FRENCH and DUNN: LOSS FROM HIGH-SEAS GILLNETTING 



salmon entered the nets IV2 h prior to the 

 beginning of the second 3-h interval. If, on the 

 other hand, we postulate that the majority of the 

 salmon entered the nets either prior to or after 

 the assumed IV2 h, an increased or decreased 

 time interval is available for salmon to drop 

 from nets during the 3-h period. Our estimate 

 of dropouts now would be either lower or 

 higher than that of the original model. A test of 

 these hypotheses was inconclusive. Dropout 

 rates were estimated at 15 and 35% for the two 

 new assumptions of time of entry into the nets, 

 but in neither instance was the rate significant. 



When fishing the experimental dropout 

 sets, we frequently observed northern fur seals, 

 CaUo7'hiuus ur.'^i)ius, and Steller sea lions, 

 Eumetopias jtibatus, swimming along the nets 

 and taking fish. We frequently observed sea- 

 birds near the nets and we often took salmon 

 sharks. Lanuia ditropis, in the gill nets. 

 Removal of fish from the nets by these animals 

 was a factor we could not differentiate from 

 actual salmon dropouts from the nets. Both 

 factors, however, lead to loss of salmon from 

 gill nets. As a matter of fact, the gill nets served 

 to collect and hold salmon for predators and 

 carrion eaters. We do not know if sea lions and 

 fur seals could catch salmon as easily through 

 their own efforts compared to having gill net- 

 caught fish available. We concluded from these 

 experiments, therefore, that there was loss of 

 salmon from gill nets that was due to predation 

 (removal of fish by predators and carrion 

 eaters) and dropouts. 



In the experiments to determine dropouts in 

 1964 and 1965, our technique involved indirect 

 methods and assumptions of equal availability 

 of salmon to units of gill nets fished during 

 varying time periods and that all nets fished 

 with equal efficiency. In 1965, the experiments 

 were limited to estimates of dropouts over a 3-h 

 period. To overcome these limitations, we 

 attempted in subsequent experiments to observe 

 salmon in the nets and determine droi)outs 

 by direct observation. 



Direct Observations (1966-69) 



1966. — In the summer of 1966, the loss of 

 salmon from gill nets was estimated by direct 

 observation. Three vessels fished with gill nets 



in July, August, and in early September along 

 three sampling lines: long. 176°22'W, 167 °W, 

 and 158°W. In this technique, as outlined 

 under the section of "Experimental Methods," 

 we observed fish in the net from small boats, 

 marked their position in the net with colored 

 tags, and determined their presence or absence 

 at subsequent time periods and at haul. 



The experimental gill nets included 4 

 shackles of 133-mm, 3 of 114-mm, 4 of 83-mm, 

 and 3 of 64-mm nets when the first 14 shackles 

 of the string were observed. The net sequence 

 was 133, 83, 114, 64 mm. In some sets not 

 all nets were observed because of inclement 

 weather or other factors. In any event, however, 

 a range of mesh sizes was observed to include 

 observations on age .1 immature salmon and on 

 the larger .2 and older immatures. 



In 1966 we made 28 observations for drop- 

 outs (Table 4). Each observation entailed one 

 to three patrols of the net, some of which were 

 not completed because of rough seas. There 

 were 28 first patrols, 27 second patrols, and 

 16 third patrols. 



Combining data from all observations, the 

 overall dropout rate was substantial, 42.5% 

 (95% confidence interval ±5.6) for periods up 

 to 11 h (Table 5). After 1 h the rate was 3.7% 

 (±5.1), and 14.9% ( ±7.9) for periods up to 2^/2 

 h. The rates for varying time periods were 

 generally similar for the two combinations of 

 mesh sizes. An exception was the 8% (±7.7) 

 dropout rate of the large mesh sizes for periods 

 up to 21/2 h compared with 22.7% (±14.6) 

 for the smaller mesh sizes. Typically, the large 

 mesh nets take age .2 and older salmon with 

 immature sockeye and chum salmon dominat- 

 ing catches in the summer. The small mesh 

 sizes take mainly age .1 immature sockeye and 

 chum salmon in the summer. Thus, overall 

 dropout rates appeared to be about the same 

 for small fish as for large fish. 



The data suggested little difference in loss 

 rates among the three vessels. The total rate of 

 loss in the large mesh nets ranged from 42.3% 

 (MV Paragon) to 48.1% (MV St. Michael). In 

 the small mesh nets, overall loss rates were 

 39.6% for the Paragon and 45.7% for the St. 

 Michael (and was 22.2% for RV George B. 

 Kelez, but only 10 fish were marked). 



853 



