RHINCALANUS GIG AS 285 



and conditions under which the hauls were necessarily carried out, it is an almost 

 negligible error. Greater commencing depths than 1 00 m. give towings varying somewhat 

 more widely from the length aimed at. A commencing depth of 180 m. gives an oblique 

 towing of about 1305 m., 70 m. too short. Even this, however, represents an error of 

 only 5*3 per cent, and it was seldom that the surface net reached so great a depth as 

 180 m. The commencing depth was more usually between 100 and 150 m. — the latter 

 depth giving a towing of about 1340 m., an error of only 2-2 per cent. 



The deep hauls, when plotted in the same way, show greater errors, which arise 

 mainly from the varying depths at which the net closes. The depth aimed at, 250- 

 100 m., gives a towing about 990 m. in length, and it may be seen from Fig. 4 that such 

 a haul as that made at St. 802, 320-70 m., gives a tow of about 1275 "">•' ^ difference 

 from that aimed at of 285 m. A towing such as that at St. 776, however (356-170 m.), 

 gives a haul of about 835 m. only, a difference from that aimed at of 155 m. 



Nevertheless, in spite of the wide variations in the depths through which the deep 

 net was towed, it was decided to make no attempt to apply a standardizing correction 

 to the catches, firstly because it is impossible to find any simple or constant correction 

 to apply, and secondly because the nets used cannot be regarded as instruments which 

 give a quantitative estimate of the plankton. The results obtained with them must be 

 looked upon as strictly qualitative, and the numbers in the catches must be taken to 

 indicate only relative abundance or relative scarcity. 



The path of the net actually traced out during an oblique towing is never under any 

 circumstances a perfectly straight oblique line, since, as already explained, the net rises 

 and falls with variations in the speed of the ship, which can never be kept absolutely 

 constant. Again, very many factors introduce inaccuracies into the results obtained with 

 towed nets of this description, such as the local swarming of the plankton, variations 

 in the depth of the plankton with the weather, and the possibility of the nets fishing 

 to some extent when being paid out or handled at the surface and so on. Hardy and 

 Gunther (1935) have drawn attention to this general aspect of plankton investigation 

 (p. 27) and it may be once more emphasized here. "We are in this work concerning 

 ourselves only with big differences ; the very nature of the distribution of the plankton 

 we are studying and the necessary limitations to our methods in the field will not allow 

 us to attempt the establishment of small differences. When we are comparing one 

 region represented by 5000 Corethron with another represented by 562,000, what does 

 it matter if that 5000 is really 7500 or 2500, or again if the 562,000 is really 281,000 or 

 743,000? " Similarly with Rhincalanus gigas, if we are comparing one region represented 

 by 15,300 R. gigas with one represented by 5730 it makes no difference to the final 

 picture of the distribution of the species if the former number is really 13,250 or 18,470, 

 and the latter number 7250 or 4320. So that even if a standardizing correction could be 

 applied to the hauls the labour of applying it would certainly not be repaid. 



As will presently be seen (pp. 286-7) we are in this report concerned with two layers of 

 water in the area studied, an upper and a lower one. The boundary between these two 

 layers of water is usually only known to the nearest 50 or 100 m., so that the length of 



