igo DISCOVERY REPORTS 



The Discovery specimens, as far as they have been examined, all agree with Skogs- 

 berg's type, G. miilleri. While adopting this specific name, I must emphasize that the 

 differences between Skogsberg's two types are very unsatisfactory. In fact, in only one 

 point is there any certain difference, and the value of this cannot be estimated until the 

 original specimens collected by M filler have been re-examined. 



Thus, as regards the length of the shell — 14-16 mm. in Miilleri and 23 mm. in 

 Agassizii — it is obvious that this character alone would be insufficient on which to base 

 a new specific diagnosis. 



The last joint of the antennule is very difficult to distinguish. Skogsberg states 

 (1920, p. 204) that it is "often somewhat withdrawn into the seventh joint and rather 

 difficult to verify with certainty". From an examination of the Discovery specimens 

 I can fully support Skogsberg's remarks. In some specimens it is impossible to be 

 certain that the eighth segment is present, while in others it is well marked. It is obvious 

 that in Midler's original specimens the apparent absence of the eighth segment might 

 be due to the small number of specimens, or to the method of preserving them. 



The number of cleaning bristles quoted by Muller is more than 200, and the words 

 he uses, "ich schatze jederseits fiber 200" (1895, p. 158), indicate clearly that this is 

 only an estimate — no attempt was made at an accurate count. Skogsberg, however 

 (1920, p. 213), states that the number of bristles present is "very difficult to determine 

 with certainty on account of the closeness of the bristles to each other ". He has counted 

 the bristles in his specimens and arrives at a figure of not more than 140. This agrees 

 with my count of Discovery specimens, but I can well believe a rough estimate of the 

 number being put as high as 200. The packing of the bristles at their bases gives a false 

 impression of their number. 



The length of the joints of the antennal endopodite cannot be relied upon, as Mfiller 

 gives two figures which do not agree. His smaller figure (Plate III, fig. 1) agrees 

 approximately with Skogsberg's statement, and with a figure by Scott (191 1) (Plate II, 

 fig. 6) of a form that Skogsberg claims (p. 217) as G. miilleri, while it is upon the larger 

 figure (Plate I, fig. 21) that Skogsberg depends for his measurements of G. agassizii. 



The last character on which Skogsberg distinguished his new species from the original 

 is the only one which cannot be questioned, except on the suggestion that Mfiller made 

 an error in his observation. He states definitely, however, " der Dorn, der bei Cypridina 

 neben der Basis der Borsten entspringt, fehlt" (1895, p. 158). In Skogberg's material, 

 and in those Discovery specimens which I have examined, stout spines are present at 

 the bases of the natatory setae of the antennal exopodite. 



GENERAL COMPARISON WITH A TYPICAL CYPRIDINID 



The Ostracodan carapace has the form of a bivalve shell, and in all Ostracods except 

 Gigantocypris the two valves of the shell can be opened in the same way as those of 

 a bivalve mollusc. To enable this to take place the two valves are joined together dorsally 

 along a certain length to form a hinge. This hinge is relatively short in those cases, 

 such as Cyprids, where the outline of the valves, as seen from the side, is elliptical. 



