GILL NET MESH SELECTION CURVES FOR PACIFIC SALMON ON THE 



HIGH SEAS 



By Alvin E. Peterson, Fishery Biologist (Research) 

 Bureau of Commercial Fisheries, Seattle, Wash. 



ABSTRACT 



Gill net mesh selection curves of normal distribution 

 were developed and applied to Pacific salmon caught by 

 research vessels on the high seas of the North Pacific 

 Ocean and the Bering Sea. Mesh selection curves were 

 constructed for pink, sockeye, and chum salmon for 

 each of four mesh sizes, 2y 2 -, 3Vi-> 41 /2-> an< l 5V4-inch. 



Catch efficiency curves for combined mesh sizes show 

 that the range for salmon lengths was covered, although 

 the coverage was not equal for all lengths. The length- 

 frequency distribution of each species was adjusted for 

 effect of gill net selectivity. Adjustments were minor. 



The Bureau of Commercial Fisheries fishes 

 experimentally for salmon with surface gill nets on 

 the high seas of the North Pacific Ocean and the 

 Bering Sea. Salmon samples taken by gill nets 

 are used to estimate abundance, distribution, racial 

 identity, and growth of salmon populations in the 

 ocean. Accurate estimates of abundance, distri- 

 bution, racial identity, and growth require un- 

 biased samples from salmon populations in the 

 ocean. Gill nets are selective; a particular mesh 

 size of gill net selects a particular size range of fish. 



To cover the range of fish sizes, four gill net 

 mesh sizes, 2K-, 3%-, 4%-, and 5%-inch, stretched 

 measure, of multifilament nylon twine are used 

 in the fishing. Selectivity studies are necessary 

 to assess the adequacy of this coverage and to 

 adjust the salmon size frequencies for any bias 

 caused by selectivity. Determining the shape 

 and extent of the mesh selection curve for each 

 mesh size and for combined mesh sizes is necessary 

 before size frequencies can be adjusted for possible 

 bias. 



Note.— Approved for publication June 25, 1964. 

 FISHERY BULLETIN: VOLUME 65, NO. 2 



A METHOD FOR DETERMINING GILL NET 

 SELECTIVITY 



Holt (1957) described a method for determining 

 gill net mesh selection curves with normal fre- 

 quency distributions. He developed normal mesh 

 selection curves for Fraser River sockeye salmon 

 (from Peterson, 1954) and for North Sea herring 

 (from Hodgson, 1933). He used the ratio of 

 catches from adjacent pairs of mesh sizes at 

 different length classes to develop parameters for 

 the normal curve. The following formulations 

 were abstracted from Holt (1957) and McCombie 

 and Fry (1960): 



C L =nP l p m -e-<*-W 8 ' 



(1) 



where C L is the number of fish of length L caught, 

 n the number of operations or the fishing dura- 

 tion, P L the number of fish of length L liable to 

 capture, p m the fishing power of the mesh at the 

 mean selection length, e the base of natural 

 logarithms, L m the mean selection length caught, and 

 S the standard deviation of distribution. The ratio 

 of catches for two meshes {A and B), differing 

 slightly in size and fishing together, can be de- 



381 



