254 BRITISH FISHERIES 



ground, the changes taking place in the abund- 

 ance of fishes on the latter could be estimated ; 

 that is, that the fishes caught in a haul taken 

 on an area of, say, loo square miles in extent 

 were a sample of all those actually present. Now, 

 such a trawl-net as has been commonly used 

 in these investigations sweeps a portion of the 

 sea-bottom of two or three miles long at the 

 most, and about twenty-five feet wide. This is 

 only a very srnall fraction of the whole area, 

 and if we are to regard the catch made as a 

 fair sample of the whole, we must assume that the 

 fishes present are uniformly scattered over the 

 whole area. But experiment shows that this is 

 far from being the case.^ Again, a trawl-net as an 

 instrument of research is far from being perfect. 

 What it does catch does not depend alone on the 

 quantity of the fish actually present on the fishing 

 ground on which it is used, but also to a great 

 extent on the exact construction and " trim " of 

 the net, on the state of wind and tide, on the skill 

 of the fisherman using the net, and on a variety 

 of circumstances well known to fishermen, but 

 generally neglected by those who have adopted the 

 method in the past. Subsequent manipulation of 

 the results obtained by statistical methods will 

 obviously not render them of any more real value, 

 if the data themselves are of such questionable 



1 Johnstone, Report Lancashire Fisheries Laboratory for 1902; 

 p. 83. 



