MORPHOLOGICAL REVISION OF THE SUBORDER. 89 



The second^ fig. 26, b, from the mid-dorsal region, has a shorter centrum with the 

 bottom line broadly rounded, but not keeled, and concave antero-posteriorly, not 

 straight as in the forms with keels. The centrospheues and centantra are present, 

 but small. The zygapophyses differ from those of the same region in Theropleii7'a 

 retroversa in being more oblique. 



The thirds fig. 26, c, a posterior lumbar, is very short and proportionately broad ; 

 the sides are marked with longitudinal ridges as in Tlieropleitra retroversa. The 

 ribs are detached, leaving separate capitular and tubercular faces. 



The ilium is similar to that of Theropleura retroversa., but larger. 



No. 4109. The two presacrals are of the form figured in Nos. 4130 and 4134, 

 but show a character that is hidden by the matrix of that specimen. Instead of the 

 articular face of the centrum being reverted and passing gradually into the side of the 

 centrum, it is reverted, but separated from the side by a prominent narrow ridge 

 which extends all around the centrum, leaving a wide space for the intercentrum. 



In the two sacrah the first has a wide recurved edge on the anterior face, but 

 the posterior fits against the anterior face of the second sacral very closely, leaving a 

 smaller interspace for a more narrow intercentrum than in Theropleura retroversa. 

 The rib of the first sacral extends rapidly downward rather than out and has a broad 

 rugose distal end. The proximal end is free from the centrum, leaving an articular 

 space extending from the neural arch almost to the mid-line below and resembles 

 very closely the same face of Poliosanriis uniformis (plate i, fig. 6). The anterior 

 edge of the proximal ends does not project beyond the anterior edge of the centrum 

 as in Dinietrodon gigas. 



The anterior caiidals have a similar ridge separating the articular face and the 

 side of the centrum, but it is not so distinct as in the presacrals. The sides of the 

 caudals are concave antero-posteriorly and vertically so there is something of a keel 

 which is rather more sharp than in Theropleura retroz'ersa. The keel of one caudal 

 has a deep, narrow longitudinal groove dividing it into two parts. The caudal ribs 

 have the capitulum and tuberculum as in Theropleura retroversa. 



Genus Elcabrosaurus baldwlnl Case (see p. 28). 



Characteristic specimens : No. 2285 American Museum, the type (plate 4, figs. 1-5). 



The axis is similar to that of Dimetrodon obtiisideiis, with a narrow anterior 

 articular face and well-developed centantra. The bottom line has a sharp ridge 

 running to the posterior edge of the centrum. 



An anterior dorsal has the bottom line narrow and sharp, approaching very 

 nearly to the condition of a true keel, but the sides of the centrum are gradually com- 

 pressed, not sharply pinched in, just below the intercentrum. The lower edge of the 

 anterior face is not bent back nor is there a wide face, showing that the intercentrum 

 was smaller than in Dimetrodon. The transverse process rises from high up on the 

 neural arch. Neither this vertebra nor the more posterior dorsal have the shortening 

 of the bottom line of the dorsals so characteristic of Di»ietrodo)i, but it is not certain 

 that the most anterior dorsals are present. 



The more posterior dorsals have the centra relatively thin and high. The 

 bottom line is thin and sharp but concave antero-posteriorly. An intercentrum of 

 this region is thin antero-posteriorly and has no trace of faces for the heads of the ribs. 



A luynbar is shorter with rounded bottom line. The sides are narrowed by a 

 deep pit on either side just below the transverse process. The transverse process 



