The vertical migration of lobster larvae was suspected from previous neuston collections taken off southwest Nova Scotia, Canada. An electronically-controlled Tucker trawl was developed for discrete depth sampling and used to study the die1 vertical distribution of lobster larvae, together with environmental variables, over Browns Bank in late August. Significant vertical migration is demonstrated for stage I lobster, which were most frequently caught between 15 and 30 m water depth during daylight but were rarely caught below 10 m at night. Stage I lobster generally stayed below light intensities of 100 to 200 U E m-2 S-'. Stage I1 and 111 lobster were collected throughout the upper 20 to 30 m of the ocean but were too rare to distinguish statistically between day and night depth-abundance patterns. Stages I, I1 and 111 lobster were all confined to the upper mixed layer, above the thermocline, whlch varied in depth with the tidal cycle. Stage IV lobster were caught almost entirely at the surface, with no significant lfference between day and night abundances. These findlngs are novel and have important implications for the ecology and dispersal of larval Hornarus arnericanus in offshore waters.