No abstract

The fish assemblage of the Canadian River in Texas historically was dominated by Hybognathus placitus and Notropis girardi. These species represented > 90% of fishes collected fiom the Canadian River in 1954-1955. Construction of two reservoirs on the Canadian River in the 1960s (Ute Reservoir, New Mexico, 1962; Lake Meredith, Texas, 1965) altered hydrologic conditions in the river and affected the fish assemblage. Downstream fiom Ute Reservoir, mean annual discharge decreased by about 38% after impoundment, but H. placitus, N. girardi, and other mainstem species still dominate the assemblage. Downstream from Lake Meredith, mean annual discharge decreased by 76% and the historic mainstem fish assemblage has been almost completely replaced by species that formerly were restricted to tributary streams. The magnitude of postimpoundment changes in the fish assemblage of the Canadian River appears to be related to the degree that discharge has declined, especially during the spawning season.

Riverine flow regimes are naturally dynamic, but become increasingly homogenized following anthropogenic flow alteration. Loss of dynamism disrupts naturally occurring structuring mechanisms within the associated biotic communities, at times causing shifts in composition. Here we considered how stream fish assemblages in two Central Texas rivers changed after alteration of flow regime by either construction of a mainstem, deep storage reservoir or flood-retaining structures. Following impoundment, number of large and small floods increased from 0.81 to 1.07 floods per year (FPY) in the upper Guadalupe River, decreased from 0.84 to 0.42 FPY in the lower Guadalupe River and decreased from 0.87 to 0.7 FPY in the San Marcos River. Historical ichthyofauna data ranging from 1938 to 2006 were used to assess changes in assemblage composition and species abundance. Assemblages did not differ in the upper Guadalupe River (Bray-Curtis index ¼ 37.4%; ANOSIM global R ¼ 0.079, p ¼ 0.08), but did differ in the lower Guadalupe River (25%; global R ¼ 0.409, p < 0.01) and San Marcos River (27%; global R ¼ 0.19, p < 0.01). In general, habitat generalist species dominated assemblages during periods of reduced flood frequencies (i.e. drought of record; following impoundment), whereas regionally endemic species (N ¼ 3) and substrate and broadcast spawning species (N ¼ 5) declined (b 1 < 0; a ¼ 0.05). Based on the results from this study, managing flows in the lower Guadalupe River to mimic historical timing of flood pulses might attenuate contemporary disruption of natural assemblage composition.

Reduced suspended‐sediment loads (i.e., turbidity) in many Midwestern prairie rivers have been hypothesized as contributing to the replacement of species that historically occupied highly turbid main‐channel habitats by visually feeding species that are competitively superior in less‐turbid waters. We examined the relationship between prey consumption and turbidity for six fish species from the Canadian River (New Mexico, Oklahoma, and Texas) and found experimental support for this hypothesis. Among species adapted to highly turbid main‐channel habitats, we found that prey consumption by the peppered chub Macrhybopsis tetranema and flathead chub Platygobio gracilis was unaffected (P > 0.12) by elevated turbidity, whereas prey consumption by the Arkansas River shiner Notropis girardi was reduced (P < 0.01). Among species characteristic of less‐turbid habitats, prey consumption by the emerald shiner N. atherinoides, red shiner Cyprinella lutrensis, and sand shiner N. stramineus was reduced (P < 0.01) by elevated turbidity. Compared with prey consumption at 0 nephelometric turbidity units (NTU), prey consumption at 4,000 NTU decreased 21% among peppered chub, 26% among flathead chub, and 59% among Arkansas River shiners, which was less than that observed among emerald (73%), red (84%), and sand shiners (89%). In general, elevated turbidity had less effect on the prey consumption of species that are adapted to highly turbid habitats than on those characteristic of less‐turbid habitats. The high suspended‐sediment loads that historically were characteristic of many prairie streams may have excluded emerald, red, and sand shiners from main‐channel habitats.

We determined swimming ability for 37 warmwater stream fishes in Texas and Louisiana and assessed relationships among swimming ability, habitat selection, and fish morphology. Mean absolute speeds (cm/s) were 17.8–81.9 cm/s for Cyprinidae, 50.9 cm/s for Characidae, 70.0 cm/s for Ictaluridae, 33.0 cm/s for Cichlidae, 40.0 cm/s for Percidae, 30.7–43.4 cm/s for Cyprinodontidae, 30.2 cm/s for Atherinidae, 15.7–18.6 cm/s for Poeciliidae, and 23.4–40.5 cm/s for Centrarchidae. Absolute swimming speeds were correlated (P < 0.01) to reported habitat descriptions (i.e., sloughs, lakes, streams, and rivers) of fishes. Absolute swimming speeds also were good predictors of spatial (i.e., runs, pools, and riffles) distributions of fishes from Independence Creek, Texas (P = 0.03), and Banita Creek, Texas (P = 0.06). Multivariate ordination of nine morphological attributes identified the linear combination of relative body depth, dorsal fin length, and caudal span width as an adequate predictor (P < 0.01) of swimming speed; however, several exceptions were noted. Understanding interactions between current velocity and swimming ability in fishes is critical to species conservation and to maintaining instream flow requirements for fishes.