Identifying sources of phenotypic variability in secondary sexual traits is critical for understanding their signaling properties, role in sexual selection, and for predicting their evolutionary dynamics. The present study tests for the effects of genotype, developmental temperature, and their interaction, on size and fluctuating asymmetry of the male sex comb, a secondary sexual character, in Drosophila bipectinata Duda. Both the size and symmetry of elements of the sex comb have been shown previously to be under sexual selection in a natural population in northeastern Australia. Two independent reciprocal crosses were conducted at 25° and 29°C between genetic lines extracted from this population that differed in the size of the first (TC1) and third (TC3) comb segments. These temperatures are within the documented range experienced by the species in nature. Additive and dominance genetic effects were detected for TC1, whereas additive genetic, and Y-chromosomal effects were detected for TC3. TC2 and TC3 decreased sharply with increasing temperature, by 10% and 22%, respectively. In contrast, positional fluctuating asymmetry (PFA) significantly increased with temperature, by up to 38%. The results (1) document an important source of environmental variance in a sexual ornament expected to reduce trait heritability in field populations, and thus act to attenuate response to sexual selection, (2) suggest that variation in ornament size reflects differences in male condition; and (3) support the general hypothesis that asymmetry in a sexual ornament is indicative of developmental instability arising from environmental stress. The “environmental heterogeneity” (EH) hypothesis is proposed, and supportive evidence for it presented, to explain negative size-FA correlations in natural populations. Data and theory challenge the use of negative size-FA correlations observed in nature to support the FA-sexual selection hypothesis, which posits that such correlations are driven by differences in genetic quality among individuals.