Abstract Factors that affect recruitment of sea lampreys Petromyzon marinus are not well understood; for example, the majority (85%) of sea lamprey eggs are washed out of the nest, and the survival rate of these eggs is unknown. We examined the role of predation and substrate on egg survival in the laboratory and egg predation and dispersion of eggs outside the nest in the field. Substrate type did not influence the numbers of eggs consumed by predators, but the proportion of available eggs consumed was higher for crayfish Orconectes spp. (74%) and creek chub Semotilus atromaculatus (47%) than for white suckers Catostomus commersonii (9%) and logperch Percina caprodes (3%; F = 11.12, df = 3, P < 0.0001). Survival of eggs incubated on silt (69.2%) and sand (50.8%) in the laboratory was significantly higher than survival on gravel (19.1%). Stream sampling indicated that sea lamprey egg density outside of nests on gravel was low (1.9 eggs/m2) compared with egg density in depositional areas (105.7 eggs/m2); eggs sampled outside the nest had near zero viability. In exclosure boxes seeded with sea lamprey eggs and deployed on gravel and silt for 7–10 d, egg retention was significantly greater on spawning gravel (31 ± 3%; mean ± SE) than on silt (6 ± 3%; P ≤ 0.0001); egg retention was affected by the presence of a 0.5-cm-mesh cover. Stream discharge appeared to increase egg loss from nests. We conclude that predation rates on sea lamprey eggs outside the nest are high and that substrate type alone has little impact on egg hatching success. Given the low retention of eggs in nests, egg mortality is probably a primary determinant of year-class strength. Production of sea lampreys outside the nest likely occurs under low stream discharge conditions from eggs deposited on spawning gravel.