274 DISCOVERY REPORTS 



which at stage III are thin rod-shaped structures, sometimes as little as 14 % of the total length but 

 which at stage IV are thick swollen structures occupying as much as 70% of the total length. It is 

 probable that fertilization takes place when the ovary/body length percentage is between 55 and 60, 

 and that the eggs are laid when this percentage is about 70. No animals with seminal vesicles intact 

 have been taken with the ovary/body length percentage in excess of 55, and none at all with this 

 percentage in excess of 7 1 . 



Spawning takes place throughout the spring, summer and autumn, 

 and perhaps to a small extent in winter. This prolonged breeding 

 season is evident both from the fact that small-size groups are always 

 present in the population, and from the almost complete overlapping 

 of all size groups which makes the interpretation of a figure such as 

 Fig. 26 impossible during the late spring and summer months. 



Fig. 26, however, indicates the presence of several recognizable broods 

 apart from the March/April one. Traces of a later brood, probably 

 hatched in April/May, are evident at Station 965 in September, and this 

 brood may account for the 5-10 mm. group at Station 2536 in July; 

 the graph for Stations 2424 and 2425 is based mainly on this brood, 

 which is still recognizable at Station 2451 in early October and at 

 Stations 983-90 later in the same month; its individuality is lost in 

 November. It is possible that with the advent of spring the larger 

 animals of the March/April brood grow at a more rapid rate than do the 

 smaller animals of the April/May brood, and thus for a month there is a 

 noticeable gap between the peaks for the two broods. 



At Station 2451 in October the large size groups appear. These are 

 from broods hatched in late spring and summer of the previous year, 

 and have spent the winter months in deep water. It is probable that 

 the stock which gives rise to the earliest spring broods never ascends to 

 the surface after its winter migration. These large animals are absent from 

 the shallow hauls in November, and evidently begin their breeding migra- 

 tion early in that month (there are always at least a few large individuals 

 in the shallow hauls in every month). The peaks of the November graphs 

 are probably due to a combination of the March April and April/May 

 broods, the point of overlap of the size ranges of each brood giving 

 a combined total in excess of the peak of either brood separately. 



The areas of greatest density of population in the vertical sections 

 (Fig. 22) for December to May are composed mainly of animals less 

 than 20 mm. in length, and the apparent shift of the centre of density 

 of the population southwards during the summer suggests that the 

 earliest breeding takes place in the most northerly part of the Sub- 

 antarctic zone, and as the summer progresses the centre of breeding moves southwards. 



It is not possible to say much about the life cycle of the L.S. race. The normal size at maturity is 

 from 75 to 80 mm., and no mature animals in excess of 90 mm. have been recorded. If the growth rate 

 is assumed to be 5 mm. per month, as in the S.N. race, then the time taken to reach full size will be 

 15-16 months. However, the growth rate may be expected to be considerably higher in summer, and 

 a 12-month life cycle seems probable on these assumptions. 



Although most of the mature specimens of the L.S. race have been taken in March, this is probably 



Fig. 27. A large non-breeding 

 specimen of S. gazellae L.S. 

 race, 100 mm. long. 



