I20 DISCOVERY REPORTS 



records of the occurrence of eggs are also taken into account, spawning begins a month or two earUer. 

 Eraser found eggs in the plankton from mid-November to March, and with this additional evidence, 

 it would therefore appear that E. sitperba lays its eggs over a period of 5I months. 



AVERAGE GROWTH RATE 



There are two ways of working out the growth rate of E. snperba : (i) by investigating the growth of 

 the developmental stages, and (2) by investigating the growth of the adolescent and adult population 

 as a whole. 



(i) Grozvth of the developmental stages. By the first method, the total catch each month was sexed, 

 measured and divided into the seven growth stages, males and females of course being grouped 

 separately. These measurements are set out in detail in the appendix. Length frequency tables were 

 made for each stage and the average monthly lengths of the stages were calculated. The average length 

 of each stage for the whole year was also worked out (see Table 5, p. no). 



These monthly and yearly averages for males and females are shown graphically in Fig. i . There is a 

 fair degree of correspondence between the lengths of the sexes at each stage, though on the whole the 

 males tend to be larger and grow more rapidly than the females. The growth of the females is slower 

 and steadier, but the same average length is reached in the final adult stage. The fall which occurs in 

 many of the curves in October and November, following a rise in September, may be due to the 

 stock being replenished, as the spring goes on, by a number of smaller specimens, whose growth has 

 been delayed by winter conditions. The general rise to a maximum in February and March, on the other 

 hand, occurs during the period of optimum conditions of food and temperature. 



In both males and females, there is sometimes considerable overlap between the average lengths 

 of the stages. This is seen in Fig. i in which these average lengths are superimposed on one another. 

 Size is again shown to be an unreliable criterion of development, a point which has already been much 

 stressed. 



In order to find out how long E. snperba takes to grow to maturity, I have worked out the frequency 

 of occurrence of the developmental stages expressed as a percentage of the catch each month. Table 

 13 A gives the actual figures^ and Table 13 B shows the months in which eggs, adolescents and adults 

 are at a maximum. 



Eraser's work has shown that the spawning season is a long one. His results indicate that the 

 "greatest production of eggs is in November-December", when two large catches were recorded. 

 But he points out that once the eggs have been spawned, there are so many factors influencing their 

 dispersal, that a rather distorted impression of their abundance may easily be obtained, if it is based 

 on records of egg catches alone. The occurrence of gravid and spawned females gives a more reliable 

 picture. My results show that these are at a maximum in February, March and April, and it seems 

 reasonable to suppose that this will also be the period for the occurrence of the majority of the eggs. 

 Rough analyses made on board the R.R.S. 'Discovery II' during the ship's last two commissions 

 indicate that this maximum for gravid females and eggs is met with a month earlier, i.e. in January, 

 and extends through February to March. Annual variation may account for this, just as an early 

 southern summer is probably the reason for the large catches of eggs recorded by Eraser in November 

 1929. My results were obtained by examining material from many seasons and localities, and may 

 therefore be regarded as being characteristic of the average catch in average circumstances. 



1 In the totals of stage 1, I have included the records of adolescents from Eraser's paper. He examined 124 young adol- 

 escents, and their measurements are given in Table 14. The majority of the lengths lie between 13 and 22 mm., and the 

 specimens can therefore be regarded as all belonging to stage 1. Males and females occur in the catch in approximately 

 equal proportions, so I have incorporated these lengths in the totals for stage i in both sexes. 



