128 LORICARIXiE. 



that we consider the list by any means complete, 

 for in such an extent of waters we have no doubt 

 more than double the number of species exist, 

 yet it may be presumed that specimens of the pre- 

 vailing kinds were obtained. The drawings from 

 which we have wTOught were made upon the spot 

 from the newly killed fishes, and are on this ac- 

 count most important, as giving the fresh tints of 

 colour which it is impossible otherwise to carry 

 away. They have been quickly executed, but pos- 

 sess evident marks of being very characteristic and 

 correct, wherever specimens for comparison have 

 accompanied them, — at the same time, minuteness 

 of execution and of the details of scaling, &c., as 

 might have from circumstances been anticipated, 

 are w^anting; and here has lain our greatest diffi- 

 culty, from these circumstances it has been im- 

 possible always to fix to our satisfaction the proper 

 station of the fish, but in all cases where this occurs, 

 the uncertainty is stated, and the species placed as 

 near as can be ascertained to where its true place 

 should be. In arrangement we have followed no 

 regular system, but have preferred keeping the fami- 

 lies separated, and at the commencement have placed 

 the Siluridae, as being characteristic of these rivers. 

 The sub-family, 



LORICARIN^, 



or mailed cat-fish, may be first noticed ; several of 

 the genera w^hich -will range here are preserved 

 in the Schomburgk Collection. These are most 



