-product of other selection pressures on the sexes. If protandry is selected for per se, theory predicts that it should be associated with seasonal environments where there is little overlap between generations. The degree of protandry should be insensitive to environmental conditions. Moreover, on the assumption that males and females grow at the same rate as larvae, a trade-off between development time and size is expected to result in a strong association between protandry and female-biased sexual size dimorphism. These predictions were tested by a combination of comparative and experimental studies on five populations of the speckled wood butterfly, Pararge aegeria, from central and south Sweden, England, Spain, and the island of Madeira. Protandry was associated with seasonal environments, as it was only exhibited in the three northernmost populations. Protandry in these populations remained largely constant in a variety of temperatures, both under direct development, when protandry results from a sex difference in development time through the egg, larval, and pupal stages, and under diapause development, when it results from a sex difference in pupal development time only. These results indicate that protandry is selected for per se through sexual selection in seasonal environments. Similar female-biased size dimorphism occurred in protandrous and non-protandrous populations alike, and hence sexual size dimorphism in P. aegeria is not a result of selection for protandry, nor the causal factor behind protandry. Protandry and sexual size dimorphism appear to be largely decoupled traits in the life history evolution of P. aegeria. This is achieved by means of variation in pupal developmental time and variation in the relative growth rates of the sexes. Variation in growth rates is likely to be a general phenomenon and may make possible independent optimization of size and development time (age at sexual maturity), and accordingly influence expected patterns of size-related trade-offs.
