FRENCH and DUNN: LOSS FROM HIGH-SEAS GILLNETTING 



was made. Additionally, because the study 

 required the use of small boats from the 

 research vessel at night, some patrols were not 

 made in rough seas. 



These studies, with minor modifications, 

 were conducted during high-seas gillnetting in 

 1966-69 and in Puget Sound in 1967. Details 

 of these modifications are explained in the 

 analysis section. 



The direct method of observation also con- 

 tained potential sources of error. For example, 

 salmon marked on one patrol may drop out but 

 be replaced by another fish, thus causing the 

 observer to miss an actual dropout. Moreover, 

 some fish may be inadvertently missed on sub- 

 sequent patrols and erroneously recorded; 

 upon hauling the net, the observer may then 

 attribute the presence of a fish to replacement. 

 Furthermore, as was the case with the indirect 

 methods, it was not possible to differentiate 

 losses due to dropout and losses due to 

 predation. 



Fallout Studies 



Observations of the number of fish falling 

 out of gill nets during net hauling operations 

 (fallouts) were made from 1965 to 1970. Fish 

 falling from nets of known mesh sizes were 

 tallied and expressed as a percentage of the 

 total catch. The number of fallouts are believed 

 to be reasonably accurate, although some may 

 have been missed during inclement weather or 

 during very large catches. 



SALMON DROPOUT RATES 

 DURING GILLNETTING 



Offshore Fishing 



Indirect Observations (1964 and 1965) 



1964. — Studies of the loss of salmon from 

 gill nets by indirect methods were conducted in 

 September and October. Four sets for these 

 studies were made south of the western Aleutian 

 Islands (one set along long. 178 °W and three 

 sets along long. 173°E); catches were princi- 

 pally of immature salmon. 



The fishing procedure provided catch data 



for approximately four 3-h periods throughout 

 the night and for approximately one 6-, 9-, and 

 12-h period. Precise gear hauling and setting 

 schedules could not be maintained, and usually 

 units of gear scheduled for a 3-h fishing period 

 were fished for shorter time periods. Results of 

 two experiments were unusable because of a 

 combination of weather and tidal currents 

 affecting the net string and because of the 

 difficulty encountered in hauling the nets. The 

 two successful experiments, however, provided 

 data on catches in 6, 9, and 12 h compared to 

 the cumulative catches in nets fished for 3-h 

 periods. 



The two sets caught 257 and 239 salmon, 

 respectively. Catches were predominately age 

 .1 immature sockeye salmon.'^ Catches by age 

 and species of fish and by mesh size in each 

 unit of gear are listed in Table 1. 



The catches in gill net units fished continu- 

 ously for periods of 6, 9, and 12 h, and the 

 cumulative catch in individual units fished 

 for short intervals (2.0-3.5 h) are compared in 

 Figure 4. Less fish were landed each day in nets 

 fished continuously for 9 and 12 h than for the 

 9- and 12-h cumulative catches of individual 

 units fished for three and four short intervals. 

 The catches each day were similar between 

 fishing methods for approximately the first 6 h. 

 Thereafter, in both experiments, the gear that 

 fished continuously caught comparatively fewer 

 fish. 



The loss from the continuously fishing units 

 (diffei-ence in total catch of the two fishing 

 methods divided by the total catch during short 

 intervals of fishing) for the two sets was 16 and 

 24% in 10 h and 43 and 20% in 12.5 h. No loss 

 was shown after 6 h. The average for the two 

 sets was 20% for 10 h and 32% for 12.5 h (Table 

 2). The loss was attributed to dropouts and 

 predation. 



1965. — In the spring of 1965. a concen- 

 trated effort was directed toward indirect 

 methods of estimating the rate of dropout of 

 salmon from gill nets. Two research vessels 

 fished on maturing sockeye salmon (primarily 



•''Age designation follows that of Koo (1962); numerals 

 indicate numbers of ocean annuli. 



849 



