several meshes) are thought to cause the 

 asymmetry. 



Other investigators besides Holt used normal 

 distribution for mesh selection curves. McCombie 

 and Fry (1960), working with Lake Huron white- 

 fish data, concluded that normal distribution best 

 describes the mesh selection curve. Using nylon 

 gill net catches of Tilapia from Lake Victoria, East 

 Africa, Garrod (1961) showed that normal dis- 

 tribution applies over most of the selection range of 

 fish lengths but not at the extremes of this range. 

 Garrod used the normal curve obtained from 

 linear regression after discarding tlio extremes 

 where the relation departs from linear. I also used 

 the procedure of discarding the extremes and then 

 applying the normal curve. 



SUMMARY 



1. A method for determining gill net selectivity 

 described by Holt (1957) was applied to experi- 

 mental gill net catches of pink, sockeye, and chum 

 salmon from the high seas of the North Pacific 

 Ocean and Bering Sea. This method develops 

 mesh selection curves for gill nets of different mesh 

 sizes from catch ratios at various fish lengths. 



2. A normal mesh selection curve, representing 

 relative catch efficiency of the mesh for different 

 length classes of fish, was constructed for each 

 mesh size, 2%-, 3%-, 4%- and 5M-inch, for each 

 species. Normal distribution can be used validly 

 when extreme sizes of fish caught by snagging and 

 tangling rather than gilling are omitted. 



3. A composite curve of relative catch effi- 

 ciencies for combined mesh sizes shows that the 

 four mesh sizes cover the range of salmon lengths. 

 AH, length classes were not caught with equal 

 efficiency. A lower catch efficiency at 44 cm. of 

 47 cm. for sockeye and chum salmon and 41 cm. for 

 pink salmon, resulted from the larger (1%-inch) 

 gap between the 3%- and 4}2-inch mesh sizes. The 

 gap between other adjacent mesh sizes was three- 

 quarter inch. 



4. The composite curve for each species was used 

 to adjust gill net catches for selectivity effect. 

 Adjustments were minor. 



ACKNOWLEDGM ENTS 



R. H. Lander helped with the mesh selection 

 theory, F. C. Cleaver made suggestions for writing 

 the report, and W. F. Royce reviewed the report. 



LITERATURE CITED 



Garrod, D. J. 



1961. The selection characteristics of nylon gill nets 

 for Tilapia esculenta Graham. Journal du Conseil, 

 vol. 26, No. 2, pp. 191-203. 



GULLAND, J. A., AND D. HARDING. 



1961. The selection of Clarias mossambicus (Peters) 

 by nylon gill nets. Journal du Conseil, vol. 26, 

 No. 2, pp. 215-222. 



Hodgson, William C. 



1933. Further experiments on the selective action of 

 commercial drift nets. Journal du Conseil, vol. 8. 

 No. 3, pp. 344-354. 



Holt, S. J. 



1957. A method of determining gear selectivity and 

 its application. ICNAF-ICES-FAO Joint Scien- 

 tific Meeting, Lisbon, Paper No. S15, 21 pp. 

 [Mimeographed.) 



Ishida, Teruo. 



1962. On the gill net mesh selectivity curve. Hok- 

 kaido Regional Fisheries Research Laboratory, 

 Bulletin No. 25, pp. 20-25. [In Japanese with 

 English summary.] 



Lander, Robert H. 



1963. Girth-length relationships in sockeye and chum 

 salmon. Transactions of the American Fisheries 

 Society, vol. 92, No. 3, pp. 305-307. 



McCombie, A. M., and F. E. J. Fry. 



1960. Selectivity of gill nets for lake whitefish, 

 Coregonus clupeaformis. Transactions of the Ameri- 

 can Fisheries Society, vol. 89, No. 2, pp. 176-184. 

 Olsen, Steinar. 



1959. Mesh selection in herring gill nets. Journal of 

 the Fisheries Research Board of Canada, vol. 16, 

 No. 3, pp. 339-349. 

 Peterson, A. E. 



1954. The selective action of gill nets on Fraser River 

 sockeye salmon. International Pacific Salmon 

 Fisheries Commission, Bulletin No. 5, 101 pp. 

 Snedecor, George W. 



1956. Statistical methods applied to experiments in 

 agriculture and biology. The Iowa State Uni- 

 versity Press, Ames, Iowa, 534 pp. 



390 



1 .S. I ISIl \\I> WILDLIFE SERVICE 



