136 PROCEEDINGS OF THE NATIONAL MUSEUM. vol, 51. 



The authors have had the privilege of further examining a hirge 

 number of species, including nearly all those obtained by the Alba- 

 tross in extensive expeditions in the North Atlantic Ocean, on both 

 coasts of South America, about the Galapagos Archipelago, in Pan- 

 ama Bay, off the west coast of North America, in Bering Sea, about 

 the Hawaiian Islands, and about the Philippine Islands.^ The 

 senior author critically studied the material in the United States 

 National Museum in 1905. On the basis of the study of these collec- 

 tions a revision of the subfamilies and genera is proposed. 



The characters heretofore used to define the genera or subgenera in 

 the large subfamily Coryphaenoidinae (=Macrourinae) have largely 

 proved to be of minor or only specific value, because of the great 

 variation shown by obviously related species as regards these char- 

 acters. Among these may be mentioned the dentition, which has 

 been used heretofore in the primary division of the subfamily. The 

 dentition of the lower jaw has proved to be of no generic value, as in 

 closely related species of Conjphaeno/'des and Z/(?7iMr?/s the teeth vary 

 from a single series to a wide band. The enlargement of the outer 

 premaxillary series, forming the heterodont dentition of " Chali- 

 nura,^'' is not a reliable generic character because of its great varia- 

 tion in different species of Coryphaenoides. But in each of three 

 distinct groups there are found two genera or subgenera, one of which 

 contains only one or two series of teeth in both jaws while the other 

 has a distinct premaxillary band. Nematonurus is thus separated 

 from Coryi^haenoides^ Ahyssicola from Coelorhynchus, and Mala- 

 cocephalus from Lionurus. Other characters are correlated with 

 this difference in dentition. The amount of rostral projection is 

 usually characteristic of the genera, but no value can apparently be 

 placed on this character in the large genera Coryphacnoides and 

 Lionurus (as here defined). The roughness of the scales is subject 

 to wide variation in the larger genera and apparently distinguishes 

 natural groups in but a single case. In the genus Lionurus the sub- 

 genus Lionurus differs from the subgenus Nezumia in the smooth 

 scales of the body. The amount of interdorsal space can only be 

 regarded as of specific value. The serration of the dorsal spine is a 

 valuable character. The serrations are found only in the Cory- 

 phaenoidinae (3=Macrourinae), and are absent in six genera of that 

 subfamily: Cynomacrurus^ Coelorhynchus., Ahyssicola^ Hymenoceph- 

 alus., Malacocephalus^ and Trachonurus. The serrations are obso- 

 lescent in certain species of the large genera C ory phaenoides and 

 Lionurus. But the strength of the spine and its trenchant anterior 

 edge are retained even in these species. The spine is comparatively 



iThe entire Philippine collection of Macrouroid fishes made by the Fisheries steamer 

 Albatross during the years 1907 to 1910 has been kindly sent to us by Dr. Hugh M. Smith, 

 Commissioner of Fisheries. 



