JJ2 DISCOVERY REPORTS 



for instance, recording thick clouds of salps at the surface so densely packed that the individuals 

 were actually touching one another (see also p. 154 below). 



In January 193 1 an attempt was made in R.R.S. 'William Scoresby ' to calculate what the natural 

 density of a typical patch might be, the observations being carried out at Station WS 540 where as 

 already mentioned (p. 148) an area of widespread and prolific patchiness was encountered. As will be 

 shown later (p. 262), when a ship towing a stern net ploughs through a surface patch in daylight the 

 patch may sink bodily away or split asunder to right and left, in either event leaving an avenue of 

 clear water astern in which the net catches nothing, or next to nothing. In view of this behaviour the 

 sampling in this instance was effected by means of a net, an ordinary i-m. diameter stramin net, which 

 instead of being towed from the stern in the usual way was rigged to fish from the head of a forward 

 boom (Fig. 16) projecting 12 ft. out from the port rail, enough warp being run out to allow it to 

 fish alongside the vessel, about midships, and well clear of the side. With this arrangement, which may 

 be referred to as lateral or alongside towing, it appears that the sundering patch, although not of 

 course the sinking one, splits away directly into the path of the approaching net with the result that 

 enormous daylight gatherings can be made, involving large animals which in the ordinary way avoid, 

 or (p. 264) are caused to avoid, our daytime surface nets. Moreover, with a net so rigged, by towing 

 on the edge of a patch and by hauling or slacking away on the guide rope, it is possible to manoeuvre 

 the net directly into the patch without the latter having suffered any major disturbance resulting from 



the passage of the vessel. 



Towing thus at Station WS 540, on the edge of a surface patch undisturbed by the moving hull, 

 the lateral net was fished for 36 sec. before emerging from the discoloured area and passing into the 

 clear water beyond. During this brief operation no less than 144,976 krill were captured, the vast 

 majority of them large euphausians of the staple whale food class. As a net of this diameter towed at 

 ii knots will fish through a column of water approximately 28 cubic yards in volume in 36 sec, 

 assuming it filters the whole of the water in its path.i the figure 144,976 represents a density of 

 approximately 5200 to the cubic yard, or say one to every 8 cubic inches which, although considerably 

 less than the apparent density, is broadly in keeping with appearances. It would mean that the krill 

 in this instance were spaced at about 2-inch intervals. This, however, must be regarded as a con- 

 servative estimate since it does not allow for the fact that an unknown but probably high percentage 

 of these large euphausians must have been taking avoiding action and (see p. 258) were successfully 

 eluding the net. In other words, the actual density may well in fact have been of the order of about 

 one to the cubic inch as eye-witness accounts suggest. 



Crowding of marine animals on a comparable scale is known from other ecological niches. It is 

 estimated, for instance, that there are 3000 Thoracophelia mucronata, a Californian beachworm, in 

 a cubic foot of sand (Fox, 1957), Beklemishev (1958c) estimating that the long and narrow belts of 

 Thalia longicatidata, sometimes seen in the tropics, are crowded 2500 to the cubic metre. 



Further evidence of close surface packing by the Antarctic krill and other euphausians is provided 

 by instances of the choking of pumps by these animals, of their being washed on board vessels in 

 stormy weather, of their being involved in mass strandings or discharged from ships' hoses while 

 washing down. In northern waters, for instance, Lachambre and Machuron (1898) record an occasion 

 when the pump used by the Swedish explorer Andree in the preparation of hydrogen for his balloon 



1 In practice it would not do so. There is loss of water through 'out-thrust' and in this particular instance there must 

 have been further reduction in filtering power owing to the gradual blockage of the meshes of the net as more and more 

 euphausians were crowded into it. The figure of 28 cubic yards is, therefore, too high and the resulting estimate of 5200 krill 

 to the cubic yard in consequence too low. Igarashi (1957) analyses the filtering rate of a plankton net mathematically, showing 

 that it decreases directly with lapse of time and inversely with the speed of towing. 



