WAHLEN: INCIDENTAL DOLPHIN MORTALITY 



second and third quarters during 1977 and 1978. 

 Additionally, pairwise i-tests were made to isolate 

 annual and quarter differences detected by the 

 above tests. 



Tests for Year Differences 



Sample statistics of mean kill by year and vessel 

 capacity (Test 1) revealed an unbalanced design and 

 suggested that cell variances were related to cell 

 means (Table 4). For each of several transformations 

 of total kill, neither the interaction between vessel 

 capacity and year nor the difference between vessel 

 capacities was significant, but the difference among 

 years was significant. 



To determine where the yearly differences oc- 

 curred, the data were pooled over vessel capacity 

 so that t-tests for differences between each pair of 

 yearly means could be made. The resulting one-way 

 classification by year was unbalanced and charac- 

 terized by significantly different cell variances (P 

 < 0.001), and suggested that the means from 1973 



Table 4.— Mean of total number of dolphins killed (k), standard 

 deviation (s), and number of dolphin sets (d) for sampled trips, by 

 year and vessel capacity (tons) for all sets made in the North In- 

 side area during quarter two. Significance probabilities (P) obtained 

 for 2-way ANOVA's on transformed values of total kill: interaction 

 (P > 0.1 180), year (P < 0.001), and vessel capacity (P > 0.7851). 



Vessel 



Year 



Table 5.— Matrix of significance probabilities 

 associated with f-tests for differences between pairs of 

 annual means of total number of dolphins killed. Signifi- 

 cant values, required by the Bonferroni adjustment to 

 be <0.0033, are indicated by "*", and nearly signifi- 

 cant values are indicated by " + ". Data are from sam- 

 pled dolphin sets made during quarter two in the North 

 Inside area. 



Year 



Year 1974 1975 1976 



1977 



1978 



through 1976 were larger than the means from 1977 

 and 1978 (Table 4). 



Results from pairwise i-tests indicated that means 

 were not significantly different within each of the 

 two periods from 1973 through 1976 and from 1977 

 through 1978 (Table 5). However, each of the means 

 from 1973 through 1976 was significantly different 

 (or nearly so) from each of the means from 1977 and 

 1978 (Table 5). Based on these results, I divided the 

 data into two periods, 1973 through 1976 and 1977 

 through 1978, for further tests within each period. 



Tests for Differences Within 

 the 1973-76 Period 



The three-way ANOVA table by year, quarter, and 

 vessel capacity (Test 2) was unbalanced and sug- 

 gested that cell variances were unequal (Table 6). 

 Furthermore, no between-f actor interactions were 

 significant. The test for year differences was incon- 

 clusive; however, there is evidence for a year effect 

 within only one cell (1976, quarter 1, small vessels), 

 and there is no consistent yearly pattern of means 

 within rows of the table. Therefore, I concluded that 

 annual means during 1973-76 were not significant- 

 ly different and, hence, I pooled over years within 

 this period. I also pooled over vessel capacity since 

 it was not a significant effect during these years. 



Before 1976, sample data were unrepresentative 

 of quarter and area, since nearly all data were ob- 

 tained from trips made during the first half of the 

 year and, thus, within the North Inside area. Data 

 sparseness in the North Outside and South areas, 



Table 6.— Mean of total number of dolphins killed (k), standard 

 deviation (s), and number of dolphin sets (d) for sampled trips, by 

 year, quarter, and vessel capacity (tons) for all sets made in the 

 North Inside area during the first two quarters of 1973-76. Signifi- 

 cance probabilities (P) obtained for a 3-way ANOVA on total kill: 

 interactions (P> 0.1374), year (P = 0.0534), quarter (P = 0.0501), 

 and vessel capacity (P = 0.8219). 



Vessel 



capacity 



(tons) 



Year 



<1,000 



>1,000 



563 



