182 BULLETIN 10 0, UNITED STATES NATIONAL MUSEUM 



Back dark brown, also upper surface of head, becoming pale on 

 sides and belly. Iris brown. About 17 to 20 dusky to blackish- 

 brown transverse lines, inclined little forward, variably broken though 

 some few more or less broken as spots. In small examples dark 

 bands become little broader on caudal peduncle and base of caudal, 

 also extend up on dorsals and anals but curving forward. Soft ver- 

 tical fins with deep blue border and submarginal blackish line, and 

 in young against submarginal blackish line rather broad brownish 

 band. Pectorals mostly pale or yellowish, occasionally variable 

 dusky. Ventrals pale, terminally wuth dusky margin. 



East Indies, Melanesia, and Polynesia. Though Giinther reports 

 it from Hawaii through Garrett, likely this may have been the form 

 later described by Jordan and Metz as Holacanthus potteri. We have 

 therefore compared the type of Holacanthus potteri with this series of 

 Philippine examples and find that it appears to differ chiefly in a 

 slightly more barred color-pattern. In Holacanthus hispinosus the 

 bands do not extend across the breast, head and predorsal region. 

 Of this we are, however, not absolutely certain, as the resulting color 

 may in various degrees be due to the action of formaline. The type 

 of Holacanthus potteri is very pale brown with dull cross bands or 

 lines, only the soft vertical fins at all dark or dusky, the caudal alone 

 showing no trace of darker bands. Now in our small examples 

 these dark cross bands are very conspicuous on the caudal of Holacan- 

 thus hisjilnosus. There are also traces of them in some of the large 

 examples. There are also traces in some specimens at least, of dark 

 or light transverse bands on the side of the head, as in Holacanthus 

 potteri, though few if any across the predorsal or breast. Also, in 

 view of this and other examples of Holacanthus potteri that we have 

 examined in all stages from freshly caught specimens to those long 

 in alcohol, we have come to the conclusion that it is a very closely 

 allied species. The armature of the preorbital and preopercle as 

 contended by Jordan and Jordan, are not valid characters. Our 

 series show them simply as variants, evidently due to age or the 

 individuals. We can find nothing to distinguish the nominal Centro- 

 pyge tutuilae. 



21848. Limbones Cove. February 8, 1909. Length 85 mm. [1115]. 



3633. Maribojoc Bay, Bohol Island. March 26, 1909. Length 82 mm. 



3896. Opol, Mindanao. August 4, 1909. Length 91 mm. 



19069. PaUian Bay, Mmdoro. December 11, 1908. Length 97 mm. 



22250. Port Matalvi, Luzon. November 22, 1908. Length 80 mm. 



22261. Tapiantana Island. September 13, 1909. Length 84 mm. 



22160. Tutu Bay, Jolo Island, first anchorage. September 19, 1909. Length 



59 mm. 

 7974. Tutu Bay. September 19, 1909. Length 88 mm. 

 21458, 22825. Danawan and Si Amil Islands. September 27, 1909. Length 62 



to 72 mm. 



