498 DISCOVERY REPORTS 



counting a very small fraction, possibly ^V or jfo, and making an estimate. The tow-net 

 sample to begin with is taken from a confined area of sea, perhaps not more than a quarter 

 of a mile tow when using a fine or medium silk. Thus by the tow-net method there are 

 two large sources of error in estimating the quantity of plankton in the area examined. 

 The sample itself may not truly represent the area as a whole owing to the patchiness 

 of the plankton ; then after having taken so large a sample there is the laboratory error in 

 taking the fraction for examination. In the stempel pipette method, or other means of 

 fractioning, all organisms may not be taken up in their true proportions. The recorder 

 method takes a smaller sample over a wide traverse of sea, eliminating to a large extent 

 both the factors of patchiness in the sea and the fractioning in the laboratory. When 

 a fraction only of the section can be examined, that fraction is not such a small one 

 as is usual in tow-net sample analysis, and the sample is deposited on the gauze by the 

 sea itself so that the organisms are presented for estimation in exactly their true pro- 

 portions. The gauze section represents the "telescoping" of the plankton in a long 

 volume of sea into a small two-dimensional space for the convenience of examination. 

 If each section is regarded as a sample we may say that the recorder method presents 

 us with a large number of small but representative samples spread over the whole line 

 of the traverse instead of a few large tow-net samples taken at intervals. 



LIMITATIONS OF THE METHOD 



As in all methods there are limitations, and before making use of the instrument for 

 any particular investigation they must be carefully weighed against the possible advan- 

 tages. There appear to be two major limitations. Firstly, in regions of rich plankton 

 the small filtration area of silk is apt to become somewhat clogged so that filtration 

 and catching power are reduced. This difficulty applies, although to a lesser extent, to 

 ordinary tow-nets. It is perhaps not so serious a limitation as it may at first sight appear ; 

 it means that whenever organisms are taken in very large numbers one must realize 

 that the actual numbers present in the sea are very much larger. The peaks on the graphs 

 made from the records should be higher than they actually appear. A study of the 

 published examples of such graphs (Hardy, 1935, 1936) will show what great differences 

 in numbers may be recorded ; clogging in fact does not conceal these differences ; there 

 is no evidence of the numbers reaching a limit to produce a more or less uniform 

 maximum deposit of plankton. It must be remembered too that continually there is a 

 fresh surface of gauze being presented to the water flow as the thick deposits of plankton 

 pass on into storage. 



The second and more serious limitation is that due to the use of only one instrument, 

 making a record at a uniform depth when some organisms are carrying out diurnal 

 vertical migrations. The influence of vertical migrations on the records is clearly shown 

 in some of the graphs obtained in deep-water areas (Hardy, 1936), but does not appear 

 to have a pronounced effect in the survey over the shallower regions of the North Sea 

 where the machines are run at a uniform depth of 10 m. (see Hardy, 1935, fig. 10). An 

 ideal arrangement would be the towing of two or more machines one below the other at 



