Waterfowl pair and brood counts and estimates of total brood utilization were obtained from 10 beaver (Castor canadensis) flowages in north-central Minnesota and compared with rank correlation techniques. Summing the data for each species, correlation between pair and brood censuses was significant ( p = 0.8304, P < 0.025). Similarly, using the flowages as replicate samples of a habitat type, correlation between pair and brood censuses was again significant (p = 0.6076, P < 0.05). These data suggest that both pair and brood censuses are suitable for developing indices to waterfowl recruitment on small wetlands in forested areas. Neither method, however, correlates significantly with total brood use of the flowages, a phenomenon not related in a cause-and-effect way to indices of population recruitment. Evaluation of waterfowl populations on wooded impoundments and beaver flowages is difficult, but amenable to ground census techniques. The demands of resource management may soon justify application of such methods to areas that cannot be adequately assessed by aircraft.