i2 4 DISCOVERY REPORTS 



Body wall. The body wall is thin, with narrow but distinct circular muscles round the siphons and 

 narrow longitudinal muscles radiating across the body from the siphons. The siphons are small and 

 wart-like. 



Parietal organs. About the centre of the left side of the body above the intestinal loop, there is a 

 small ovoid parietal organ and another at a corresponding point on the right side. This species has no 

 atrial organs. 



Tentacles. In large specimens there are about twelve large compound tentacles and a few very 

 small ones between them. 



Dorsal tubercle. The dorsal tubercle has a C-shaped slit with the horns turned inwards or 

 spirally coiled. 



Branchial sac. On the margin of the wide dorsal lamina there is a series of distinct teeth. The 

 longitudinal bars of a large specimen are arranged as follows: dorsal line 5 (22) 3 (21) 2 (23) 3 (24) 6 

 (18) 3 (22) 2 (17) 7 endostyle. 



Gut. The gut is bent into a long flat U-shaped loop lying far down in the ventral half of the left 

 side. At the posterior end of the body is the curved oesophagus, which leads to the short pear-shaped 

 stomach with walls bearing twenty or more narrow folds. The intestine and rectum are narrow, and 

 bent so that the anus lies close to the oesophagus. 



Gonads. The gonads are long and narrow, the left one within the intestinal loop and the right one 

 at a corresponding position on the opposite side. In each gonad the ovary is long and continuous, and 

 the testis is developed as a series of about four swellings along the length of the gonad. 



Remarks. I agree with van Name (1945) that there is no reason for separating P. bouvetensis, which 

 Michaelsen (1904) described from the neighbourhood of Bouvet Island, from the much better-known 

 P. turqueti of the Graham Land region. Owing to an error of Michaelsen's it was originally thought 

 that the gonads of P. bouvetensis differed from those of related species, but now that this mistake has 

 been put right, P. bouvetensis and P. turqueti are indistinguishable. 



Distribution. Antarctic (Bouvet Island, Graham Land region, South Shetland Islands, PVictoria 

 Land as P. scotti Herdman, 1910, PEnderby Land, PMacRobertson Land as P. legumen Kott, 1954). 



The relationship of P. legumen, P. georgiana and P. bouvetensis 

 These three species constitute a natural and closely related group, but their validity as separate 

 species has been questioned. Of recent authors, Arnback (1938) and van Name (1945) have both 

 maintained the three species, although van Name was a little doubtful whether P. legumen and 

 P. georgiana should be separated. Kott (1954), however, definitely identifies P. turqueti with P. legu- 

 men, after examining material from Enderby Land and MacRobertson Land, which apparently showed 

 a great variety of external form. I cannot agree with this, as the presence of atrial organs readily 

 distinguishes P. legumen and from P. bouvetensis (syn. P. turqueti). Moreover, the locality of the 

 specimens which she identified as P. legumen was in the high Antarctic, very different from the well- 

 localized area of distribution of P. legumen in the Subantarctic. 



The rich 'Discovery' material shows, I think, that the three forms are best regarded as three 

 separate species. One of these, P. legumen, is confined to the Subantarctic waters of the Falkland 

 Islands, Patagonian Shelf and Magellan region. The other two are Antarctic species: P. georgiana, 

 probably with a very local distribution round South Georgia, and P. bouvetensis, which is circumpolar. 

 The two latter species are more closely related to one another than to P. legumen. This close relation- 

 ship is shown by both having parietal organs but no atrial organs, and by the relatively long stalk. 



The forms that we have been considering present a case of the breaking up of a population into 

 species which, at their present stage of evolution, still show their close relationship. 



