348 



Fishery Bulletin 93(2). 1995 



The stocks of offshore spotted dolphins are not dis- 

 tinct in external morphology and therefore could not 

 be noted as proportions in the observer data. There- 

 fore, I estimated kill for each stock by using these 

 stratified numbers of sets to estimate total dolphin 

 kill in each stock area, then prorated the total kill 

 by the observed proportion of offshore spotted dol- 

 phins in the kill. The same method as in Smith ( 1983) 

 was used for prorating species proportions and will 

 be explained in more detail below. 



Because eastern spinner and whitebelly spinner 

 dolphins are morphologically distinct, estimates of 

 kill for those two stocks were prorated from the esti- 

 mated total dolphin kill (summed across the two geo- 

 graphic areas) by observed proportions of the two 

 stocks in the kill. Observed proportions of dolphins 

 killed from species other than spotted and spinner 

 were reported as "other dolphins." Therefore, esti- 

 mates were made for this pooled category in the same 

 way as for the stocks of spinner dolphins. Dolphins 

 in this category were primarily common dolphins, 

 Delphinus delphis (Smith, 1983). 



A second modification was possible because IATTC 

 provided the number of sets additionally stratified 

 by tuna catch and vessel size (Table 1 ). Estimates of 

 the number of sets in each stratum were made by 



multiplying Punsley's (1983) estimates of total dol- 

 phin sets by the proportions of known sets (from the 

 IATTC logbook database) that were in each stock 

 area, set type, and vessel size category. 6 This is likely 

 a substantial improvement in accuracy over the pro- 

 ration method based on less data used by Lo and 

 Smith (1986). 



A third modification to the Lo and Smith (1986) 

 method was a change in the way that the variance 

 was calculated. They derived analytical equations for 

 estimating the variance of total dolphin mortality in 

 each year. More important for management is the 

 variance of the estimate of dolphins killed in each 

 stock in each year, which cannot be obtained from 

 the equations in Lo and Smith (1986). Observations 

 of stock and species proportions of the kill were not 

 systematically collected until the observer program 

 started in 1971 (Edwards, 1989). Therefore, Smith 

 (1983) used the observed proportions from 1971 to 

 1972 for the entire 1959-72 time period, with the 

 exception that all spinner dolphins killed before 1969 

 were assumed to be from the eastern stock and that 

 no whitebelly spinner dolphins were killed until 1969. 

 This was assumed because no dolphin sets occurred 

 in the whitebelly stock area before 1968, and only a 

 very small number occurred in 1968 (Punsley, 1983). 



