Testing Theories of Clutch Overlap with the Reproductive Ecology of Heterandria Formosa

Abstract

We used laboratory and field studies of the reproductive ecology of the poeciliid fish Heterandria formosa to examine the possible mechanisms that have driven the evolution of superfetation (simultaneous provision of maternal care to sets of dependent offspring in diverse stages of development). Increased food levels in the laboratory increased the number of broods that a female produced by increasing the level of superfetation. Brood sizes were determined by female body size and were unaffected by food level. Offspring size was unaffected by food level but varied with female age. Different females had different age—specific offspring sizes. Brood size and offspring size were not correlated, even when female size was held apart. In the field the level of superfetation increased with the total number of embryos, even when female size was held apart. Larger, older females displayed higher levels of superfetation than smaller, younger females. Superfetation appeared to increase offspring production rate without compromising brood size or individual offspring size. The correlations among reproductive traits that were displayed by H. formosa were quite different from those found in poeciliid fish without superfetation. Heterandria exhibited a demographic regime characterized by overlapping generations but high adult mortality rates relative to juvenile mortality. We used these results to test three general hypotheses for the evolution of clutch overlap. The results support Burley's (1980) hypothesis that clutch overlap is an adaptation to maximize offspring production rate in the face of limited maternal effort. We modified this hypothesis to specify production of large offspring. In poeciliids, the critical tests of this modification will involve comparing the reproductive and demographic patterns observed in species such as Heterandria formosa, which occupies an extreme position with regard to clutch overlap, with species such as Gambusia affinis, which does not display clutch overlap, yet is often syntopic with Heterandria. Similar comparative studies in demography could address the question of why clutch overlap is not more common.