296 
BULLETIN OF THE BUREAU OF FISHERIES 
irregularities in growth curves resulting from the selective action of nets are on the 
whole small. 
Creaser (1926) in connection with his study of the growth of the sunfish ( Eupo - 
motis gibbosus) observed that gill nets are “particularly selective in their collecting” 
and advocated the use of a wide range of mesh size in gathering material for biological 
investigation, but included no data on the selective action of gill nets. 
Koelz (1926) pointed out that in the Lake Ontario chub {Leucichthys spp.) fishery 
an increase of a quarter inch in the stretched measure of the mesh (from 2% to 2% 
inches) reduced the number of fish taken by more than half. 
Pritchard (1928) presented data on the selective action of six different sizes 
of mesh as that action affected the individual weight of Lake Ontario chubs {Leuci- 
chthys spp.). (The mesh size of his gill nets ranged from 1 % to 3 inches, stretched 
measure.) On the basis of his observations he stated, “The difference of one-quarter 
of an inch in the size of mesh may mean to the fishermen either a profitable or a 
‘starvation’ industry.” In a later paper (Pritchard, 1931) he presented the numbers 
of chubs of all species taken in nets of 12 different sizes of mesh (1% to 5 inches). 
Some data were included on the average lengths of the fish taken in different mesh 
nets. The data of Pritchard’s 1931 paper agree with his earlier findings in indicating 
that a very small increase in the size of mesh may produce a great decrease in the 
number of fish taken. 
Hart (1928) published detailed data on the length distribution of pike-perch 
and saugers of Lake Nipigon and Lake Abitibi taken in 6 different sizes of mesh 
ranging from 1 % to 4% inches, stretched measure. Although Hart confined his 
discussion to the question of the proper legal mesh size in the commercial fishery, 
his data on the length distribution of the catches in different mesh size show clearly 
that a single size of mesh can take these spiny-rayed fish over a considerable length 
range. In a later paper (Hart, 1931) the same author pointed out that observed 
differences in the growth rate of the whitefish {Coregonus clupeaformis ) in different 
parts of Lake Ontario could be explained in part by the tendency for the gill nets 
used in collecting samples from some localities to take only the larger fish of the 
younger age groups while pound-net samples were largely free from such selection. 
A more detailed consideration of the selective action of gill nets on whitefish appeared 
in Hart’s (1932) study of the population statistics of Shakespeare Island Lake. 
(His experimental gear included 11 sizes of mesh, 1% to 5 inches.) Hart found a 
correlation of 0.84 ±0.01 between size of mesh and size of fish in Shakespeare Island 
Lake, while in Nipigon the correlation between the same variables was 0.51 ±0.02. 
He pointed out further that “large fish may be taken in small meshed nets although 
small fish are practically never taken in nets of coarse mesh.” 
Lechler (1929), referring to the selectivity of gill nets used in the fishery for the 
Reinanke {Coregonus jera) held that the selective action of the net is quite sharp. 
He stated: “Die Zusammensetzung der Fange nach den Jahresklassen und damit 
die Grosse der Fische ist eine Funktion des Umfangs und direkt von der Netzmas- 
chenweite abhangig. ” Further experiences caused him to modify this view slightly, 
for in a later publication (Lechler, 1930) he observed, “Die Fangigkeit der Netze 
andert sich je nach der Beschaffenheit eines Bestandes.” 
Wright (1929) eliminated the younger age groups from his growth data on the 
rock bass {Ambloplites rupestris ) as the selective action of the nets caused these 
early groups to have too high average values for length. 
