Community Ecology of Stream Fishes: Concepts, Approaches, and Techniques

<em>Abstract</em>.—Studies demonstrating the mechanisms regulating biodiversity are uncommon. Streams and rivers worldwide display a longitudinal gradient in fish biodiversity, and most prior research has used correlative evidence to infer that higher downstream diversity is produced by factors facilitating greater niche separation. We combine 20 years of fish abundance samples from a representative southern Appalachian stream with critical swimming velocity experiments to provide direct evidence that a shifting hydrodynamic barrier affects this gradient in Coweeta Creek. We observed increased diversity in multiple sites, produced by species immigrating upstream during periodic droughts (1985–1988, 1999–2002) and a highly significant positive relationship (<em>r^</em>2 = 0.77) between drought (Palmer index) and Shannon-Weiner diversity. Resident fish generally had smaller standard lengths during drought periods. Critical velocity measurements showed that residents could tolerate faster water velocities than drought immigrants and that upstream velocities in nondrought years were faster than those in some downstream sites. These data support the hypothesis that local fish diversity in this system is limited by the ability of drought immigrants to pass an upstream hydrodynamic barrier. Alternative hypotheses such as temperature differences between sites, increased capture efficiency during droughts, and increased productivity during droughts were not supported by the data. The relationship between drought and diversity in this system is counterintuitive because drought should reduce resource availability and produce lower rather than higher diversity. These results highlight the important relationship between natural flow variation and maintenance of local diversity and demonstrate the need for long-term ecological data.

Swimming performance of delta smelt: maximum performance, and behavioral and kinematic limitations on swimming at submaximal velocities.

Swimming performance, measured as critical swimming velocity (Ucrit) and endurance, and swimming behavior and kinematics were measured in delta smelt Hypomesus transpacificus, a threatened estuarine planktivore. Most fish (58 % of the Ucrit test group) were capable of achieving and sustaining moderately high velocities: mean Ucrit was 27.6&plusmn;5.1 cm s-1 (s.d.). Ucrit was not affected by either acclimation temperature (12&shy;21 &deg;C) or fish size (3.2&shy;6.8 cm standard length) and was generally comparable with values measured for other similarly sized fishes. The remaining 42 % of the fish failed to swim at velocities above 10&shy;15 cm s-1. Interestingly, of the fish that provided a Ucrit measurement, 62 % experienced at least one temporary swimming failure between 10 and 20 cm s-1. Endurance was highly variable and, for all velocities, not normally distributed; the only significant decrease, from 6 h to 64 min, occurred between 10 and 15 cm s-1. Kinematic analyses of stroke frequency, stroke amplitude, stride length, glide frequency, glide duration, proportion of time spent stroking and the number of strokes between successive glides showed that delta smelt employed three velocity-dependent swimming gaits: a discontinuous 'stroke-and-glide' swimming behavior below 10 cm s-1; a continuous swimming behavior above 15 cm s-1 and up to Ucrit; and a discontinuous 'burst-and-glide' swimming behavior at velocities above Ucrit. Swimming failure at velocities between 10 and 20 cm s-1 coincided with the transition from 'stroke-and-glide' swimming to continuous swimming; delta smelt were unable or unwilling to swim steadily in the flume within this transition velocity range. These results underscore the importance of monitoring and quantifying behavior in experiments intended as physiological performance tests of whole animals.