No. 535] THE SAUROPODOUS DINOSAURS 401 



ceeded in size by some of the rhinoceroses, which again 

 have very angulated legs. It is to be noted, too, that most 

 of the genera mentioned by Dr. Matthew as being gigantic 

 mammals of diverse stocks are extinct, and there might 

 be some differences of opinion as to the exact form of 

 their legs and even of the feet. 



Nor is it certain that the type of limb found in the ele- 

 phant is the result of the great size of the animal. The 

 limb has probably not changed greatly since Middle 

 Eocene times, when elephants were small animals; al- 

 though its angulation has probably become somewhat 

 reduced, the femur relatively somewhat longer, and the 

 digits a little shorter. The femur of Mceritherium, sup- 

 posed to be a relative or even an ancestor of the ele- 

 phants, from the Middle Eocene of Egypt, is only eleven 

 inches long and is as straight as that of the living ele- 

 phant. The type of leg possessed by the elephant is 

 rather primitive in its structure and is found in such 

 animals as are slow of movement. All animals have 

 some means of protecting themselves more or less effec- 

 tively from their enemies. Horses, deer and oxen have 

 developed angulated limits, with short humeri and femora 

 and elongated metapodials, combinations that confer the 

 power of rapid flight. The elephants have found their 

 safety in their great size; and extensive modifications of 

 their legs have not been required. 



As is well known, the lizards have straight-shafted 

 femora, which are longer than the tibiae, and the digits 

 are long. As, however, the land tortoises may have short 

 digit igrade feet, it is difficult to see why some lizards 

 might not develop such feet. And it is not improbable 

 that such a lizard might attain a great size and continue 

 to retain its straight femora, short feet, and angulated 

 limbs. 



Tn short, it seems to the writer that Dr. Matthew's 

 proposition as regards the significance of the straight 

 femur is loaded with so many conditions and subject to 

 so many exceptions, known and supposable, that it must 



