2l8 EDWIN CHAPIN STARKS. 



anchylosed to the maxillary. In the Acanthuridse and Siganidse 

 these elements are immovably attached to each other but are 

 held together only by connective tissue. 



The peculiar pelvic girdle of Siganus has its counterpart in 

 the berycoid fishes, but in this connection probably means 

 nothing. The modification of the girdle is brought about by a 

 development upward of the middle portion so that a chamber is 

 inclosed between the opposing sides. 



Summary. 



Siganus stands rather widely away from any known form. 

 The possession of the peculiar prepalatine element and the two 

 spines to each ventral preclude a close relationship to any living 

 fishes. Though its relationship to the acanthuroid fishes is not 

 close it apparently was descended from some form near that 

 stock, and the condition of the maxillary elements, particularly 

 their attachment to the cranium, indicate a relationship in the 

 plectognathous direction rather than in the Chaetodontoid. 



If this be true Siganus can only be an off-shoot at one side 

 from some acanthuroid form having the plectognathous articula- 

 tion of the upper jaw to the cranium. The acanthuroid fishes 

 are in a more direct line with the plectognathous fishes and Sig- 

 anus could not stand between them. 



The plectognathous fishes show degeneration from the acan- 

 thuroid stock by a series of continuous and ever increasing 

 steps. Siganus, on the contrary, shows development in the 

 direction of higher specialization. 



The characters of Siganus are apparently of sufficient value 

 to entitle it to independent superfamily rank at least, or to a rank 

 coordinate with that of the acanthuroid fishes. 



