338 DISCOVERY REPORTS 



surface of each is even and only minutely hispid ; and in most of them the papillae are 

 less than 3 mm. high, but in two they are 5 mm. or more high and agree closely with 

 those of the holotype. As regards the spiculation, the larger megascleres in most of 

 these nine specimens have the same type of base figured by Thiele {loc. cit.) for 

 the holotype, a few show traces of the annulation recorded above for the first group 

 of specimens, and in a few this annulation is fairly strongly marked. In some of the 

 specimens, also, there is no trace of the polytylote megascleres. We see in this group 

 therefore a series of typical specimens, together with all transitions between them and 

 the group of abnormal specimens described above. 



The differences between the typical form and Hentschel's var. sifnplex are so slight, 

 especially when compared with those found in the present specimens, that I can see no 

 reason at all for maintaining this variety. In fact, if we compare the relation between the 

 holotype and Hentschel's specimen with the relations between the various individuals 

 recorded under P. jnammillaris, or any other species of Polymastia, there is not the 

 slightest justification for the introduction of this variety. 



It is probable, to judge by the written description, that Brdndsted's (1923) P. gramdosa 

 from the Auckland and Campbell Islands will also prove to be synonymous with 

 P. isidis. 



Distribution. Falkland Islands; Wilhelm Land. 



Polymastia invaginata, Kirkpatrick. 



P. invaginata, Kirkpatrick, 1907, p. 271; /(/., 1908, p. 15, pi. xii, fig. i, pi. xiv, figs. 5-15; 

 P. invaginata, var. gaiissi, Hentschel, 1914, p. 49, pi. v, fig. 4; P. invaginata. Burton, 1929, 

 p. 446. 



Occurrence. St. 42: South Georgia, 120-204 m.; St. 156: South Georgia, 200-236111.; St. 167: 

 South Orkneys, 244-344111.; St. 177: South Shetlands, 1080 m. 



Remarks. There are three specimens of this species which, although typical in all 

 essential respects, difi"er in appearance and in the disposition of the spicules ; and these 

 diff'erences seem to have a bearing on the problem of the extrusion of spicules. The 

 first specimen is cupola-shaped, 2 cm. across and i cm. high, with smooth minutely 

 hispid surface and normal spiculation. That is to say, the large megascleres are 

 arranged in regular radial bundles and the smaller spicules in a dermal palisade and in 

 stellate groups within the choanosome. There are no large spicules protruding at the 

 surface. The second specimen is sub-spherical, 2 cm. in diameter, and in this the surface 

 is smooth, or only minutely hispid, for the most part, but bears several patches of long 

 protruding spicules. The general appearance is, in fact, very like that of a long-haired 

 mammal in the last stages of shedding its coat. Moreover, the spicules composing these 

 shaggy patches are not firmly embedded in the tissues, but may be removed with the 

 greatest ease, and without the use of the slightest force, with a pair of forceps. 

 Internally, the spiculation is not quite normal. The radial bundles are normal, but the 

 characteristic stellate groups of small megascleres are not to be seen, these spicules being 

 irregularly scattered through the choanosome. The third specimen, incomplete but 



