25 6 DISCOVERY REPORTS 



Pareuchaeta antarctica is shown by the St. WS 53 series, Fig. 131, to occur irregularly 

 in denser concentrations of at most half a mile or a mile across, and possibly much less. 



Parathemisto gaiidichandi was shown to be very patchy in its distribution, particularly 

 in the St. 150 series, where there occurred four very distinct patches, which must have 

 been less than half a mile across. The nature of its distribution can best be shown in 

 graphic form (Fig. 133). The Amphipod Vibilia antarctica will be discussed in relation 

 to Salpa fusiformis. Small patches of Cyllopus and Primno are shown in Fig. 132; the 

 distribution of the latter in relation to time of day is interesting in view of Mackintosh's 

 discovery that it apparently makes a vertical migration upwards at daylight; this is 

 discussed on p. 198. 



Euphansia superba is shown to be more patchy in its distribution than any of the other 

 forms investigated. In the series at St. 150 seven very dense concentrations occurred, 

 separated by samples of comparatively small numbers. In series St. WS 53 the numbers 

 taken of this species were not so great, and the patchiness, although considerable, was 

 not so marked. The two series are illustrated in Fig. 134. Over part of the St. 150 series 

 it is seen that the samples alternate between ones of very high catch and ones of com- 

 paratively low catch, and it might be suggested that this indicates some fault in the 

 catching power of one of the nets, which were used alternatively, but on examining the 

 series further it is at once seen that this is not the cause, since on two occasions the 

 patches elsewhere in the series are separated by two low catches instead of one — i.e. 

 whilst heavy samples R, T, V and X are separated by single light samples S, U and W, 

 the heavy samples O and R, and X and AA are separated by two light samples P and Q, 

 and Y and Z, respectively. The nets were most carefully examined throughout the 

 experiment, and except for the damage and replacement at AW were in perfect order and 

 of exactly similar catching power. 



From the fact that there is usually such a sharp contrast between a heavy sample and 

 light samples taken on either side of it, it would appear that the size of the patches of 

 E. superba must be considerably less across than the length of the tow of the net, other- 

 wise one would expect to find such patches more frequently being sampled by two 

 adjacent net hauls instead of one. Only in the case of nets AA and BB may two nets 

 have sampled the same dense patch. 1 



1 If we had a large number of observations it should be possible to calculate the approximate size of the 

 patches, or rather to say that they should not be greater than a certain size, assuming the patches to be circular 

 in shape. Professor G. C. Steward has kindly given us the following expression for such a problem: 



n = —-N, 



2.b 



where n is the number of patches which extend across a division between two net hauls, N is the total number 

 of patches observed, a is the radius of the patch which is assumed to be circular, and b the length of tow of 

 each net haul. Now unfortunately the number of observations we are dealing with in the present instance 

 is far too small to place any great reliance upon the result of such a calculation, but it does give one a pro- 

 visional and very approximate idea of their size. Thus in St. 150 there are seven dense concentrations, one of 

 which falls across two nets (AA and BB) and the length of tow is o-6 mile, so that the radius of a patch is 

 approximately not greater than 96 yards, or a patch say is not more than 200 yards across. Now it is possible 

 that one or more of these half-mile samples may have cut across more than one patch — for example, net O 



