Kurylo, van Nest, and Knepper (1997 Journal of Comparative Psychology 111 126-134) have recently shown that hooded rats are able to judge the global orientation of an array of elements if orientation is signalled by the perceptual-grouping principle of proximity, but not if it is signalled by element alignment. Using a procedure designed to overcome some potential problems with the experiment of Kurylo et al, we found the same distinction in the perceptual processing of Australian sea lions. The sea lions were able to judge the orientation of arrays containing strong proximity and similarity information, but performed at chance levels judging arrays in which element alignment signalled global orientation. Human subjects were able to judge all three pattern types quickly and accurately. This is strong evidence of a qualitative distinction in the way in which perceptual grouping operates in humans and the non-human species tested. Whether this distinction is a consequence of evolutionary or experiential factors is a question for future research, but the mere fact of a qualitative difference holds important implications for our understanding of perceptual grouping.