It is assumed that many insects oviposit on food patches which present limited resources to developing offspring, and that this factor plays a role in determining the number of eggs a female lays per food patch. The role of such limited resources in the number of eggs laid by the bruchid beetleStator limbatus(Horn) on seeds ofCercidium floridum(Benth.) in the Sonoran Desert was assessed. Distributions of eggs per seed from field samples were examined first. These distributions were significantly different from Poisson distributions, suggesting nonrandom oviposition. On intact seeds with eggs, the most common number of eggs was 3 and 93% of the seeds had 1–10 eggs. The eggs on these seeds suffered various mortalities. The egg parasiteUscana semifumipennis(Trichogrammatidae) attacked eggs density-dependently (frequency of attack = 0.02 eggs per seed + 0.2). Density-independent mortality factors included embryonic deaths (0.20 ± 0.02 of all eggs laid), death of first instal's at the seed coat (0.31 ± 0.06 of larvae), and death of larvae inside the seed (0.12 ± 0.03 of larvae that entered the seed). Given these factors, the optimal number of eggs per seed that a female should lay to maximize the yield of offspring from a seed is 15, which is significantly greater than the number of eggs per seed commonly observed. At such high densities of eggs per seed, larval competition was not a mortality factor that could favor fewer eggs per seed. Very few larvae ever survived parasitism and embryonic development, and penetrated the seed coat. Hence, larval competition was minimal. Females laid few eggs per seed in the laboratory, as observed in nature, when seed availability was increased. It is suggested that oviposition byS. limbatusis determined by egg parasitism favoring few eggs per seed and seed availability. The implications of egg parasitism rather than larval competition in determining clutch size is also discussed.