third uropod rami was such that it could best be described as “coarse”, very definite and deep. So that P. australis was definitely present, and also a species which it seemed could only be P. gracilipes. Having regard to the cosmopolitan distribution of P. gaudichaudii, there seems no good reason why P. gracilipes should not be found in the Southern Hemisphere. The possibility that these specimens could be P. japonica was not supported by Bovallius's descriptions and figures, which show considerable differences in the uropods—especially since the 3rd uropod rami of P. japonica are equal in length, and the inner ramus of the 2nd uropod is, according to the description, serrate on both margins instead of only one as in these specimens. There is, however, considerable similarity between the two species. Should further work lead to the conclusion that they are synonymous, P. gracilipes has precedence of nomenclature.

A third set of specimens was distinguishable within the material. These were animals with a generally larger size (up to 18 mm.) with distinct dorsal spination, and no sign of serration or pectination on the inner ramus or peduncle of the 3rd uropod. They proved to be identical with Stebbing's E. thomsoni and so with P. gaudichaudii, into which species Barnard's key separates them.

From a practical viewpoint, separation proved much easier than would be expected when dealing with large numbers. The doubts which arise mostly do not concern confusion of P. gracilipes and P. australis but P. gracilipes and P. gaudichaudii. Small specimens of P. australis showed large coarse pectination; larger specimens of P. gaudichaudii had no pectination; but many P. gracilipes, particularly females, were so devoid of pectination on the inner ramus of the 3rd uropod as to raise doubts about their correct identification.

Here, however, comparison of the station lists shows that P. gaudichaudii specimens, as we had identified them, fell almost completely outside that 50-fathom zone; and that our P. australis and P. gracilipes fell into clearly defined areas on the other side of the P. gaudichaudii zone with only odd incursions into it. It does appear likely that P. australis tends to be concentrated a little in the 50–100 fathom area, but evidence on this point is not conclusive. There was definitely a clear separation of P. australis and P. gracilipes from P. gaudichaudii. Graphs of diurnal variation and plots of temperature and salinity to show habitat grouping show the same phenomenon. There is, in view of this evidence, no doubt in my mind that we are dealing at least with three separate populations of animals, showing some differences in morphology and considerable differences in ecology.

It is extremely difficult to set down in words further crucial differences than these: P. australis may be separated from P. gracilipes and P. gaudichaudii by its stockier build and the serration of the 3rd uropod peduncle; and its pleon segments, excluding telson, are in ratio to total length of 1:3:8, whereas P. gracilipes shows a ratio of 1:3.0. That is, in a specimen of approximately 10 mm. the three pleon segments in P. australis are 2 ½ mm. long, and in a similarly sized P. gracilipes they are nearer 3 ⅓ mm. P. gaudichaudii may be separated from P. gracilipes and P. australis by the lack of serration along the inner margin of the 3rd uropod inner ramus and peduncle, and by the extremely obvious dorsal spination of the adults. (But as noted above, young P. gaudichaudii and young and female P. gracilipes may be hard to distinguish.) There are numerous differences in the specimens which I have figured, but they are all small and few, if any, could not be attributed to sex or age differences. I have, however,