366 ZEEMAN AND GRUNEWALD 



Differential injury and recovery of specific size classes of 

 entrained phytoplankton could be important to community struc- 

 ture in the receiving waters. The small nannoplankton are thought to 

 be more important to primary production than the larger net 

 plankton (Pomeroy, 1974). Rodhe, Vollenweider, and Nauwerck 

 (1958) found that nannoplankton were responsible for more than 

 90% of production in spring and early summer and that net plankton 

 became more important in summer and fall. Similar results were 

 presented by Goldman and Wetzel (1963). 



As far as we know, the application of multivariate analysis of 

 variance (MANOVA) to a data set of this type is novel, but 

 multivariate analyses in general have been successfully applied to 

 algal communities (Allen and Skagen, 1973). 



Since there were multiple measures for each sample (i.e., 

 size-fractionated productivities), the chances for correlation among 

 the variables was high (Van de Geer, 1971). Thus a separate analysis 

 for each variable would not constitute independent tests. The 

 MANOVA technique then provides an important analytical tool, and 

 separate analyses should be avoided unless the MANOVA null 

 hypothesis has been rejected (Cooley and Lohnes, 1971). The 

 MANOVA procedure assumes a multivariate normal distribution for 

 variables with equality of dispersions. 



MATERIALS AND METHODS 



Site Description 



The Kewaunee Nuclear Power Plant is located on the shore of 

 Lake Michigan, ~43.5 km east— southeast of Green Bay, Wise. The 

 facility has once-through cooling and can produce a net output of 

 540 MW(e). Ambient temperatures ranged from 0.5 to 18.3°C and 

 AT's, from to 14.5°C. The maximum discharge temperature 

 recorded during this study was 28.8°C. 



Sampling 



Sampling was initiated on Aug. 26, 1975, and concluded on 

 July 23, 1976. Samples were collected from a precondenser station 

 in the forebay (FRB) and a postcondenser station in the discharge 

 flume (KD) on eighteen dates. Two other stations were sampled on 

 six dates from August to November, one at the intake structure 

 (KWI) and one at a reference station (REF) about 4 km north of the 

 plant. Figure 1 shows the location of the plant and the sampling 

 sites. 



