SMITH: SIZE CHANGES OF THREE DOLPHIN POPULATIONS 



1980 to determine the effect of sea state and sun 

 position on the visibility of dolphin schools di- 

 rectly on the trackline. Data from this experi- 

 ment, which have not yet been completely ana- 

 lyzed, will be of use in the design of future 

 surveys and in the evaluation of earlier sur- 

 veys. 



INCIDENTAL KILL ESTIMATES 



Incidental kill (K, ) of dolphins in year t is esti- 

 mated by multiplying the mean kill of dolphins 

 per set in year t ( KPS,) by the total number of net 

 sets involving dolphins made by the tuna fleet in 

 year t (NSETS,), as 



K, = KPS, NSETS,. 



(6) 



These estimates are obtained for each year with 

 the data stratified by vessel fish-carrying capac- 

 ity, amount of tuna caught in the net set, and 

 geographic location of the set, following the gen- 

 eral approach described by Lo et al. (1982). 



Kill rate information is available from a lim- 

 ited set of tuna fishing trips in the 1960's and 

 from a more extensive set in the 1970's collected 

 by scientific observers placed aboard a large pro- 

 portion of the U.S. fishing vessels. To illustrate 

 the data, some mean kill rates, stratified by 

 amount of tuna caught, are shown in Table 2. 

 Higher kill rates are apparent in successful (>% 

 ton tuna caught) than in unsuccessful (<% ton 

 tuna caught) sets, as are marked declines in kill 

 rates over time. Numbers of dolphin sets and 

 fishing trips on which observations of numbers 

 of dolphins killed were made are shown in Table 

 3. 



Observations of the numbers of dolphins killed 

 in the 1960's were made by both the crew and the 

 scientists. Although few observations were made, 

 there is no consistent difference between kill 



Table 2.— Observed mean kill of dolphins per net 

 set (KPS) by U.S. tuna purse seiners 1964-78, for 

 successful and unsuccessful net sets, with sample 

 sizes (iV), from NMFS records. 



rates reported by both types of observers (59 and 

 52, respectively); this suggests the presence of a 

 noncrew-member observer had no significant 

 effect on the kill rate of dolphins in the 1960's. 



All data on kill rates of dolphins for the period 

 1971-78 were collected by noncrew-member sci- 

 entists, precluding a direct comparison of kill 

 rates between fishing trips with (observed) and 

 without (unobserved) scientific observers for this 

 period. Groom, 5 however, reported dolphin kill 

 rates on a fishing trip in 1979 with no scientific 

 observer on board; his kill rates were about 4 

 times higher than the average rate in 1979 for 

 scientist-observed trips and were approximately 

 20 times higher than on previous and succeeding 

 observed fishing trips by the same vessel and 

 captain. This difference in mean kill rates was 

 due to the significantly lower proportion of sets 

 with few dolphins killed on Groom's trip than on 

 the scientist-observed trips. For instance, the 

 proportion of sets with zero dolphins killed was 

 0.23 on Croom's trip with 0.76 for observed trips. 



Although limited information is available, it 

 appears that kill rates on some unobserved ves- 

 sels were higher in the late 1970's, and that this 

 could result in the observed kill rates being lower 

 than the actual rates. If there has been an "ob- 

 server effect," it most likely occurred in the late 

 1970's, because regulations were adopted in the 

 United States in 1976 requiring the use of cer- 

 tain dolphin-release procedures, and because sci- 

 entific observers were then used to collect regu- 

 lation compliance information. If kill estimates 

 for the last few years were revised with this in 

 mind, it would only slightly affect the calcula- 

 tions presented here, since the large number of 

 animals killed through 1975 tends to dominate in 

 Equation (2). However, such revisions to the kill 

 estimates could markedly change our perception 

 of the current rate of change of these populations. 



In addition to the known direct kill of dolphins 

 in the fishery, research has been conducted to 

 estimate both the number of dolphins injured 

 and released alive from the purse seines, and the 

 possible number of dolphins which, while not ex- 

 hibiting injuries, die or suffer reduced viability 

 from stress of capture and handling in the purse 

 seining operation. Observations of the number of 

 injured dolphins have been made aboard tuna 

 vessels since 1975; estimates of the number in- 



5 Croom, M. M. 1980. The tuna-porpoise problem: Man- 

 agement aspects of a fishery. M.S. Internship Rep., Marine 

 Resour. Manage. Program, Oreg. State Univ., Corvallis, 41 p. 



