F-5 



Little evidence is available on the rate of flow of the various deep water masses 

 although Marr (p. 212) does ijive evidence that in places the Antarctic Bottom Water could 

 flow north with considerable force. 



In a recent re-examination of the distribution of efcgs and larvae ,Voronina (1974) states 

 that spawning occurs in oceanic waters (cites DolzhonkoT 1973 and Hakarov 1973) and also at 

 South Georgia, a mar.fp-nal area of its distribution range (cites Mackintosh 1973 and Makarov 

 1972 )i/ and suggests that spawning probably occurs over most of the range of the species. 

 Voronina suggests that the Antarctic Bottom Water rather than carrying the eggs and larvae 

 does in fact form the physical lower limit for the sinking eggs. Successful development is 

 assumed to be a direct fxinction of depth and a maximum depth of 1 8OO ra is ouggcsted as being 

 the greatest vertical distance through which the larvae can migrate using the limited food 

 resources in their yolk sacs. 



The 1 800 m boundary layer does encompass most of the localities in which early larvae 

 have been found although Voronina does mention that some of the stations from which they were 

 taken in the Scotia Sea fall outside this limit. Another criticism is that in many cases the 

 larvae will not arrive in the surface water until late in the season after the peak of primary 

 production. Adolescent and adult E^ superba have been shovm to be capable of feeding on 

 detritus (Pavlov 1974) and there would seem no reason why the larvae should not do the same. 

 In that case there is only an advantage in being at the surface during the primary production 

 season. Thus althou^ the 1 8OO m boiindary layer may for the most part describe the northern 

 limit of successful spawning this is probably a coincidence. 



In his analysis of the life cycle of E^ superba . Mackintosh (1972) to a great extent 

 avoids the direct issue of oceanic versus shelf area spawning and concentrates on the dis- 

 tribution of the subsequent developmental stages. However, the implication of his conclusions 

 is that he, like Voronina, considers that spaiiming occurs over much of the ocean as well as 

 the shelf area. He makes a strong case for the early larvae being distributed in areas where 

 low sea temperatures extend to great depth (see also Marr, Fig, 81 ). The major areas he 

 identifies as holding early larvae are the Ilhst Wind Drift, Bransfield Strait and the diver- 

 gence zones he describes in the Scotia Sea and IVeddell Drift. 



Vertical distribution of later stages 



In a long and detailed analysis of the "Discovery" samples, Marr (1962) suggests that the 

 major concentrations of _E^ superba occur in the top 100 m and within this stratum most fre- 

 quently in the top 10 m. In his analysis Marr is careful to minimise error due to avoidance, 

 a major problem with the 1 m .diameter and smaller nets from which his samples were obtained. 

 Althou^ massed in the surface waters for most of the day and night Marr does sug.gest that 

 there is a pattern of diurnal migration but was unable to specify the conditions under which 

 it occurred. The supporting evidence that Marr cites, based on the depths at which whales 

 and birds feed, is not particularly helpful in this context as it only confirms that krill 

 occur near the surfsuie at some time during the day or night (a fact known from the net hauls). 

 What is perhaps more important is that there is a clear indication (Marr 1962, Table 35) that 

 adolescent and post larval krill are not present in si^ificant numbers in the V/arm Deep Water. 



Recent Russian research using echosounders has shown a tendency for sexually mature krill 

 to undergo diurnal vertical migration in the top 80 m of the water column (Shevtsov and Makarov 

 1969), and Pavlov (1969| 1974) has obtained results indicating that this vertical migration 



1/ Ifekarov (1972) refers to spawning in the frontal zone between Weddell and Bellingshausen 

 water which would be oceanic. Mackintosh (1972) does not describe spavming as occurring near 

 South Georgia but considers it possible although Marr (I962) considers that no successful 

 spawning occurs in the region. 



