WAHLEN: INCIDENTAL DOLPHIN MORTALITY 



in stratification decisions than for the 1973-76 data. 

 The four- way test on 1977-78 data (Test 3) was not 

 helpful in resolving the question of quarter differ- 

 ences because of the interactions between quarter 

 and each of the other three factors. However, pair- 

 wise t-tests for differences between quarterly means 

 in the North Inside area during 1977-78, pooled over 



Table 9.— Matrix of significance pro- 

 babilities associated with f-tests for dif- 

 ferences between pairs of quarterly 

 means of total number of dolphins 

 killed. No significant values, required by 

 the Bonferroni adjustment to be 

 <0.0083, were attained. Data are from 

 sampled dolphin sets made in the North 

 Inside area from 1977 through 1978. 



year and vessel capacity, detected no significant 

 quarter differences (Table 9). Based on that result, 

 I pooled over all quarters in the North Inside and 

 North Outside areas. Finally, I pooled over all 

 quarters in the South area because of the small sam- 

 ple sizes (Table 3). 



Thus, after pooling over year, vessel capacity, and 

 quarter as indicated above, only three strata re- 

 mained for the 1977-78 data: 1) North Inside, 2) 

 North Outside, and 3) South. 



Estimates 



I obtained annual estimates of the total number 

 of dolphins killed by summing estimates for each of 

 three or five strata, depending on the year. Esti- 

 mates for a stratum were computed as the product 

 of (a) total number of dolphin sets (Table 10) and 

 (b) the corresponding total kill-per-set ratio (Table 

 11), increased by (c) the observed total number of 

 dolphins killed during experimental-gear trips (Table 



Table 10.- 



-Estimated number of dolphin sets (D) and number of trips (N) for the 

 population of trips, by year within strata. 



North Inside 



Table 11.— By-trip means of total number of dolphins killed (k) and of number of 

 dolphin sets (d), total kill-per-set ratio (R), estimated standard error of the total kill- 

 per-set ratio (s), and number of sampled trips (n), by strata. 



565 



