of the Province of Ceara, in (he North of Brazil. 81 



split very readily, and almost all of them present parts of a 

 fish in a more or less perfect state. But hy far the greater 

 number of them are so much broken that it is with consider- 

 able difficulty tolerably perfect specimens can be obtained. 

 The spot which these stones occupy is not above an hundred 

 yards square, and almost no other stone is mixed with them; 

 but on every side of this deposit the ground is covered with 

 little rounded sandstones, similar to the rock of which the 

 Serra is composed. Besides this, I afterwards visited other 

 deposits ; one half a league to the south of it ; one at a place 

 called Maeape, five leagues to the east of Jardim ; and another 

 at Mundo Novo, three leagues to the west ; all perfectly simi- 

 lar to the one I have described, being all situated on the de- 

 clivity of the low hills which stand between the valley and the 

 Serra, and all occupying places which are almost altogether 

 free from other kinds of stone. From these places I have ob- 

 tained a suite of specimens, embracing upwards of a dozen spe- 

 cies of fossil fish.* They vary in size from those of a few inches 

 in length, to others which must have been several feet ; and 

 all of them, so far as my limited knowledge of the subject 

 allows me to judge, except two species, belong to the order 

 Cycloidece of M. Agassiz. The most abundant species is one 

 of those which do not belong to this order. Of it I possess a 

 nearly perfect specimen, about a foot and a half long, but, 

 judging from other fragments of the same species, it must 

 have attained a much larger size.t It has the head very much 

 elongated, and the scales of the back and abdomen are an- 

 gular, while those of the sides consist of but one row of long 

 narrow ones arranged vertically. Of the other species I only 

 possess the tail and a very small part of the body. It differs 

 from the last, in appearing to be entirely covered with small 

 angular scales. Both of them, I have no doubt, belong to tlie 

 order Ganoidece of M. Agassiz. 



though surrounded by others of sandstone, seems to make it probable that 

 the former occur in a bed or layer of detached concretions in the sandstone 

 rock of the Serra. — J. E. Bowman. 



* AgMfiz make* them but seven species, and refers three of them to the 

 Ctenoid group. — J. E. I!. 



t The fish here described is the AspidorhynchusComptoni. Agass. ,) . I ',. I ; 



VOL. XXX. NO. MX. JANUARY 1841. V 



