Emigration and immigration of decapod larvae from estuaries depend on timing of larvae occurrence in the water column relative to the tidal, tidal amplitude and day cycles. The phase relation of these natural cycles varies with tidal regime and geographically, resulting in different time-patterns of hatching of first stage larvae and of presence of late stage larvae in the water column. Vertical migration behaviour according to phase of tide also controls transport inside estuaries. These mechanisms were investigated in a field study conducted on the northwest coast of Portugal where neap ebb tides occur during the night around the quarters of the moon. Flux of decapod larvae through one sampling station was measured during one lunar month at the Canal de Mira (Ria de Aveiro) in the spring of 1990. The sampling programme was comprised of a set of 25-h fixed station studies, separated by 25-h intervals during which no sampling took place. Plankton samples were collected with a pump every hour at three depths. Current velocity and direction at the standard depths, as well as height of the water column, were also measured every hour. Hourly instantaneous flux of larvae through a 1-m-wide vertical section of the Canal de Mira was calculated for the most abundant forms. A total number of 13 combinations of species and larval stages were analyzed, belonging to the families Atelecyclidae, Pirimelidae, Portunidae, Pilumnidae, Grapsidae, Palaemonidae, Crangonidae and Thalassinidae. Patterns of net larval flux along the lunar month could be grouped into three types. Type 1 includes first zoeas that were consistently exported to the sea. Type 2 comprises late zoeas, megalops and juveniles that were consistently imported into the estuary. First zoeas that were imported during some of the 25-h studies but were exported during the others were included in Type 3; in species of this type import periods appeared to alternate with export periods according to lunar phase. Flux of Type 1 larvae followed a semi-lunar pattern. Release activity of Type 1 zoeas took place during the night and started during neap tides around the quarters of the moon, but maximum releases occurred 3–4 h after high tide of average amplitude tides, 3–4 days after the quadratures. These observations agree with the hypothesis that hatching is timed to occur on ebb tides of the largest possible amplitude so that larvae are easily dispersed from areas with a high density of predator fishes. However, based on other observations on the Portuguese coast, it cannot be ruled out that hatching might depend on a minimum number of hours of darkness experienced by the females. Larvae included in Type 2 comprise forms that may have been retained inside the estuary for the entire larval phase, as well as one form that was imported from shelf waters. No semi-lunar pattern of import was detected in this last form. Fluctuations of net flux observed in Type 3 larvae, as well in other forms that were not included in any of the types, were more difficult to explain. These larvae were first zoeas of species belonging to different taxonomic, morphological and ecological groups and may show a diversity of adaptations to the way of life of the adults. Imports and exports of larvae depended not only on time-patterns of abundance, but also on time-patterns of larval vertical distribution. As a general rule, larval stages showed patterns of depth distribution that were consistent with vertical migration rhythmic behaviours synchronized with the tidal cycle. Though the effect was not always statistically significant, first-stage larvae were closer to the surface during ebb, especially during the night, enhancing seaward transport. On the contrary, later zoeal stages, megalops and juveniles were usually closer to the surface during flood, suggesting migration to the water column during this phase of the tide and landward transport.