28 Alvin L. Braswell and Ray E. Ashton, Jr. 



and Neuse rivers may be acting to isolate the N. lewisi in the Trent 

 River subdrainage from the main population in the Neuse River basin. 



Regrettably, water fluctuations made it impossible to record stream 

 widths, depths, and flow rates under stable conditions for most collect- 

 ing sites. A stream index based on such measurements would likely be 

 most valuable if the measurements were taken under dry-weather, 

 summer conditions when oxygen levels are more critical for N. lewisi. 

 Under prevailing winter and spring conditions, stream width was the 

 most reliable of the three measurements and was used to formulate the 

 stream categories in Table 1. Stream size appears to be a limiting factor 

 for N. lewisi occurrence only in the case of small streams. Cursory 

 observations on Piedmont streams indicate that N. lewisi does not occur 

 in streams where water flow ceases under normal summer dry-weather 

 conditions. Based on our data, N. lewisi seems to survive better in 

 streams wider than 15 m, deeper than 100 cm, and with a main channel 

 flow rate greater than 10 cm/ sec. This assessment agrees with Hecht's 

 (1958) general theory of the streams N. lewisi occupies. 



Although capture success indicates that N. lewisi prefers clay or 

 hard soil substrate, a variety of conditions appear suitable (Table 3). 

 Siltation levels may be a more critical factor, since heavy siltation could 

 reduce the availability of food and cover, and hinder reproduction by 

 smothering nests and eggs. Generally, siltation seemed light to moderate 

 in most N. lewisi sites. 



Capture success rates do not support the contention of Fedak 

 (1971) and Martof et al. (1980) that N. lewisi spends the winter in leaf 

 beds. That hypothesis probably arose because N. lewisi can be captured 

 regularly by seine and dipnet only when they are in such habitat (see 

 Ashton, this publication). However, too much speculation about habitat 

 preferences based on trapping success is unwarranted, since the distance 

 N. lewisi will travel to a baited trap is unknown. 



A few places where N. lewisi was expected but not collected during 

 our study include: the Neuse River from east of Raleigh to the vicinity 

 of Smithfield; Swift Creek in Johnston County; and the Tar River from 

 Rocky Mount to about 20 km downstream. To what extent municipal 

 and industrial effluents may be acting to cause such apparent gaps in 

 the distribution is unknown. 



The Little River from eastern Wake County to the vicinity of 

 Goldsboro, and the Trent River in Jones County, appear to have the 

 best N. lewisi populations in the Neuse River drainage. Virtually the 

 entire Tar River from Granville County to central Pitt County, with the 

 exception of Rocky Mount Reservoir and the 20 km section below 

 Rocky Mount, appears to have a healthy population. Other Tar River 

 drainage streams with apparent good populations include Fishing and 



