In late May 1992, cod and haddock eggs and larvae were sampled at discrete depths using MOCNESS on the southern flank of Georges Bank when seasonal stratification of the water column was developing. A deeper, stratified site (80-m bottom) was sampled over 7 days in relation to the temporal evolution of vertical structure and compared to a nearby shoal, mixed site (50 m). Sampling also was conducted to the west along two transects in the Great South Channel. Larval biochemical composition was measured and nutritional condition inferred from RNA:DNA ratios.During the period 22–24 May 1992, surface warming resulted in a temperature gradient from 6 to 10°C (0.6 δt units) in the upper 20 m at the stratified site. Gadid eggs (93% haddock) were most abundant (6–14 per 100 m3, on average) in the surface 20 m, with maximum density just above the base of the thermocline at 20-10 m. The few recently-hatched larvae (5–6 mm) caught were broadly distributed in the water column, with maximum average densities (1–3 per 100 m3) deeper at 10–40 m. The naupliar and copepodite stages of Pseudocalanus spp. were principal prey of these larvae, and also were more abundant (3–7 prey 1−1) in the surface 20 m, with a maximum density at 20-10 m. A moderate storm on 25 May mixed the upper part of the water column to 5–6°C. The day after the storm, gadid eggs and copepods had highest densities in the surface 10 m.At the shoal (mixed site) egg densities were low (1–3 per 100 m3) and distributed evenly through the water column. Cod and haddock larvae were larger (7–8 mm modal length) and more abundant than at the stratified site. At night, they were evenly distributed at 4–6 per 100 m3 on average, but by day, maximum density increased to 6–16 per 100 m3 in the deepest strata sampled (40–30 m). The larger larvae at this site preyed predominantly on copepodite stages of Pseudocalanus spp., which were uniformly distributed in the water column.Of all the larvae sampled for biochemical analysis, 10% of haddock and 1.5% of cod had ratios below 4.1. Based on laboratory studies, fish larvae with RNA:DNA ratios below 4.1 are considered to have been in poor condition. There were no larvae in poor condition at the Great South Channel Transect, where their mean size was greatest; no larvae were collected at the stratified site for comparison. At the mixed site, cod larvae sampled before and after the storm had identical ratios, while haddock larvae had significantly higher ratios after the storm. Larvae from different sampling depths had significantly different ratios, inferring that they remained together long enough to acquire a unique signature.