112 



THE CERATOPSIA 



in fenestratus is further forward in relation to the horn base than in st ember gii but not as far forward 

 as in Trkeratops and Torosaurus. This again may be a mechanical evolutionary trend due in part to 

 a fore and aft thickening of the base of the horn, the Kirtland fenestratus being later in time than 

 sternbergii from the Fruitland beds. Vascular impressions cover almost the entire cranial roof as 

 they do the horns, indicative of the advanced age of the individual. It seems evident, therefore, that 

 the skull of fenestratus does represent a species distinct from the earlier sternbergti, although the 

 specific character which gives it its name, fenestratus, may be considered pathologic and therefore not 

 of taxonomic value. However, the general appearance of the skull, the form of the nasal horn, the 

 relative position of the orbit and supraorbital horns, the development and direction of the epijugal 

 horns, and possibly the apparently greater number of epoccipitals, together with the inclusion in a 

 later formation, are collectively sufficient to confirm the specific distinction which Wiman claims. 



Fig. 40. — Lateral view of skull and skeleton of Pentaceratofs fenestratus, in Upsala, Sweden (modified from Wiman), 

 1/28 natural size, c, left coracoid; /, left femur; /', right femur; / (below) left fibula; ti ', right humerus; i, ischium; 

 il, ilium; p, pubis; r', right ulna; s, left scapula; /, right scapula; sqj, squamosal fenestra (pathologic); /, left tibia; 

 u , right radius. 



Specimen B, the skeleton, has no bones in common with specimen A, the skull, hence the rela- 

 tionship between the two is assumed, although highly probable. The two animals seem to be of 

 approximately the same size, the direct point of comparison being the lower jaw and tooth row of 

 specimen B and the upper jaw of A. Wiman's restoration of the skull in outline as applied to the 

 skeleton gives a most remarkable result, showing the crest extending backward over the entire pre- 

 sacral series of vertebrae, except for the last two or three. In his drawing, however, the crest is 

 flattened down so that its long axis is more nearly in line with that of the cranium than in either 

 of the American Museum specimens of Pentaceratofs, and, I am sure, in fenestratus itself. In our 

 own revision of Wiman's reconstruction (Fig. 40), this flattening of the crest is modified and the 

 head posed somewhat differently, which gives an altered, but not less remarkable, appearance to 

 the animal. One doubts whether the crest could be made to lie flat on the back as in Wiman's 

 figure. In all of these long-crested and doubtless related forms, Chasmosaurus, Pentaceratofs, and 

 Torosaurus, the crest did extend well over the back; in Chasmosaurus it reached to presacral XIII, 

 two-thirds of the back; in Pentaceratofs to presacral XVIII. In Torosaurus, the vertebral column 

 is unknown, but if it was equivalent to that of a Triceratofs of equal body size, such as Triceratofs 

 brevicornus the crest would reach to beyond the 1 6th vertebra, more as in Pentaceratofs. 



The skeleton lay in a somewhat argillaceous sandstone with numerous carbonaceous plant 



