Studies of phenotypic selection in natural populations often concentrate only on short time periods and do not quantify selection intensities. We quantified temporal and microspatial variation in the intensities of natural and sexual selection for body size in the yellow dung fly over 2 years. Female fecundity selection intensity remained approximately constant over the season with an overall mean ± SE of 0.187 ± 0.014. Selection intensity for male reproductive success, defined as eggs obtained by mating males, did not differ from zero, indicating there was no assortative mating by size. Sexual selection intensity for male mating success favouring large males was variable but overall strong in the two years (0.499 ± 0.053 and 0.510 ± 0.051). As theoretically expected for male–male competition, sexual selection intensity increased with competitor density and reached an asymptote at about 250 males per pat; it also decreased with time in spring and increased again in autumn as a function of density. Small males had the best chance of obtaining a female at very low male densities. Greater selection intensity for large size in males than females is consistent with, and might be responsible for, the observed sexual size dimorphism in this species, as males are larger. The seasonal pattern of mean male body size (smallest at the beginning and end of the season) most likely reflects mere environmental (primarily temperature) influences on phenotypic size.

Summary 1.Maternal adult diet and body size influence the fecundity of a female and possibly the quality and the performance of her offspring via egg size or egg quality. In laboratory experiments, negative effects in the offspring generation have often been obscured by optimal rearing conditions. 2. To estimate these effects in the Yellow Dung Fly, Scathophaga stercoraria, how maternal body size and adult nutritional status affected her fecundity, longevity and egg size were first investigated. 3. Second, it was investigated how female age and adult nutritional experience, mediated through the effects of egg size or egg quality, influenced the performance of offspring at different larval densities. 4. Maternal size was less important than maternal adult feeding in increasing reproductive output. Without food restriction, large females had larger clutch sizes and higher oviposition rates, whereas under food restriction this advantage was reversed in favour of small females. 5. Offspring from mothers reared under nutritional stress experienced reduced fitness in terms of egg mortality and survival to adult emergence. If the offspring from lowquality eggs survived, the transmitted maternal food deficiency only affected adult male body size under stressful larval environments. 6. Smaller egg sizes due to maternal age only slightly affected the performance of the offspring under all larval conditions.

The allozyme genetic variability of various species is correlated with a variety of morphological, physiological and fitness-related traits. In particular, temperature can affect the fitness of insects through its influence on enzyme function. We examined the seasonal (12 days over 1 year) and daily (nine samples over each day) allozyme variation at the phosphoglucomutase (PGM) locus in one population of yellow dung flies (Scathophaga stercoraria; Diptera: Scathophagidae). PGM is of central functional importance in the mobilization of glycogen reserves for flight, and has been shown to affect larval growth at different temperatures in the laboratory. Based on a sample of over 3000 flies, we found a quadratic relationship, with a minimum at approximately 12 degrees C, between the frequency of the most common allele and temperature, primarily mediated by seasonal temperature variation. This could be caused by behavioural responses over the short-term, but over the year either variable viability or sexual selection probably operates on this locus, maintaining the existing polymorphism. These results call for further work on the functional differences between PGM allozyme genotypes.

Summary 1.We quantitatively compare the fitness function linking male mating success in the field to body size and the allometric relationship linking testis length (estimating sperm production) to body size, both empirically derived, for the yellow dung fly Scathophaga stercoraria (Diptera: Scathophagidae). 2. We find that both relationships are similarly non-linear (quadratic) and increasing. The most parsimonious interpretation of this congruence is that the evolution of testis size (and hence sperm numbers) evolutionarily 'tracks' the number of copulations a male of a given size can expect, resulting in the particular non-linear testis size-body size allometry evident. 3. We offer a general, two-tiered statistical approach for this kind of quantitative comparison.