386 
BULLETIN OF THE BUREAU OF FISHERIES. 
It would follow from these considerations that the net coefficient can not be 
accurately determined. It varies even during the same haul and becomes greater 
toward the end of the haul than at its beginning, because the net becomes progres- 
sively clogged. 
The coefficient of the net used in our investigation was determined by comparing 
the amount of plankton taken by the net with that taken by the pump. The 
experiment was made on a small pond of stagnant water 2 m. deep and with a 
plankton content of about 15 cm. 3 per cubic meter. The coefficient computed for 
a new net was 1.40. The volume of plankton per cubic meter of water has been 
calculated from the actual data, using the formula M=qCtv, where for all observa- 
tions q was supposed to be equal to the reciprocal of 1.40, 0.714, and the velocity 
v was one-half meter per second. The data obtained simultaneously with net and 
pumping methods are given in Table 18. The data of the pumping method rep- 
resent the average volume of plankton at each station calculated from the deter- 
minations of the volume of plankton at various depths from surface to the bottom. 
The comparison of the average volume of plankton thus calculated with that 
obtained by continuously lifting the hose of the pump from the bottom to the 
surface and pumping 50 liters showed an insignificant difference between the 
results of the two methods, not to exceed +3 per cent. The table shows that in 
most cases the amount of plankton obtained by the pumping method is greatly 
in excess of that taken by the net. 
The differences between the data vary from + 1.25 per cent to — 70.5 per cent 
of the pump data. All cases where the amount of plankton taken with the net is 
greater than that obtained with the pump occur in shallow water (Table 18, sta- 
tions 35 and 114). This wide range of fluctuation indicates that in most cases the 
net coefficient ought to be greater than 1.40, but that in the shallow water (stations 
35 and 114) it must be less. Since the filtering capacity of the net depends not only 
on the depth to which the net has been lowered and the amount of plankton, but 
also on the kind of plankton, it seems very difficult to find a definite relation between 
the net coefficient and the condition at which the haul is made. That the coeffi- 
cient varies with different depths is apparent in Table 18. 
The most instructive case may be found at stations 112 to 114, where the differ- 
ence between the net and pump data in the cases of deep hauls varies from —61.4 
to —70.5 per cent, whereas in the case of shallow water the difference is only +1.2 
per cent. In order to avoid the error resulting from the adoption of one coefficient 
irrespective of the age of the net and of seasonal, local, quantitative, and qualita- 
tive differences in the catch, Kofoid (1903) decided to assign an empirical coeffi- 
cient to each catch. This coefficient was decided upon after an analysis of the 
plankton and in view of its quantity, the basis for such an estimation being the 
coefficient test by the pumping method made under conditions most nearly 
approaching those of the catch in question. Obviously, this method, depending 
on a personal estimation, involves a source of error of uncertain extent. Conse- 
quently, it is impossible to determine the coefficient of the net without comparison 
with the results of the pumping method. Therefore it would be more practical 
to use the plankton net only for qualitative collections and to abandon this method 
entirely in all quantitative investigations of inland waters. The correct quanti- 
tative data can be obtained only by the pump method. 
