366 PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 



noticed by us. According to Mazatlan fishermen, it readies a larger 

 size than as yet observed by collectors, still retaining its characterisric 

 coloration. 



56. Pomacentrus quadrigutta Gill. 



{Hypsypops dorsalis Gill, Proc. Ac. Nat. Sci. Phila. 1862, 147 (adult) : Pomacenlrns 

 quadrigutta Gill, Proc. Ac. Nat. Sci. Phila. 1852, 149: Pomataprion dorsaHs G\U, Proc. 

 Ac. Nat. Sci. Pbila. 1803, 21G: not Pomacentrus dorsalis Gill. Proc. Acad. Nat. S.-;i. 

 Phila. 1859, 29; a Chinese species.) 



3657. (Type of Pomacentrus quadrigutta.) A very young example, 

 less than one inch in length. 



The type of Rijpsypops dorsalis (43G9) has now gone to decay. 



57. Glyphidodon declivifrons (Gill) J. & G. 



(EuscMstodus declivifrons Gill, Proc. Ac. Nat. Sci. Phila. 18G2, 145, 146; 1833, 219: 

 EvscMstodus concolor Gill, 1. c. 18G2, 145, foot-note = £'».sc/((8^o<7».5 analoyus Gill, 1. c. 

 1883, 219, Aspiuwall.) 



9332. (Tj^pes of Euscliistodus declivifrons.) About ten young exam- 

 ples, 2 to 4 inches in length. 



30744. A large example, 5^ inches in length, in good condition. On 

 this specimen the dark bands have all disappeared. 



58. Glyphidodon saxatilis (L.) Lac. 



(Glyphidodon trosclielii Gill, Proc. Ac. Nat. Sci. Phila. 18G2, 150 ; 18G3, 21:0.) 

 8173, 8180. (Types of Glyphidodon troschelii.) Many young speci- 

 mens. 



■39. Chromis atrilobata Gill. 



(Chromis atriJohata Gill, Proc. Ac. Nat. Sci. Phila. 1862, 149; 1863, 220.) 



3075. (Type of Chromis atrilobata.) A half-grown specimen, in bad 

 condition. 



Ko second specimen of this species has yet been obtained. It may 

 be identical with the Brazilian Chromis marf/inatus, as suggested by 

 Dr. Giinther, but it is certainly premature to unite the two on tLe ba.sis 

 of our present knowledge. A few species of shore-fishes are certainly 

 common to the fauujie of Brazil and Lower California, but the suppo- 

 sition is against identity in any individual case. Much injury has been 

 done to our knowledge of geographical distribution by the random 

 identification of specimens with closely related species belonging to 

 some other fauna. Of 50 species of marine fishes given by Dr. Giinther 

 (Trans. Zool. Soc. London, 1809, 385-392) as common to both sides of the 

 Isthmus of Panama, at least 11 have been incorrectly identified and 

 are not found on both coasts, the identity of 18 more is doubtful and 

 must be verified, while but 21 of the list can be positively stated to be 

 specifically identical. A large number not included in this list are also 

 certainly identical, but in this case it is better to retain some doubtful 

 species than to make many doubtful identifications. 



We may notice that the green coloration of the type of Chromis 

 atrilobata (Proc. Ac. Nat. Sci. Phila. 1863, 220) seems to have come 

 from the copper tank in which it has been kept. 



