GENUS CHKLONOBIA. 387 



(PL 14, fig. 1 a) of the shell of C. testudinaria, with the 

 transverse section (PL 15, fig. 1) of the same species: in 

 this latter figure, a is the outer lamina, and bab the three 

 compartments of the rostrum. But when the outer lamina 

 is worn away, as is always the case with the upper part of 

 the walls in C. caretta (PL 14, fig. 2), the two fissures 

 separating the three compartments of the compounded 

 rostrum, are plainly exhibited on the outside of the upper 

 part of the shell. On the internal surface, the sutures 

 separating the three compartments are always open, except 

 in the upper part of the sheath, above the attachment of the 

 opercular membrane, where they are partially obliterated 

 by a thin continuous layer of shell. That these three por- 

 tions of the rostrum are in their essential nature compart- 

 ments, is well shown in C. patiila and testudinaria, by the 

 sheath or inner lamina having loopholes or channels (such 

 as before described) in the middle of each, and on each line 

 of suture. From the number of these channels, seven 

 altogether, (the two between the compound rostrum and 

 lateral compartments being counted,) the sheath of the com- 

 pound rostrum is reduced to mere flattened pillars between 

 the several channels. By slight violence, the rostrum breaks 

 into the three portions ; and the sutures between them are 

 found to be marked on both sides by sinuous, slight, 

 calcareous riclges, those on opposite sides locking into each \ 

 these represent the septa on the edges of the radii and alae 

 and their recipient furrows, in the ordinary compartments. 

 The outline of the middle compartment of the three (best 

 seen in section PL 15, fig. 1, a), much resembles that of 

 the carina ; in fact, if we suppose the growth of the two alas 

 of the carina to have been arrested, no essential difference can 

 be pointed out : in this rudimentary compartment, therefore, 

 we have a rostrum, characterised as in the sub-family of the 

 Chthamalinse. In the two little rostro-lateral compartments, 

 moreover, (b b), a slight swelling on the side opposite to 

 the large existing radius, shows that if the development of 

 these compartments had not been prevented, each, probably, 

 would have had, exactly as in the Chthamalinae, a radius on 

 both sides. In the introduction to the Balanidae, I have 

 argued, from the structure here described and from some 



