850 AMPHIBIA. 



The 1st yertebra has the usual two lateral facets, Plate LXXYIII, fig. 1 a a, for 

 articulation with the occipital condyles with an odontoid-like process (b), similar to 

 that figured by Mivart^ in Amphiuma, fitting into an articular surface between the 

 occipital condyles. In one specimen, the parasphenoid sent up a small nodule 

 between the posterior margins of the exoccipitals and entered into the middle of the 

 anterior margin of the cavity. In this instance, fig. 2, the intercondyloid process of 

 the atlas was divided into two rounded articular facets b, which corresponded to two 

 median articular surfaces, ^g. 25 s, internal to the facets of the condyles. I have 

 observed this arrangement in two cases. In specimens in which the odontoid-like 

 process is not very long, either one or both of the lateral facets of the 1st vertebra 

 are confluent with it, but, when it is of moderate size, each surface is sharply 

 defined. The spinous process of the 1st vertebra, fig. 3 c, has its extremity dilated 

 and covered with a broad rough porous plate (h) or crest corresponding to the 

 porous glandular layer of the back, and a sharp anterior edge (d) and a deeply 

 concave posterior surface (e) for the reception of the spinous process of the 2nd 

 vertebra. In the succeeding vertebrae, figs. 6 to 15, the spinous process becomes 

 much laterally compressed, and the lateral margins of the posterior concavity, 

 figs. 8^ and 12^, become distinct plates, from which the posterior zygapophyses (/) 

 depend. 



The most peculiar feature of the vertebrae is the existence on the summit of the 

 spinous process of the dorsal, sacral, and first two or three caudal vertebra of the 

 rough osseous expansion or crest, already referred to, and conforming to the outline 

 of each process, and which appears to be peculiar to this newt. It is pointed ante- 

 riorly and bifurcate posteriorly for the reception of the anterior extremity of the 

 process immediately behind it. The anterior and posterior zygapophysial facets 

 have the general characters of these structures in the Urodela. There are no hypa- 

 pophysial ridges in the trunk vertebrae. Transverse processes are not indicated in 

 the 1st vertebra (fig. 6), but a process of this nature bearing a rib {m and j9), 

 figs. 7 to 14, is developed in each of the vertebrae from the 2nd to the 2nd caudal. 

 In the rest of the caudal vertebrae, the transverse processes, figs. 15 and 16^, also 

 assume the general character distinctive of the Urodela. In the dorsal region the 

 two transverse processes, superior and inferior, are placed side by side in the usual 

 manner and united by their edges, each terminating in an articular surface for 

 the head and tubercle of the rib. The base of each process is deeply concave 

 posteriorly, and is perforated by a vascular canal, as in many other genera of 

 Urodela.^ In the caudal region, figs. 15 to 18, the transverse processes are short, 

 their vertical plates having a somewhat wavy margin connected to the anterior and 

 posterior articular processes by a long longitudinal lamella and by a ridge to the 

 posterior inferior margin of the hypaxial arch. There is so great a similarity at 

 first sight between the dorsal and ventral arches, that when a single caudal 

 vertebra is taken up, one hesitates at first to decide between the arches. In the 



1 Proc. Zool. Soc, 1870, fig. 19, p. 277. 



2 Mivart, L c, p. 271. 



