in either plankton collection (0.2%). Grubby ranked third among 

 impinged species and was collected predominantly from December through 

 May (Table 4) . Length frequency histograms from both trawl and 

 impingement programs were very similar, with major peaks occurring 

 between 60-100 mm (Fig. 16). 



30 



40 



50 



60 



70 



80 



90 



100 



I 10 



120 



130 



140 



150 



160 



170 



180 



190 



200 



X/AVJ 



^AAAAAAA.V>P?WWWa 



•ssssssss^^^^ 



S222S2^^^22 



^vwvvvwza 



XXXXXXXXl 



VAWi 



] 

 I 



200 400 600 



FREQUENCY 



^Mr,r,^v^Mr>^s^>r,MM^^^iCiCM)CiC/\ 



wxxxxx^xxxxxxxxx] 



^^^Vy^A<v.A,^^vvv^<'v'^a 



^^^^ffiX2XX2S2 



^VwVwVWN 



S22ZZ3 





800 



1 00 200 300 



FREQUENCY 



400 



Figure 16. Length frequency of M. aenaeus in trawl (left) and impingement 

 (right) collections from October 1976 through September 1983. 



Descriptive models of grubby had various combinations of harmonic 

 components (Table 10). In general, all models had 12-mo cycle terms; 

 those from NR and TT also had significant 6-yr cycle term. Other 

 statistically significant cycles were 6-, 4-, and 3-mo. Cooling-water 

 flow was a significant multiplicative variable in models describing 

 impingement counts. Entrainment and impingement models were the most 

 accurate with R^ greater than 0.89 (Table 10). Models of the trawl data 

 were less accurate with R^ values ranging from 0.29 for BR to 0.68 for 

 JC. Model errors were high for trawl and low for entrainment and 

 impingement models and few 1983 data points were beyond the 95% 

 confidence limits (Figs. 17). In addition, forecast errors of the trawl 

 models ranged from 38% at JC to 294% at NR, but most of the data points 

 were within the 95% confidence limits except for NR (Fig. 18). At this 



31 



