Larval supply affects the structure of marine communities. Cross-shelf transport of larvae by upwelling and downwelling may cause variation in larval supply. We tested the hypothesis that slow-swimming bivalve larvae are swept offshore during upwelling and shoreward during downwelling. We sampled a transect at Duck, North Carolina during a period when currents shifted from upwelling to downwelling and back to upwelling. During each shift, nearshore water was exchanged with offshore water and currents were 10 to 100 times faster than larval swimming speeds. Larval Spisula solidissima and Ensis directus were found below the thermocline and, contrary to prediction, were swept onshore during upwelling and offshore during downwelling. When S. solidissima larvae were found above the thermocline, cross-shelf transport was as predicted. Larval Tellina spp. and Mulinia lateralis remained within 5 km of shore despite cross-shelf currents and the exchange of nearshore waters with offshore waters. They did not behave as passive particles; they were not swept offshore by upwelling or onshore by downwelling. For these taxa the hypothesis was rejected. These larvae may have remained close to shore by using behaviors analogous to those displayed by animals concentrated at convergent fronts. Given the relatively slow swimming speed of bivalve larvae, there is no reason to expect that any larval type is swept offshore by upwelling. The effect of upwelling and downwelling on larval distributions varies with larval behavior and vertical distribution. Without careful sampling, one cannot invoke offshore transport of larvae by upwelling as a cause of variations in larval settlement.