SPATIAL VARIATION OF FISH COMMUNITIES IN THE UPPER MISSISSIPPI RIVER BASIN John H. Chick 1, Mark A. Pegg 2, Todd M. Koel 3, 1 Illinois Natural History Survey, Great Rivers Field Station; 2 Illinois Natural History Survey, Illinois River Biological Station; 3 National Park Service, Yellowstone National Park. Since 1991, researchers with the Long Term Resource Monitoring Program have collected fish abundance and composition data from six regional trend areas (RTA) in the Upper Mississippi River System (UMRS): Mississippi River navigation pools 4, 8, 13, and 26, the La Grange pool of the Illinois River, and an open river reach on the Mississippi River near Cape Girardeau, MO. These six RTAs were chosen to represent the range of conditions present throughout the UMRS and are meant to allow for system wide inferences about fish population and community dynamics. To test the adequacy of these six RTAs in representing the variation of fish communities present in the UMRS, additional sampling was conducted during 2000 from areas above and below (out-pools) three RTAs, and from navigation pools 19 and 20. Cluster analysis and non-metric multi- dimensional scaling of species composition (presence absence) and community structure data revealed two major spatial groups of pools, and four sub-groups. Northern and southern pools formed the two major pool groups. The sub-groups identified included open river reaches, southern pools, northern pools, and the La Grange reach of the Illinois River. Out pools generally grouped with the closest RTA, but pools 19 and 20 grouped with southern pools in terms of species composition with northern pools in terms of community structure. This analysis suggests the six RTAs may be adequate to capture the major spatial differences in fish communities within the UMRS, namely southern versus northern pools, but additional research is needed to clarify how fish communities in pools 19 and 20 contrast with other areas in the UMRS. Keywords: Upper Mississippi River System, fish communities, spatial variation, multivariate statistical analysis