56 BULLETIN 110, UNITED STATES NATIONAL MUSEUM. 



In going over a lot of tracings and drawings of Theropodous dinosaur bones 

 made for Professor Marsh, a slip of paper in Marsh's handwriting was found, on 

 which he had written instructions to the draftsman for the composition of the Antro- 

 demus {Allosaurus) fore Umb (fig. 58) which reads as follows: 



Fore limb Allosaurus. 



1. Enlarged scapula (1933 as 83.5 is to 100). 



2. Make coracoid to correspond (See Phillips, p. 208). 



3. Draw humerus (1894 nat. size). 



4. Make radius 9i-incbes long. 



5. Make ulna 9J-inches long + olecranon. 



6. Enlarged foot as 83.5 is to 100. 



The number 1933 is the catalogue number of the Peabody Museum originally 

 given to the type-specimen of Ceratosaurus jmsicornis Marsh before its transfer to 

 the United States National Museum. The coracoid, which is not present with the 

 Ceratosaurus skeleton, was evidently modified from a figure of that bone in Phillips 

 Geology. The humerus is from another specimen, and the radius, ulna, and foot 

 were evidently drawn from those bones pertaining to the type-specimen of Cerato- 

 saurus nnsicornis Marsh, shown in figure 60, and plate 27, figure 3. 



From this brief review it becomes at once evident that this figure is not to be 

 relied upon as being of the genus Antrodemus, and it may therefore be dismissed 

 from further consideration in that connection. 



In the American Museum of Natural History in New York is a beautifully 

 moimted skelton of Antrodemus {Allosaurus), (pi. 16, fig. 1) the fore limbs and feet of 

 which are partially restored. In a letter, bearing the date 1909, Mr. Walter Granger 

 of the American Museum staff wrote me regarding their composition as follows: 



In reconstructing our own fore limb of Allosaurus we had scapula and coracoid, ulna and one claw. 

 The humerus we got from a cast of one in Kansas University; the remainder of the limb and foot was 

 modeled from the Ceratosaurus specimen (Type, No. 4735, U.S.N.M.), which was borrowed from your 

 Museum. 



Following the Ceratosaurus as a pattern the foot was given four digits, whereas 

 the specimen before me shows quite conclusively that Antrodemus had but three 

 digits, with a reduced metacarpal III. The phalangial formula in the New York 

 specimen, is correct, and the relative proportions of the various segments of the 

 limbs are entirely in accord with the associated material forming the basis of the 

 present discussion. In view of the limited fore limb and foot material available at 

 the time of reconstruction this limb and foot of Antrodemus, those in charge of 

 this work are to be congratulated upon their close approximation to the facts as re- 

 vealed by this more recent material. 



Remove the fourth digit, replace metacarpal III by metacarpal IV and insert 

 the carpal bones and the limb would be quite in accord with the fossil specimen 

 before me. This change has now been made. 



Hay ' in 1908 in commenting upon this New York specimen concluded that 

 it had been wrongly identified as Allosaurus, because of the great size of the hand 

 in relation to the other segments of the limb, being led into this error ^ by relying 

 on the authenticity of the figure of the limb and foot published by Marsh. 



1 Proe. U. S. Nat. Mus., vol. 35, 1908, p. 355. ^ Science, vol. 30, No. 759, 1909, p. 93. 



