110 PROCEEDINGS OF THE ACADEMY OF 



absence of palatine teeth. When the perfect concordance of Leptoscoptis and 

 Dactyloscopus in so many and so most characteristic features is then recalled, 

 can the demonstration of the pertinence of the two genera to the same natural 

 family be rendered more evident ? Yet this peculiar modification of the ven- 

 tral fins is most remarkable in a member of this family ; it is certainly one 

 that the naturalist would not a priori expect to be found. Dr. Gunther, 

 doubtless influenced by such considerations, has not noticed the Dactyloscopi 

 in his catalogue of Acanthopterygian Fishes, as a member of his group of 

 Uranoscopina. Having never seen it, he probably, notwithstanding the com- 

 parisons and observations recorded in the original description, considered it 

 to be a Blennoid ; the only character that it possesses in common with that 

 family is the structure of the ventral fins. We again repeat, that not only 

 the preponderance but the totality of its characters, with that sole exception, 

 decides that its legitimate affinities are to the Leptoscopi. But to vindicate 

 our appreciation of the importance of this character in the present family, we 

 have proposed to institute for the genus a distinct subfamily. 



Nor is the peculiar modification of the ventral fins the only character which 

 is generally indicative of family rank, but here of much less value. The 

 Uranoscopina? have a coecal stomach and a moderate number of pyloric coeca, 

 the number in the species examined ranging from eight to twelve, but in 

 Leptoscopus the coeca are entirely absent. This is undoubtedly a character- 

 istic of as much importance as the peculiarity of the ventral fins. Yet Dr. 

 Gunther has with propriety retained the species as characterized among his 

 Uranoscopina, a group which is equivalent to the family of Uranoscopoids as 

 here admitted after the substraction of the Dactyloscopi. 



It will be observed that I have always compared the Dactyloscopi to the 

 Leptoscopi, and asserted that both belong to the same family. If ever a divi- 

 sion of the Uranoscopoids should be made, or if any forms now referred to it 

 are abstracted, they would properly be the Leptoscopina? and Dactyloscopina? 

 together. Those groups resemble each other in their elongated body covered 

 with moderate scales, their median lateral line, the long dorsal and anal fins, 

 and the smooth head. In all of these respects they differ from the Uranosco- 

 pina?. The latter have also pancreatic coeca, while doubtless all the former 

 have none. The Leptoscopinae and Dactyloscopinse will, therefore, be probably 

 referred by some future naturalist to a distinct family, but I am not myself 

 prepared at present to adopt such, and entertain some doubt whether such a 

 separation would be ever justifiable. In my former remarks on Dactyloseopus 

 I have observed that " had either the peculiarity of dentition or of the ventral 

 fins singly distinguished the Dactyloscopi from the Leptoscopi, both might 

 possibly have been naturally placed in the same tribe or subfamily.'' I have 

 further remarked, that "notwithstanding the abnormal and blennoid struc- 

 ture of the ventrals, and the absence of the vomerine and palatine teeth, the 

 Dactyloseopince appear to be almost as much related to the Leptoscopinoz as 

 the latter are to the Uranoscopina, properly so called." Since the not unex- 

 pected discovery of Crapatalus, I will now express my belief in the much greater 

 affinity between the Dactyloscopins and Leptoscopina? than that of the latter 

 to the Uranoscopina?. The subfamilies are now indeed distinguished by almost 

 only one character, but I still retain them. 



From the preceding observations, it is apparent that there are few groups of 

 such intrinsic interest, as well as of such importance for the proper informa- 

 tion of the value of certain characters. We are taught not to place too great 

 reliance on any one character, as such might cause us to violate the principle 

 of natural classification. The close identity of general form, so characteristic 

 of the entire family, and the combination of characters so numerous and so 

 peculiar, forbid the naturalist from the consideration of any one of the groups 

 referred to as the representative of different or distinct families, notwithstand- 

 ing the absence of pyloric coeca in one, and the additional modification of the 



[May, 



