406 BULLETIN OF THE BUREAU OF FISHERIES 



even more impressive, Johnstone (1908, p. 163) has calculated that on the basis 

 of this haul "every drop of sea water from this part of Kiel Bay contained some 

 200 diatoms;" and though by Hensen's (1S87) calculations less than one-tenth as 

 many diatoms as this are present on the average in the West Baltic, their numbers 

 are sufficiently appalling when extended to any considerable sea area. 35 



During the years that have passed since Hensen's pioneer studies in this field 

 many similar counts have been made in the Baltic and in various parts of the North 

 Sea, with the details of which it is unnecessary to delay here. 30 Lohmann, for in- 

 stance (1908, Table B), has recorded some very large counts by the centrifuge 

 method, including 7,SOO,000.000 Skeletonema per cubic meter in June, 1906, with 

 another individual catch of about 2,000,000,000 diatoms in Kiel Bay on April 11, 1906. 

 As still another example of the results of this modern method, the accuracy of which 

 leaves little to be desired, though, as Gran (1915) himself points out, it is not of 

 universal application, I may quote his own average of about 228,000,000 diatoms 

 per cubic meter in the surface waters of the Skager-Rak for Februarj', 1912. 



The centrifuge, however, is not the "last word" in quantitative determination of 

 the phytoplankton, for E. J. Allen (1919) has recently essayed the following totally 

 novel procedure: To a small sample of sea water (0.5 cubic centimeters) he added a 

 large amount (1,500 cubic centimeters) of a nutrient solution that had previously been 

 found suited for the cultivation of marine diatoms (Allen and Nelson, 1910; E. J. 

 Allen 1914). The culture was then examined after a period of incubation, where- 

 upon he found a total of 232 different kinds of organisms. A second experiment 

 yielded similar results. Since now it is obvious, to use his own words (E. J. Allen, 

 1919, p. 4), that each of these organisms "must have been represented by at least 

 one individual or unit, either cell or spore, in the original K cubic centimeter of 

 sea water from which the experiment was started," the latter must have contained 

 at least 464 organisms (mostly diatoms) per cubic centimeter — that is, 464,000 per 

 liter — and probably, as he calculates, as much as 1,000,000 per liter for the part of 

 the English Channel whence his sea-water sample was taken. How much more 

 effective this method is than centrifuging, even for such comparatively large organ- 

 isms as diatoms (for which the culture method is particularly well adapted, as 

 indicated by their great predominance in the final product), is illustrated by the fact 

 that whereas the two culture experiments call, respectively, for 378,000 and 290,000 

 diatoms as the absolute minimum per liter, centrifuging a similar sea-water sample 

 at the beginning of the experiment revealed only about one-thirtieth as many. Nor 

 can even the method of the culture medium be relied on to give a total census of the 

 phytoplankton. because it is by no means certain that the nutritive fluid employed 

 was as suitable for the growth and reproduction of peridinians, infusorians, coccoli- 

 thophorids, etc., as it was for diatoms. In short, as Herdman says (1920, p. 819), 

 "every new method devised seems to multip'y many times the probable total 

 population of the sea." 



3a There has been much discussion as to the reliability of numerical results yielded by nets of the "Hensen" type, owing to 

 uncertainty as to their coefficient of filtration. In the present connection it is enough to point out that in any case the ostensible 

 results are always smaller, never larger, than they should be. 



» For details of such I may refer the reader to Hensen (1887) himself, Driver (190S), Lohmann (1903 and IMS), and Qran (1915). 



