F-20 



The major factors that have been positively identified as being associated with swarming 

 are therefore light and feeding activity. Reactions of planktonic animals to light have been 

 well documented by Gushing (1951)1 Stasenko (I967) and Burukovskii (196?) describe krill being 

 attracted to a red light source but moving away from a strong white light source while Ivanov 

 (1969) has drawn attention to the luminescence of krill in swarms. Fiirther information is 

 clearly needed to identify a possible wide variety of reactions by krill to light. 



The suggestion by Pavlov (I969) that krill swarm as part of a feeding cycle implies 

 dependence to a great extent on food availability - the more food that is available the more 

 rapidly the animals become replete and swarm. As mentioned earlier a direct result would be 

 that swarms would tend to appear during the early summer and be most frequent in abundance 

 during the summer peak of primary production. Hardy and Gunther (1935) showed that in the 

 South Georgia area the densest concentrations of phytoplankton were not coincident with the 

 largest concentrations of krill (or other zooplankton) (see Table 4«l)» 



This general pattern is confirmed by Avilov et al. (I969) who found the largest concen- 

 trations of krill in areas of the Scotia Sea where the mean biomass of phytoplankton was 

 I-5 ml/m^. In areas where the mean biomass of phytoplankton was less than O.5 ml/m-^ or more 

 than 5 ml/m they found few krill concentrations. This observation is not confirmed by Shust 

 (1969) who reports numerous krill swarms in areas where the phytoplankton density varied from 

 0.1 to 8 ml/fflJ, He does however relate duration and size of swarm to phytoplankton standing 

 crop — swarms persisting longer in areas of high phytoplankton standing crop. It is possible 

 that the inverse relationship reported by Hardy and Gunther (1935) and Avilov et al. (I969) 

 is due to grazing effect whereby there is a time lag before the herbivores eat down the phy- 

 toplankton (Gushing et al. I963 and Gushing 1975)* Tha direct relationship in that case wovild 

 be expected during the initial phytoplankton bloom in areas where overwintering herbivorous 

 zooplankton concentrations occur. Although this explanation is not proven in this instancei 

 it does fit the facts and therefore indicates that studies investigating this interaction, and 

 thus by implication other factors that might control krill raicrodistribution, should take 

 accovmt of short-term changes. 



Swarm Size 



Estimates of swarm size have been made by both visual observation and analysis of echo- 

 sounder records. Marr (I962) published a series of drawings of the horizontal outline of 

 several swarms that were observed near to the surface. These swarms were all quite amorphous 

 in shape and varied in size from a few to several hundred metres across and were continually 

 changing shape in an amoeboid fashion. Nemoto et_ al. (in press) describe surface swarms as 

 being generally round or oval in shape and that some changed shape as a result of wind stress 

 becoming more oval with the greatest axis at right angles to the wind direction. 



A comparison of swarm size between areas is made by Nemoto et al. (in press) who con- 

 sidered that swarms in the Scotia Sea area were on average smaller than those in the Queen 

 Maud Land area. 



In the vertical plane swann size has been mapped using echosounders. Certain of the 

 small swarms are so dense that it is impossible to d^efine the vertical size. However, where 

 density within the swarm is lower, authors have mapped the vertical distribution and size of 

 swarms. In an analysis of this type, Shevtsov and Makarov (I969) describe various patterns 

 in the changing shape of swarms observed in the Scotia Sea. From their observations it 

 appears that there are a few common patterns throiighout the area. The following types of 

 vertical movement and configuration are described in Table 6.6. 



Formation of two layers (Table 6.6) during the day and amalgamation at night with a 

 tendency for adult krill to be at the deeper levels seem fairly standard. A much greater 

 vertical movement was observed "by Mohr (1976) who, using an echosounder, followed a concen- 

 tration of krill in the vicinity of the South Sandwich Islands for 6 days in April I976. At 

 midnight the krill were massed in the top 20 m but by O6OO hours they had migrated to a depth 

 of 70-110 m, to slowly migrate back to the surface during the following 10 to 12 hours. 

 Since all these observations were made in a fairly brief period of time during one season, 

 more information will be required to amplify these observations and conclusions. 



