ABSTRACT: We examined larval availability and settlement of the intertidal mussel Perna perna simultaneously at different spatial and temporal scales using a nested design at 2 sites, 3 km apart on the south coast of South Africa. Each site had 3 locations (300 m apart) where 5 artificial settler collectors were placed about 20 cm apart. Collectors were replaced on temporal scales varying from fortnightly (for 16 mo) to daily (2 series of 15 to 20 d). Each intertidal location was paired with an inshore location (these too were 300 m apart) within 500 m of the shore, where larval availability was measured by 3 vertical plankton hauls collected on the same dates as for settler sampling. There was strong temporal variation in abundances of larvae and settlers, and no correlation (r always < 0.14) was found between the two. Larvae were abundant only at the start of sampling and rare for the rest of the study, while distinct peaks in settler numbers occurred later. No spatial effect was detected for larval availability, while there was strong spatial variation in settlement at the location level. These results indicate that, on scales of 100s of m to km, delivery of larvae from the nearshore water column onto the shore is strongly differential, with some locations consistently receiving more settlers than others. We conclude that, at these sites, the patchiness in settlement observed on scales of 100s of m depends on differential delivery, rather than differential offshore distribution of larvae. We suggest that differential delivery is due to the effect of nearshore bottom topography on local hydrodynamics.