PLANKTON 263 



deal with relatively small volumes (one, three or five litres) 

 of the water, and it must remain doubtful whether the 

 same organisms in the same quantity would have been 

 present in the next bucketful of water that might have 

 been taken from the sea. 



Even if we had no hope of attaining to greater accuracy 

 our present planktonic results are of some value. Although 

 estimates which may be 50 per cent, wrong in either direction 

 do not justify us in calculating exactly the number of 

 organisms or of potential food present per area of sea or 

 volume of water, they do give us a useful approximation. 

 Even if 100 per cent, out, doubling or halving the estimated 

 number is a relatively small variation compared with the 

 much larger increases and reductions, amounting to, it may 

 be, ten thousand times in the case of Diatoms, ten to fifty 

 times in the Dinoflagellates and five to twenty times in 

 Copepoda, which we find between adjacent months — and 

 even greater differences if we take groups of months—in a 

 survey of the seasonal variations of the plankton. 



Successive improvements and additions to Hensen's 

 methods in collecting plankton have been made by Lohmann, 

 Apstein, Gran, and others, such as pumping up water of 

 different layers through a hose-pipe and filtering it through 

 felt, filter-paper, and other materials which retain much of 

 the micro -plankton that escapes through the meshes of the 

 finest silk. Use has even been made of the extraordinarily 

 minute and beautifully regular natural filter spun by the 

 pelagic animal Appendicular ia for the capture of its own 

 food. This grid-like trap, when dissected out and examined 

 under the microscope, reveals ^ surprising assemblage of 

 the smallest Protozoa and Protophyta, less than thirty 

 micro -miUimetres in diameter, which would all pass easily 

 through the meshes of our finest silk nets. That the 

 regularity of the meshes in the silk rapidly deteriorates with 

 use is seen from a comparison of Plate XXI, Figs. 1 and 2. 



The latest refinement in capturing the minutest -known 



