28 OSTEOLOGY OP SCOMBROID FISHES 



has developed into a strong, swift swimmer, the bones of the tail have 

 naturally developed for strength and the caudal rays have clasped the 

 hypural more firmly. 



Considering still other characters, another relationship is suggested. 

 In the Scombridae the first interhEemal is scarcely enlarged and is not 

 closely attached to the haemal process of the first caudal vertebra. The 

 haemal process of the first caudal vertebra is in no way differentiated 

 in shape or size from the last haemal processes (united parapophyses) 

 of the last abdominal vertebra. Consequently when the ribs and inter- 

 haemals are not attached it is impossible to distinguish between the ab- 

 dominal and the caudal vertebrae. In these characters Naucrates and 

 Seriola most resemble the Scombridae. The size and development of the 

 first interha?mal and the last one or two pairs of parapophyses united 

 to form haemal arches, scarcely differentiated from the haemal arches of 

 the caudal vertebrae, are the same, except that in the Scombridae there 

 are always a number of abdominal haemal arches rather than one or 

 two (Scomber, which has the fewest in its family, has five or six). 



These problems are touched upon not only because they are sug- 

 gestive of relationship, but also because they show so well the difficulty 

 of selecting characters that have meaning, and the danger of giving to 

 what may be minor characters undue importance. On the whole, how- 

 ever, Oligoplitcs and Scomberoides seem to approach the Scombridae 

 more closely than do other Carangoid forms. Not only is the general 

 appearance, with a large number of more or less disconnected finlets 

 to the anal and dorsal fins, suggestive of the Scombridae, but the pre- 

 maxillaries are non-protractile, the vertebrae are slightly increased in 

 number and the epiotics are broadly in contact within the cranium.* 



*The character of the union of the epiotics within the cranial cavity has not the 

 value in the Carangidae that it has in the Scombridae. In the Carangidae the 

 character seems to be in a transitional stage, and its stability has been lost in 

 changing from the broadly united epiotics of the Scombridae to the separated ones 

 in the Percoid fishes. In the Percoid fishes (these need further study for this 

 character; ten forms were examined) the epiotics are either separated by a gener- 

 ous area of cartilage below the supraoccipital, which extends downward at least 

 half way between them, or the supraoccipital almost entirely separates them, as in 

 the examples of the Carangoids where the epiotics are separated to the greatest 

 extent. In two closely related Carangoids the condition of the epiotics may show 

 considerable variation, so not much reliance can be placed upon them. In Olig- 

 oplitcs and Scomberoides, however, where the epiotics are in close and broad con- 

 tact and the supraoccipital sends no entering wedge between them, the character no 

 doubt has some value in showing Scombroid affinities. 



