Basically, the evolution of butterfly size may be understood in terms of economy and investment of the energy that members of a species will predictably be able to process, involving:

(1) the amount of resources available per unit time, where relevant resources take the shape of food (foodplant characteristics, perhaps size or biomass per unit area), and climate (temperature and water, the latter probably with an indirect effect on size through foodplant ecology);

(2) the way in which the energy invested in reproduction is shaped as egg size, egg number (fecundity, partially resulting from the negative interspecific allometry between egg size to adult size), and egg concentration in space and time (clustering).

In support of this idea, the interspecific relationships between egg size and body size in butterflies (Papilionoidea and Hesperiidae), and between size and egg and larval development time, larval trophic specificity, foodplant structure, climate, and phenology were investigated based on a sample of more than 1180 species. The independent contrasts method was used to avoid taxonomy-dependent results. Egg size is allometrically related to adult wing length by a slope of 0.43.

Based on a subset of species, fecundity is correlated to adult body size, and there is evidence for a compromise between egg number and egg size (relative to adult size) across species.

Butterfly size increases in correlation to the mean annual temperature of the species geographic range, but decreases in relation to increased aridity (or the length of the dry season). Larger butterflies tend to have longer larval development times, use large or structurally complex host plants, and are more likely to lay their eggs in batches, irrespective of climate. Larger
eggs tend to develop more slowly, and give rise to larvae with longer developmental periods that will result in larger adults. No evidence was found to support a relationship between butterfly body size and polyphagy. A complex pattern of interrelationships links body size (and egg size) to other traits, although correlations other than that between egg size and body size are generally low. The results suggest the necessity of separating climate and seasonality into components that are relevant to insect life histories in comparative studies.

http://uam.es/personal_pdi/ciencias/egb/download/bodysize.pdf