Observations of animal distributions have revealed that population densities have tracked resource quality. Various models have been proposed to account for such “resource matching.” For example, the Ideal Free Distribution (Fretwell and Lucas 1970) model employs habitat selection rules which assume mobile animals evaluate available habitat patches and select the highest quality patch first. I examined movement patterns of northern crayfish (Orconectes virilis) in response to habitat patches of different quality to test this assumption. I found that animals were more likely to leave a low quality patch than a high quality patch even when there were no other patches available, suggesting that the quality of encountered resources plays a significant role in subsequent decisions about movement. However, many animals did not leave the first patch they encountered, even when better habitat existed elsewhere and was well within their ability to sample it. Finally, not all crayfish selected the best site they encountered. My studies demonstrate that the response to resources is complex, and IFDs and other distribution models may rely on overly simplified assumptions about habitat selection behavior.