520 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



that the influence of this loss, on the constitution of the 

 catches and the results given by numbering, was of no 

 essential importance, since the mass of the forms slipping 

 through was only small in comparison to the quantities 

 caught by the bolting silk ; and, in any case, llensen gave 

 his numbers as a minimal value, recognising that a loss 

 must occur. Kofoid in 1897 (8) through new investiga- 

 tions in the fresh waters of North America came to the 

 conclusion that the loss which enters into the results, 

 when nets of " Miillergaze " are used, w r as much greater 

 than Hensen had supposed. By using filters of hardened 

 paper, he demonstrated that only 2-50 per cent, of the 

 organisms w^ere caught by the net. Lohmann (13 ), when 

 investigating the Appendicularia, also found that the 

 quantity of small forms going through the net must be 

 of far greater importance than was formerly supposed. 

 Nothing shows the loss more distinctly that the investiga- 

 tion of the food of the plankton organisms themselves. 

 One finds in their alimentary canal the remains of the 

 smallest diatoms, Peridinians, Coccolithophoridae and 

 SilicofLagellates, of which an ordinary net used m the 

 same water in which the " devourers " (Pteropods and 

 Appendicularia, &c), lived, would contain none or few. 

 As showing the importance of this loss, Lohmann 

 mentions the fact that the Coccolithophoridae, which 

 play a great part as food for the plankton animals of the 

 North Sea, are almost unrecorded in the tables when 

 bolting silk nets are used. 



The most favourable organism on which to study 

 the food of the plankton animal is the Appendicularian, 

 which does not take food directly into the alimentary 

 canal, but secretes a special structure, the " house/' for 

 the purpose of catching its food. This is a perfectly 

 transparent structure, and under the microscope can be 



