582 LENAT 



The substrate at each benthic station is heterogeneous. The most 

 common type of substrate is a firm clay— sand, but patches of rubble 

 or detritus occur frequently. Macrophytes invaded Belews Lake in 

 year 5 and occurred at all stations (except one) in year 6. 

 Potamageton diversifolias and Najas sp. were the most common 

 species. The density of macrophytes in the discharge area was much 

 less than that observed in the main lake area. At discharge stations 

 macrophytes usually occurred only as scattered patches in depths of 

 to 0.5 m. 



The most complete temperature data are for years 5 and 6 and 

 are presented in Table 1. Temperatures as high as 31° C occurred 

 during the periods sampled. Dissolved oxygen (not shown) was 

 usually near saturation and was not considered to be limiting. 



METHODS 



The 14 littoral benthic sampling stations at Belews Lake are 

 shown in Fig. 1. Stations are located to form two longitudinal series, 

 one in the main lake basin (65M, 8, 4, 1907, 2, 1906, and 1116) and 

 one in the discharge basin (7, 5, 1904B, 1904 A, and 1904). Two 

 additional stations were located in the main lake basin, near the 

 cooling water intake (3) and the ash basin outfall (1). Samples were 

 taken quarterly in August, October, January, and April. 



A box-type sampler was used which sampled approximately 0.2 

 m^ at depths of 0.3 to 0.7 m. In all sample processing, screens with a 

 mesh aperture of 300 n were used. Samples were preserved in 95% 

 ethyl alcohol and stored in quart mason jars. 



Samples were stained with rose bengal and sorted under a 7x 

 dissecting microscope. For the Oligochaeta and Chironomidae, 

 mounting with CMC-10 or polyvinyl lactophenol and identification 

 under a compound microscope were required. 



Analysis of benthic collections includes computation of 

 Shannon— Wiener diversity index (Wilhm, 1967) and percent similar- 

 ity coefficients (Johnson and Brinkhurst, 1971). 



Four samples were taken at each station, and an effort was made 

 to sample all substrate types. Since this procedure generates very 

 high variability among replicates, statistical analysis treats the four 

 replicates as a composite sample. Comparisons are made, not among 

 stations, but between two station groups. 



The basic statistical design was a complete randomized-block 

 procedure. All analyses were run on log-transformed data. Interac- 

 tion terms were used as a divisor in computing F values. This 

 procedure was facilitated by use of the program FACTOR of the 



