1910.] CAUDAL KTX OF THE TELEOSTOMI. (J 1 9 



lust verfcel)ral segment ; thus there is an julviineed foiiture (the 

 former) and a lower feature (the hvtter) both present in the same 

 caudal — in other words, an intermediate condition in specialization 

 is present, which is consistent with the intermediate p(Jsition the 

 division occupies among the Acaiithopterygii. 



It would, therefore, seem that each individual division has its 

 own grades of specialization within itself, and in this minor 

 cliissitication the structure of the tail does coirespond with the 

 general structure of the fish in i-espect to specialized featvu-es in 

 general. In the Perciformes, the Herranidje are among the lenst 

 specialized of the several families, and the Labrid;e are at the 

 other end of the sc;de ; and it has been pouited out that memljers 

 of the former family have less specialized tails, by far, than those 

 of the latter ; thus caudal features correspond with specialized 

 features in general, within this division. Exactly the same 

 thing occurs among the Scleroparei, where the least specialized 

 members ai'e the Scorpajnidas and one of the most specialized 

 the Triglidfe ; the Scoi'pa^nidaj have lowly and the Ti-iglida^ 

 highly specialized caudal tins. 



The Scleroparei are a moi-e specialized di\ision than the Perci- 

 formes, and yet Scorpcena^ among the former, has a less specialized 

 tail than that of JMhriis in the latter division. Thus while the 

 caudal tin may not be dejjended upon as a feature for classification 

 in a broad sense, it may be useful within the scope of a single 

 division. 



P I s T II o M I. 



Mastacembelus sp. ? (Plate L. fig. 31.) 



This fish possesses a continuous undifferentiated median fin ; 

 the urostyle is well-marked and free, and equal in length to a 

 vertebra. The limits of the caudal fin proper are probably best 

 fixed here by those dermotrichia supported by hypural bones, of 

 which two are associated with the last vertebral segment, the 

 anterior one being obviously composed of three fused collaterally, 

 and one with the penultimate vertebra. All the remaining fin- 

 rays of the median fin are borne by radials ; no epurals ai'e 

 present. A trace remains of the neural arch to the last vertebra. 

 Thus the tail is definitely lion)ocercal ; in regard to this, it is 

 strange that authoi's refer to it an a gephyrocercal tail. The only 

 possible foundation for so naming this type is the apparent 

 symmetry due to the confluent tins ; it mu.st, however, be insisted 

 on that gephyrocercy indicates true secondaiy synnnetry, i. e. an 

 al)breviated protocercal condition. It is obvious that such a tail 

 as that of Mastacembelus cannot be referred to the same category 

 as that of Fierasfer. 



Pediculati. 

 Loniius PiscATORius. (Plate L. fig. 32.) 



This fish may be said to illustrate the extremost type of 

 advanced homoceivv. Practically caily one hypural is present. 



