306 henry b. ward: 



the only uncertain factor, and since the organisms which escape 

 thus are the smallest, their volumetric importance may be 

 slight; but they must be investigated numerically by other 

 methods as the numerical estimations made from catches of 

 the vertical net are evidently most open to question. 



As the vertical net does not collect all the material in the 

 column of water through which it passes, various means have 

 been adopted to ascertain the portion of water actually strained 

 or the coefficient of the net. Hensen's earlier, extremely com- 

 plicated method was pointed out by Iteighard (94a) and Ward 

 (96b) to be open to question, and the former proposed an experi- 

 mental method ( !)4b) for precise determination of the efficiency of 

 the net. Hensen (95) advocates the use of a tin plate covering the 

 mouth of the net except a small opening in the center. By 

 counting the number of individuals of a well marked species 

 caught under these conditions and comparing with the number 

 caught l)y the full opening of the net, its coefficient may easily 

 be obtained. The method, however, evidently affords more 

 opportunity for error than that proposed by Birge(97a) who fil- 

 tered the entire column of water in a tin cylinder having the di- 

 ameter of the net opening in order to ascertain the coefficient of 

 the net. This was found to be about two, and the difference 

 between maximum and minimum hauls of the net was no 

 greater than that shown bv the column of water in the cylinder 

 at sucessive tests. 



For Crustacea alone Birge found that the clogging of the net 

 in an IS m. haul did not markedly affect its coefficient over 

 that employed for the o m. haul until after the rapid increase 

 of the phytoplankton in July. For the short haul the clogging 

 made at no time any visible alteration in the coefficient which 

 in the opinion of Birge is furthermore one of the most constant 

 factors, and quite as accurately determined as any other. How- 

 ever, Frenzel (97a) is inclined to think the coefficient decidedly 

 variable. Kofoid (97a) ascertained the coefficient of the net 

 in use at the Illinois station, according to the original method 

 of Ilensen, to be 1.32; experimentally it was shown to vary 



