1903. J FROM EAST AFKICA AND ZANZIBAR. 357 



the species in question, or draw an arbitrary dividing line. In 

 the latter case, I think we must say that the typical Staurodoris 

 has simply pinnate branchiae and valve-like tubercles closing the 

 rhinophorial and branchial pockets. One or other of these 

 features may be absent or obscure in a species which otherwise 

 possesses the generic characteristics ; but when both are absent, 

 as in 0., the animal, I think, must be referred to Archidoris, 

 Again, Staurodoris pseudoverrucosa has the dorsal tubercles some- 

 times connected by ridges, and the same phenomenon is found 

 in Garstang's Doris maculata (which appears to be a Staurodoris)^ 

 and, sporadically, in the true St. verrucosa. But, as will be seen 

 from a species described below, Halgerda loasinensis, this character 

 brings Staurodoris very near to Halgerda. 



It will be well to examine the value of the chief points by 

 which the subfamilies and genera under consideration can be 

 differentiated. They are as follows : — 



{a) The dorsal surface and general texture. The back is rarely 

 quite smooth, as it is in Chromodoris ; Halgerda and Asteronotus 

 have the skin smooth but raised into ridges or lumps. As a rule, 

 the surface is covered with projections which may be either 

 minute granulations [Platydoris, Discodoris, &c.), papillse {Thor- 

 disa, &c.), tubercles [Archidoris, &c.), clavate tubercles or warts 

 [Staurodoris), compound tubercles [Trippa, Fi-acassa). There is 

 sometimes a ridge down the centre of the back. Some genera, 

 notably Platydoris, are exceedingly hard ; others, such as Trippa, 

 are so soft as to be almost gelatinous. On the whole, these 

 external characters of the skin and texture form a fairly good 

 indication of relationship. Platydoris, in which I should be 

 disposed to include Hoplodoris, forms a distinct natural group, 

 and the warty or tuberculate forms [Archidoris, Staurodoris) also 

 hang together*. 



(6) Rhinophores and branchife. Neither the rhinophores 

 themselves, nor the pockets into which they are retractile, seem 

 to offer good generic characters, though they may often serve to 

 distinguish species. But even within a species there may be 

 vaiiety : Archidoris tuherculrda has the rims of the rhinophore 

 pocket sometimes smooth and sometimes tuberculate. 



The branchiae also are disappointing as a means of classification. 

 For instance, it does not seem possible to unite Staurodoris with 

 the other genera having simply pinnate branchiie {^Chromodoris, 

 Gasella, Ceratosoma, Sjjhmrodoris, Holla, Thorunna, Rostanga), 

 and these simple branchi^ often show a tendency to divide at 

 the tip and become, strictly speaking, bipinnate. Similarly, we 

 cannot bring together bipinnate forms and oppose them to the 

 tripinnate. Perhaps the distinction between ample and scanty 

 branchiae will prove to be of generic importance. The branchial 

 pocket is of various shapes : round, crenulate, or stellate. However, 



* I find it liai-d to agree with Prof. Bergli's criticism of his own family Archi- 

 doricIida3 that it will prove " ganz unhaltbar und kiinstlich." With the exception 

 of Peltodoris the other forms seem to lismg' well together. 



