94 



been generally supposed that on an area where the physi- 

 cal conditions, depth of water, and nature of bottom, are 

 similar, the distribution of fish is pretty uniform. The 

 distribution of fishes will depend primarily on the distri- 

 bution of their food, and how variable this may be Avill 

 be seen by taking any area of cockle-bearing sands, ten 

 miles by ten miles, and observing how irregularly cockle 

 beds may be distributed over this area, though the 

 physical conditions are closely similar in every part. 



It will appear, then, that to obtain results which will 

 give more than merely general conclusions, such as we 

 have given with regard to the Blackpool and Mersey 

 grounds, must require very frequent trawling observations 

 and the employment of averages based on rather large 

 series of figures. By considering averages only and cal- 

 culating these from many data, some of the possible errors 

 which we have indicated above as inseparable from 

 isolated observations, might be eliminated. 



Such considerations as we have dealt with above appear, 

 then, to indicate that if a series of trawling observations 

 is to be undertaken with the view of determining the 

 changes in time in the fish population of a fishing ground, 

 these things must be borne in mind : — 



(1) The nets employed must be chosen and used so that 

 their fishing power Avill be equal — unless one net is 

 employed throughout the series of observations. 



(2) Due regard must be paid to the differences which 

 the varying states of the tide must make in the catches, 

 or the obsei^ations must be made under the same 

 conditions of tide. 



(3) Finally, the results should be based on the averages 

 of several observations taken at neighbouring places so 

 that the differences due to irregular distribution in space 

 may be as far as possible eliminated. 



