238 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
pull. Hensen takes this pull component accordingly as 
d sin a, and finds that the reckoning can then be satis- 
factorily carried out which is not possible with s. sin a. 
The net consists of a conical part, whose filtering 
area is C, and a cylindrical bucket whose filtering area 
is H. ‘The equation runs 
V(s—d) . Obes Ew(d) + Cw [(1+sin q)d] 
If F is multiplied by cos a and the product added to 
C, one obtains N= C + E cos a with only a very unim- 
portant error, as conical area for the whole of the filtering 
tissue and the final equation becomes— 
V(s—d) . O=N.. W [(1+sin q)d] 
This is applied in the same way as the equation for 
the cylindrical net, trial values of d being taken until 
both sides agree. 
Such was the theoretical method employed by 
Hensen. It is certainly extremely ingenious, but at the 
same time very laborious, and finally, as Hensen himself 
shows, the coefficient @ is only approximately correct, 
and must be corrected empirically. 
It is far better, therefore, that the whole calculation 
be made empirically, if possible, and this is both easy 
and reliable, that is, of course, in comparison with the 
theoretical method that has been described, which I feel 
to be unsatisfactory. Certain points, however, remain 
to be considered. It appears that the coefficient of 
filtration for a net can and does vary within rather too 
wide limits. This has been emphasised by Kofoid, who 
pointed out that the stoppage due to deposits of organisms 
clogging the meshes altered considerably the coefficient. 
This can take place very easily in the sea when using 
No. 20 silk, and the catches are made at times in spring 
or autumn when diatoms are very abundant, In tropical 
