224 BULLETIN 34, UNITED STATES NATIONAL MUSEUM. 



is present as a cartilagiuoiis band ; the prefrontal bones are wanting ; 

 the orbitosplienoids are large and exj)anded laterally in front, so as to 

 form part of the palatal surface. The carpus is cartilaginous, and there 

 are no hind legs or jjelvic arch. There are external branchise, which 

 consist of branching processes of the integument of the epibranchial 

 elements. The latter are separated by branchial fissures of the walls 

 of the pharynx. 



In the genus Siren the cranial extremity of the ceratohyal is free from 

 the cranium, but is connected with the stapes by a strong ligament. In 

 this respect this genus resembles the adults of the true salamanders, or 

 Pseudosauria, rather th an the other perennibranchiate forms, or the Tre 

 motodera and Amphiumoidea. In its four epibranchial cartilages, how- 

 ever, it resembles the larvce of the Pseudosauria, as also in the presence 

 of a second basibranchial, connected with the first anteriorly, and ex- 

 panding posteriorly. This mixture of characters of the adults and of 

 the larvae of pseudosaurian urodela has a significance which I will fur- 

 ther illustrate. 



1 have already pointed out (American Naturalist, 1885, p. 245) that 

 palaeontology shows that the order of Trachystomata is a degenerate 

 type, if the structure of its skull, limb-arches, and limbs be considered. 

 I have also reason to believe that there are indications of a retrograde 

 metamorphosis to be found in the history of its branchial apparatus. I 

 was for a long time at a loss to account for the curious condition which 

 I had observed in the branchiae of the sirens. The fringes are frequently 

 in a state of apparent partial atrophy and inclosed in a common dermal 

 investment of the branchial ramus, or all the rami are covered by a 

 common investment, so as to be absolutely functiouless and immovable 

 This character observed in the Fseudobranchus striatus, gave origin to 

 its separation from the genus Siren. The character is, however, common 

 to the Siren lacertina at a certain age, and the real difference between 

 the genera depends on the different number of the digits and pharyn- 

 geal fissures in the two. 



I have also observed that the functiouless condition of the branchiae 

 is universal in young individuals of the Siren lacertina of five and six 

 inches in length, and that in a specimen of a little over three inches 

 they are entirely rudimentary and subepidermal. I have, in fact, no- 

 ticed that it is only in large adult specimens that the branchiae are fully 

 developed in structure and function. The inference from the specimens 

 certainly is that the branchiae are in the sirens not a larval character, 

 as in other perennibranchiate Batrachia, but a character of maturity. 

 Of (jourse only direct observation can show whether sirens have 

 branchiae on exclusion from the egg ; but it is not probable that they 

 differ so much from other members of their class as to be without them. 

 Nevertheless, it is evident that the branchiae soon become functiouless, 

 so that the animal is almost if not exclusively an air breather, and that 

 functional activity is not resumed till a more advanced age. That Sirens 



