Male sexual ornaments often evolve rapidly and are thought to be costly, thus contributing to sexual size dimorphism. However, little is known about their developmental costs, and even less about costs associated with structural complexity. Here, we quantified the size and complexity of three morphologically elaborate sexually dimorphic male ornaments that starkly differ across sepsid fly species (Diptera: Sepsidae). Male forelegs range from being unmodified, like in most females, to being adorned with spines and large cuticular protrusions.  The 4th abdominal sternites are either unmodified or are converted into complex de novo appendages. Male genital claspers range from small and simple to large and complex (e.g. bifurcated). We tracked the development of 18 sepsid species from egg to adult to determine larval feeding and pupal metamorphosis times of both sexes. We then statistically explored whether pupal and adult body size, ornament size, and/or ornament complexity are correlated with sex-specific development times. Larval growth and foraging periods of male and female larvae did not differ, but the time spent in the pupal stage was ca. 5% longer for sepsid males despite emerging 9% smaller than females on average. Surprisingly, we found no evidence that sexual trait complexity prolongs pupal development beyond some effects of trait size. Evolving more complex traits thus does not incur developmental costs.

Sex-specific larval and pupal durations of 18 sepsid fly species were measured for all ca. 1,500 individuals raised in the R. Meier Laboratory of the National University of Singapore at constant humidity and temperature (25C). Subsequently, the morphological complexity of three male sexual ornaments was estimated for a subset of 5–20 (on average 10) adult males and females derived from several cohorts per species so as to represent the entire intra-specific variance in development time and body size. These specimens were dissected (removal of fore-femurs, sternites, and claspers for males; fore-femora and sternites for females) and imaged with a digital camera at high resolution. The images were then traced to produce detailed two-dimensional outlines of all the above-named traits. Individual trait size was measured by its total pixel count from each image using Adobe Photoshop CS6. The file contains overall species means of all traits, which were further also expressed relative to the corresponding female trait to correct for dimorphism.

Microsoft Excel and/or any statistical software (R, SPSS, SAS, etc).