all catches at a particular station or stations x 15 min 

 total time for the above tows 



As previously mentioned, the CPUE used for the trawl evaluation was 



simply the unadjusted catch (in terms of tow time) from the standardized 



0.69-km tow. The two methods of adjusting the catch on a time basis add 



an unknown amount of variability to the data as well as being internally 



inconsistent within the Annual Reports. For a benthic fish such as the 



winter flounder it is most reasonable to assume that the catch would 



depend mostly on the bottom ground covered by a trawl tow. Since trawling 



a standardized distance between two fixed points allows the ground 



covered by the trawl to remain the same for all tows (assuming all other 



factors affecting efficiency remain the same which is not entirely 



accurate) , then the manipulation by time is neither necessary nor 



justified. Factors such as tide, wind, and amount of debris in the net 



account for most of the variability in tow time at a station. Thus the 



trawl data which can be used with most confidence in any analysis is 



that from October 1977 -through the present without correction for time. 



Finally, a second inconsistency was noted between the analyses used 



in the Winter Flounder section and the Fish Ecology section. The former 



had In (CPUE + 1) transformed data used with a parametric ANOVA to 



examine for differences among stations and time periods. The latter 



stated that In (CPUE + 1) values were known to be non-normally distributed 



and used a Kruskal-Wallis non-parametric ANOVA for similar examinations 



(NUSCo 1982). However, a previous evaluation (Millstone Ecological 



Advisory Committe Meeting Minutes, Nov. 29-30, 1979) noted that the 



winter flounder trawl data from 1973 through 1978 did approximate a 



normal distribution when In transformed and could be used with parametric 



84 





