Fish assemblage structure in lotic environments is a product of interactions between the habitat and the biota, but little is known about how deep pool habitat conditions affect distributional patterns of fish occupying them in larger warmwater streams. This study describes relationships between the habitat and the fish assemblages in deep pools of the Upper Tombigbee River, Mississippi. Pools exhibited an increase in size from headwaters to mouth. The change in the structure of fish assemblages was related significantly to increases in pool size while independent of time or other environmental conditions. A small amount of the variation in structure of fish assemblages in deep pools was accounted for by the measured environmental variables. This suggests other factors such as biotic interactions play an additional role in the forming the observed distributional patterns in fishes occupying deep pools.

The diversity of fish species found in warmwater stream systems provides a perplexing challenge when selecting species for Instream Flow Incremental Methodology (IFIM) studies. An often-suggested approach has been to use habitat guilds to incorporate the diversity found in these systems. My goal is to determine the feasibility of developing habitat suitability criteria (HSC) for the entire fish assemblage in the North and South Fork Shenandoah River, Virginia, using habitat guilds. I examined the strengths and weaknesses of direct underwater observation via snorkeling and throwable anode electrofishing to sample fish habitat use (e.g., depth, velocity, distance to cover, dominant and subdominant substrate, cover, and embeddedness) indicates that using the data collected from both techniques may produce better criteria than using just one of the two sampling techniques. To develop habitat suitability criteria using habitat guilds I placed each species a priori into a guild based on a hypothesized guild structure. Transitional life stages with significantly different habitat use were placed separately into the guild structure. The four guilds (riffle, fast generalist, pool-run, and pool-cover) were found to be significantly different from each other using the data collected for the species assigned to the guilds. Criteria were then developed for representative species from each guild and the entire guilds. Criteria developed for depth, velocity, Froude number, cover, distance to cover, substrate, and embeddedness were used to estimate a habitat response function (i.e., the relations between usable habitat and stream flow) for a representative species from each guild, the guild itself, and for a second species from each guild for comparisons. Both the representative species and guild criteria showed similar habitat response functions for the riffle guild, fast generalist guild, and pool-run guild. However, neither set of criteria performed well for the pool-cover guild. For guilds, other than pool-cover, either the guild or the representative species approach may be a viable option to developing habitat suitability criteria. The transferability tests were performed to determine if criteria developed in the North Fork Shenandoah River, Virginia would transfer to the South Fork Shenandoah River, Virginia. Only criteria for the margined madtom (Noturus insignis) and the juvenile smallmouth bass (Micropterus dolomieu) transferred for both suitable and optimal habitat. Criteria for mottled sculpin (Cottus bairdi), Cyprinella sp. (spotfin and satinfin shiners), river chub (Nocomis micropogon), adult and juvenile redbreast sunfish (Lepomis auritus), and adult smallmouth bass did not transfer. Only the pool-cover guild criteria transferred for both suitable and optimal habitat, while riffle guild, fast generalist guild, and pool-run guild criteria did not transfer. I recommend the use of site-specific criteria for the South Fork Shenandoah or different variable combinations. / Master of Science