Egg size has been shown to vary among females in many species of marine invertebrates and fishes. Usually, however, only a single reproductive event is measured for each female, leaving open the possibility that egg size could vary significantly among broods from the same female. Here I measured egg size from 3 to 7 broods per female from 2 calyptraeid species, Crepidula atrasolea and C. ustulatulina, in 2 experiments. In a preliminary experiment, females were raised from hatching under constant temperature in incubators; in a second larger experiment, they were raised under varying temperatures on the laboratory bench top. Under both conditions, there were significant differences in egg size among broods from each female, as well as significant differences among females from the bench top. Estimates of repeatability, which give an upper boundary on the heritability of a trait, were significantly reduced when among brood variation was included in the calculations, suggesting that caution should be used in interpreting among-female components of variation in studies that measure only a single brood per female. Egg size was not related to female size, and there was an effect of brood order on egg size in only C. ustulatulina from the bench top. Eggs produced in the incubators were larger than those from the bench top and they were more variable within broods, resulting in a lower repeatability of egg size from incubator animals than from bench top animals. The significant variation in egg size at all levels of the analysis suggests that such variation should be included in models of egg size evolution, and that the evolutionary dynamics of egg size may be more complicated than reflected in current models of life-history evolution.