398 BULLETIN OF THE BUREAU OF FISHERIES 



plankton have been made by methods the reliability of which has steadily increased 

 through the medium of successive trial and criticism. Inasmuch as our Gulf of 

 Maine studies touch only the edge of this field, I may simply refer the reader to 

 Hensen himself, to Lohmann(1903 and 1911), to Steuer (1910), and especially to the 

 summaries by Johnstone (1908) and by Gran (1915) , 28 for general accounts of such 

 undertakings. Much of the earlier work of this sort was robbed of part of its value 

 by the impossibility of determining how much of the vertical column of water fished 

 through by the net was actually filtered by it. But thanks to Lohmann's (1911) 

 demonstration that satisfactory counts of many of the most important pelagic 

 plants could be obtained by centrifuging a water sample obtained with an ordinary 

 water bottle, and to Gran's (1912a; 1915) discovery of a satisfactory preservative 

 (Flemming's fluid) for such samples, a simple but exact method for quantitative 

 plankton work is now available, which it is to be hoped American biologists will 

 soon adopt. 



While this method gives far more reliable results for the smaller planktonic 

 plants, "many of the larger species," as Lebour (1917, p. 135) points out, "do not 

 get into the water samples in anything like a representative number," and as a rule 

 this method is quite worthless for the larger animal plankton. In fact, no one 

 collecting apparatus can be expected to be equally satisfactory for all the members 

 of the plankton, large as well as small. 29 



Horizontal hauls with ordinary tow nets yield useful information as to the 

 relative abundance of phytoplankton, but only if hedged about by the same pre- 

 cautions as are necessary for the zooplankton (p. 79), the need of which is now 

 universally recognized. For example, we face the impossibility of insuring that 

 all the tows shall fish through an equal column of water, because it is practically 

 impossible to keep even a steamer moving at a uniform rate at the low speed that 

 towing requires. The uncertainty introduced by imperfect filtration is much more 

 serious for phytoplankton than for zooplankton, for the much finer-meshed nets 

 that must be employed become clogged much sooner and to a greater degree. This 

 is especially the case when Phaeocystis and certain diatoms swarm (that is, just 

 when information on their abundance is most to be desired), for they often clog 

 the silk so thoroughly that the nets become quite impervious to water after a few 

 minutes, so that the catch becomes the product of the first part of the tow only. 



There is also the problem of a method of estimating the amount of phyto- 

 plankton caught, on the one hand sufficiently accurate for the results to be instruc- 

 tive and on the other rapid enough to deal in a practical manner with the large 

 amounts which horizontal tows at the surface often yield. The total volume — 

 simplest and easiest measure — is estimated by the same method as for the zoo- 

 plankton, described above (p. 81), and entails the same sources of error, the worst 

 being the uncertainty as to what proportion of the measured volume represents the 

 actual plankton and how much of its bulk is due to the spaces between its members. 



! * W. E. Allen (1921) has recently formulated a formidable list of sources of error inherent in all collections of plankton taken 

 with tow nets. 



** Lebour's (1917) tables give instructive examples of the discrepancy between net hauls and collections made with the water 

 bottle ofl Plymouth, England. 



