Currently used niche measures of overlap and breadth are shown to be inappropriate for analyzing field data. Previously proposed measures assume resources are equally available, but this assumption is unwarranted in field studies. Comparisons of different measures of either breadth or overlap are also not valid. A measure of niche breadth is derived by asking how likely it is that a species' utilization of resources could have been drawn from the available resources in the environment. An overlap measure is derived in a similar manner and represents the likelihood that the utilization of resources by one species is identical to the utilization by another species. Tests for comparing different measures are given.