Document Type


Publication Date



Twenty-two field sites in W-central Mississippi were surveyed for southern red- belly dace Phoxinus erythrogaster to characterize habitat and ichthyofaunal associations of these 'endangered' disjunct populations. Phoxinus erythrogaster occurred at only six sites. Data assessed similarity among all sites and described microhabitat associations of P erythro- gaster. Five species, including Semotilus atromaculatus, Campostoma anomalum, Etheostoma whipplei, P erythrogaster and Pimephales notatus, comprised 95% of the total number of in- dividuals collected at P erythrogaster sites. Standardized abundance (number per seine haul) for S. atromaculatus was consistently high among all P erythrogaster sites, whereas standard- ized abundance for C. anomalum, P erythrogaster and P notatus varied from site to site. Increased standardized abundance of P erythrogaster was inversely related to abundances of C. anomalum and P notatus and was highest in the absence of C. anomalum and P notatus. Thus competition among C. anomalum, P erythrogaster and P notatus may be linked to the limited occurrence of P erythrogaster in W-central Mississippi. A UPGMA cluster analysis of ichthyofaunal composition (chord distance) showed high similarity among the six sites where P erythrogaster was present. These sites were typically characterized as narrow meandering streams with gravel, pebble and sand substrata, slow-flowing pools with clear cool (22 C) water and steep banks. A discriminant function analysis indicated that the presence of P erythrogaster is associated with stream reaches with high proportions of gravel, detritus, snag, pebble, small cobble and small boulder (Canonical R = 0.73, P = 0.018). A test of the reliability of the predictive model resulted in 87% of the cases being correctly classified (P < 0.01). Declining habitat quality due to erosion and increased urban development was evident at a number of sites including two historic localities where P erythrogaster no longer occurs. Phoxinus erythrogaster abundance was low along degraded stream reaches and typi- cally higher in areas where habitat has been less impacted. Clustering of sites where P ery- throgaster was present, based on habitat similarity (chord distance), was not as strong as similarity based on ichthyofaunal composition. There was significant congruence between the two dendrogram matrices (Mantel z = 314.69, P = 0.014) but the relationship between corresponding similarity values was low (r,, = 0.334). Thus, the ability to predict sites similar in ichthyofaunal composition based on habitat similarity was low. This suggests that although specific habitat features are important to P erythrogaster, additional factors such as declining habitat quality and interspecific competition may be influencing the occurrence of P ery- throgaster in W-central Mississippi.

Journal Name

American Midland Naturalist


Copyright permission granted by The University of Notre Dame.