Few studies of diving predators have explicitly addressed the 3-dimensional nature of interactions between predators and prey at the spatial and temporal scales relevant to an individual predator's search behavior Here, we present a new method for examining such interactions using the results from an acoustic survey of krill availability to foraging penguins Analyses of fine-scale krill distributions within a 1852 X 1852 X 50 m volume of ocean revealed substantial prey patchiness in all dimensions. Our survey detected the presence of at least 6 krill aggregations in the survey area. The surface distribution of penguins was associated with the edges of these aggregations and was nonrandomly associated with kl-ill dellsitirs above 0.1 krill m-' in the 30 to 40 m depth layer. The latter association was masked when krill abundance was integrated over the entire water column. Given that lilean daytlme dlve depths for chinstrap penguins fall betwccn 30 and 40 m, our data suggest penguins may fail to detect or choose to pass by shallow, denser prey aggregations and successfully forage on deeper, more homogeneously distributed krill offering higher encounter probabilities per unit volume searched. Thrse findings reveal biologically important features of prey patchiness that cannot be addressed within the limitations of a primarily 2-dimensional analysls of predator-prey distributions. We emphasize that if we hope to gain a predictive understanding of the foraging behavior of dlving predators, then we must consider fine-scale, 3-dimensional patterns of prey patchiness when assessing the availability of prey to diving predators.