254 DISCOVERY REPORTS 



very uncertain when such work will again be possible, so that it seems the more desirable 

 that the data, and a possible interpretation of them, should be published without delay. 

 The methods we adopted were: vertical hauls with the Gran International Net from 

 ICO to o m. ; vertical hauls with a modified form of Harvey's apparatus (Harvey, i934«) 

 and centrifuging of water samples. 



The routine hauls with the Gran Net (N 50 V) of 50 cm. diameter at the mouth, 

 and made of the finest grade of bolting silk, had been fished throughout the previous 

 work of the Discovery investigations. By analysing the catches by the well-known 

 Hensen's methods it was possible to gain some idea of the grosser quantitative changes. 

 The method is very useful for qualitative purposes, as it provides a large amount of 

 material in good condition in a short space of time without the necessity for having a 

 phytoplankton specialist on board to deal with the samples immediately. It was there- 

 fore particularly valuable during the pioneer stages of the investigations when we had 

 little knowledge of the general distribution of the phytoplankton, and had a limited 

 staff distributed over two ships and a shore station. It was realized from the first, 

 however, that such hauls can only provide a very rough idea of even the grossest quan- 

 titative changes (Hardy, in Hardy and Gunther, 1935, pp. 26, 27, 40; Hart, 1934, 

 pp. 15-17). Therefore, as soon as it became possible to adopt better methods, we fished 

 the Gran Net mainly to ensure an abundant supply of material for subsequent taxonomic 

 work. It may still provide the best means of studying the general distribution of some 

 of the larger and rarer diatoms (Hart, 1937), but apart from such special studies the 

 analytical work has been concentrated on the other two methods. 



Harvey's method consists essentially in applying the assimilatory pigment extract 

 colour match, first introduced into marine plankton work by Kreps (Kreps and Verj- 

 binskaya, 1930), to the catch obtained from a measuring net. Harvey (1934 «, p- 762) 

 tells us that Nansen was the first to suggest the use of a measuring net for plankton 

 studies. The co-ordination of the two ideas and the elaboration of a successful working 

 technique are, however, quite new. We found certain structural modifications necessary 

 to suit our own special conditions, but the dimensions, working parts, and silk nets were 

 identical with those of Harvey's own model. Our subsequent treatment of the catches 

 by digestion with 80 % acetone and direct visual comparison of the coloured extract 

 with the nickel sulphate mixture were carried out exactly as described in Harvey's first 

 account of the method (19340, pp. 770-1). Quantitative counting was not attempted, 

 but during the third and fourth commissions all the catches were examined micro- 

 scopically at sea and the dominant species noted. During the fifth commission all the 

 catches were subjected to a more thorough microscopic examination, usually when 

 fresh. A ' qualitative count' was made from a large wet mount prepared from the well- 

 mixed sample, which usually involved the examination of some thirty fields of the 

 microscope, but varied considerably according to the size of the catch. The numbers 

 of the leading forms were then reduced to percentages. 



Our modifications of Harvey's original design and method of fishing the apparatus 

 were introduced to increase its strength and reliability, even at the cost of some loss of 



