N. lewisi Habitats and Behavior 85 



No. 1). Two-hundred seventy-one positive sightings were made during 

 this period. Nine adult N. lewisi were captured and released within the 

 same 25 m 2 area in Site No. 1 from 15 January 1980 to 28 August 1980. 

 We consequently were unable to identify individuals, but could monitor 

 microhabitat use and movement within the same area under similar 

 stream conditions. Two-hundred seventy-four sightings of tagged ani- 

 mals were made during this period. Similarly, on 21 February 1979 five 

 post-hatchling larvae, 41 to 46 mm total length (TL), were captured by 

 dip-netting, then were tagged and released in the Little River in Johns- 

 ton County (Site No. 2). One-hundred thirty-eight sightings were made 

 of these animals. 



All tagged animals were monitored bi-weekly during all three study 

 periods. When a tagged animal was located, data were taken on ambient 

 temperature, flow rate, turbidity, dissolved oxygen, carbon dioxide, and 

 pH, both at the animal's current location and at the site where the 

 tagged animal had previously been found. Standard data were collected 

 at a single control station at each study site throughout the study. This 

 allowed us to evaluate changes in overall stream conditions and com- 

 pare them with varying conditions in the microhabitat. Weather data, 

 including maximum and minimum air and water temperatures and pre- 

 cipitation, were monitored at or near the control site. Other data were 

 obtained from the U.S. Weather Bureau, Raleigh-Durham Airport, 

 approximately 17 km from Site No. 1 and 27 km from Site No. 2. 



The study areas were mapped for depth, bottom types, amount and 

 types of rock cover, and leaf bed development. Very little change in 

 these features took place in both areas throughout the study. General 

 water quality at both study sites was analyzed monthly from April to 

 December 1979 (Table 1). Water analysis was done using a Mach DR- 

 EL/2 with spectrophotometer. 



Mercury reached peaks of 4.18 mg/ 1 (x=1.07) at Site No. 1 and 

 5.09 mg/ 1 (x=2.17) at Site No. 2. These levels indicated that mercury 

 was a potential pollutant. High levels of nitrates and sulfates were also 

 observed, notably in the May and June samplings, which followed peri- 

 ods of rain and subsequent runoff from surrounding farmlands. Con- 

 centrations of all other chemicals tested for were within normal limits. 



Two surveys of aquatic invertebrates, one along a transect (Table 2) 

 and the other in a 10 m 2 quadrat (Table 3), were made at Site No. 1 to 

 quantify potential food items in microhabitats used by N. lewisi. Aqua- 

 tic vertebrates (Table 4) were collected by seine, trap, and electroshock. 



Standard SAS programs, including PROC-F reg, categorical data 

 and PROCORR using Perison, Sperman, and Kendall were used to 

 determine significance of specific correlations between behavior and 

 physical factors within the microhabitat. Although tests showed many 

 significant correlations between environmental factors, movement rate, 



