CAUSES OF PATCHINESS 233 



ment and staging of a very large number of samples covering the later developmental phases (Figs. 

 39-42) demonstrating in sample after sample a normal length or stage frequency in which the value of 

 the mode, as in the purely larval samples, is persistently high. 



As with the larval samples. Figs. 38-42 are again arranged in order of advancing time, beginning 

 with July, when the first adolescents appear, and ending with April, 21 months later, when the last 

 spent females are found in the plankton, the material on which the length and stage frequencies 

 are based having been chosen from the Weddell or northern zone, since it is there we have obser- 

 vations and measurements covering every month of the year. The length frequencies, expressed in 

 2-mm, groups from July up to the beginning of the second half of November, and in 4-mm. groups 

 from then onwards, appear in the upper part of each figure, the stage frequencies, where determined, 

 in the lower. The number of individuals measured, or both staged and measured, from each sample 

 is shown by the large unbracketed figures, the figures in brackets showing, where determined, the 

 dominant stage of the female in each sample. Thus, following Bargmann (1945, pp. 1 13-15), i, 2, 3, 4 

 and 5 indicate females in from very early to late adolescence, 6 indicating paired females, 7 ( = Barg- 

 mann 7 a) gravid females, and 8 (= Bargmann 7B) spent females. 



The normal length frequencies and high modal values repeatedly encountered in the Weddell 

 samples are, of course, equally phenomena of the East Wind drift and of the cold northward encroaching 

 tongues of water to which it gives rise. In fact wherever and in whatever developmental phase this 

 species had been fairly sampled the samples obtained, as the many developmental diagrams presented 

 in the distributional section show,^ point everywhere to its congregating in discrete groups in which 

 the length frequencies are persistently normal and the modal values persistently high. 



The final condition of the 24-28 month old swarms, incorporating the data from the East Wind 

 drift, is shown in Fig. 43 in which the length frequencies reveal a tendency for the animals in 

 the Weddell swarms to be somewhat larger than their contemporaries in the East Wind zone, sug- 

 gesting that, taking the krill population as a whole, it is in the lower latitudes that the adults attain 

 their greatest length. Even Fig. 43, however, probably does not reveal the true condition of the 

 adult swarms. It is based on townet samples and since there is an escape factor involved, particularly 

 (p. 262) where the largest animals are concerned, it may well be that in both Weddell and East Wind 

 samples we have repeatedly been recording modal lengths of distinctly lower value than perhaps would 

 be found in the swarms in their natural state. This is distinctly suggested by the condition of the adult 

 samples (p. 143, Fig. 14) collected by Ruud and Crimp from whales' stomachs. 



Turning again to the larvae, it will be seen that with the advance of the spawning season the age 

 pattern, or modal values, of the surface swarms (Figs. 34-7), particularly in those encountered 

 in the Weddell stream from March to June, becomes increasingly complex, the surface population 

 as a whole becoming more and more heterogeneous as successive batches of eggs hatch out and the 

 existing swarms become augmented by others of later generation. In the East Wind drift on the other 

 hand, because of the compressed spawning season and very slow growth-rate (p. 355) characteristic 

 of these high latitudes, the corresponding age pattern from January to April is largely homogeneous. 



It would appear then, from the random sampling of our nets, that the patchiness of E. superba 

 arises from the congregating of this species from hatching onwards in a system of swarms, each 

 discrete and complete in itself, the individuals of it, especially as larvae, all, or very largely, in the 

 same developmental phase. Or to put it in another way, the enormous variation in catch-figures 

 our nets reveal, as Motoda and Anraku (1955) have put it, springs from the 'statistically non-random 

 distribution ' of the organisms we have been sampling. It would appear, too, that at any given time 

 or place an individual swarm need not necessarily be composed of euphausians of the same age as that 



^ See particularly p. 397, Fig. 136. 



26 nM 



