114 BULLETIN 110, UNITED STATES NATIONAL MUSEUM. 



the skin. It is a relatively small subquadrangular plate of bone, 58 by 70 mm., 

 with a comparative!}' smooth ventral and a roughened dorsal surface. The under 

 surface is gently concave in the direction of its shortest diameter, with a low longi- 

 tudinal swelling extending through the middle of its longest diameter. The rough- 

 ening of the external surface is without definite pattern. 



THE MOUNTED SKELETON OP CERATOSAURUS NASICORNIS MARSH. 



Plates 29-30. 



In plate 30 is a drawing of the mounted skeleton of Ceratosaurus nasicornis 

 which clearly shows the original and restored parts, the latter being indicated in 

 outline only. A comparison of this illustration with the first skeletal restoration 

 (pi. 27, upper figure) would indicate that the skeleton was more perfect at the time 

 Marsh made his reconstruction than it is now, for during the period of 18 years, 

 from 1892 to 1910, the skeleton worked out in relief in blocks of sandstone suffered 

 some breakage and loss of pieces due to packing and sliipping from New Haven to 

 Washington and one or two subsequent moves since its arrival in the United 

 States National Museum; but, on the other hand, many of the bones shaded in full 

 in the first drawing of this animal were never present and have been drawn from 

 the opposite element when a paired bone, or from bones of an allied form like Antro- 

 demus. The small bones of the pes and many of the manus, together with the 

 humerus and coracoid, are examples of the latter, while the loss of parts of the 

 caudal vertebrae, chevrons, and cervical vertebrae illustrate the former reason. 

 I doubt very much that Marsh ever had the distal ends of the pubes and ischia and 

 the portions missing from the ends of the ilia. 



This specimen was collected in 1883 and 1884 by the late M. P. Felch from 

 quarry No. 1, a quarry made classic by the successful researches; thereby Mudge 

 in 1876; Williston and Lakes in 1877; Felch from 1883 to 1888; and Hatcher and 

 Utterback in 1901 and 1902. Quarry No. 1 is situated on the west side of Oil 

 Creek (Four Mile Creek) at the entrance to "Garden Park," and some 9 or 10 miles 

 east by north of Canon City, Fremont County, Colorado. 



Quarry maps made at the time the skeleton was disinterred show the position 

 in which all of the bones of the skeleton were found. It was inclosed in a very 

 hard sandstone, with the skull and axial part of the skeleton very much compressed. 

 This lateral crushing to a great extent determined the bas-relief method of mount- 

 ing as being the best adapted to this particular specimen. Upon assembling the 

 blocks containing the skeleton preparatory to mountmg it was found that some 

 had been cleaned on the right side and others on the left, and a few of the more 

 important elements, as the skull, atlas, axis, and limb bones, had been entirely 

 removed from the matrix. It was finally decided that the right would make the 

 besf exhibition side, and this is well shown in the reproduced photograph of the 

 specimen shown in plate 29. 



The pose of the specimen was largely determined by the position of the bones 

 as originally found. This applies particularly to the left femur, which was held by 

 the rock in its true position with respect to the pelvis. It is directed baclcft'ard at 

 such an angle as to demand a walking stride for bipedal locomotion. An attempt 

 was therefore made to carry out the idea of a rapid walking motion and to make 



