231 



type hy the addition of extra phalanges in the three middle toes and 

 such an addition can hardly have come about except by a sudden 

 jump from 2 to 3, 3 to 4, and 4 to 5. 



In the present paper I wish to discuss rather how the formula, 

 2. 3, 4, 5, 3, become reduced to 2, 3, 3, 3, 3, when the feet instead of 

 being placed by the sides of the body came to be placed underneath. 

 Such a reduction has taken place with the shorteiiing of the toes in the 

 Chelonia and an exactly similar reduction has taken place in the an- 

 cestors of the Mammals, and fortimately though we know little of the 

 evolution of the Chelonia we can trace the steps of mammalian descent 

 with confidence. 



In the early mammal -like reptiles such as the Pelycosaurs we have 

 the digits almost typically reptilian with the formula 2, 3, 4, 5, 3. 

 In the South African rather later Carnivorous groups, the Thero- 

 cephalia and the Gorgonopsia. the for- 

 mula is still 2, 3, 4, 5, 3 but the 2nd 

 phalanx of the 3rd toe and the 2nd and 

 3rd phalanges of the 4th toe are much 

 reduced and the toes thus become 

 subequal in length. The figure given 

 of the manus of the Gorgonopsian 

 Scymnognathus tigriceps shows the 

 condition in the early mammalian an- 

 cestors. In some TherocephaHans e. q. 

 Aelurosaurus the reduced phalanges are 

 even more plate-hke. In the Droma- 

 sauria, the Anomodontia and the Cyno- 

 dontia we find the reduced phalanges 

 lost and the feet with the typical mam- 

 malian formula 2, 3, 3, 3, 3. 



As the increase in number of the 

 phalanges most probably took place by 

 a sudden jump it is also more pro- 

 bable that the reduced phalanges sud- 

 denly disappeared than that they became more and more reduced. 

 No trace of the missing phalanges has ever been detected in mam- 

 mahan embryos. 



In the heav}^ bodied Pareiasaurus we find the digital formula 

 reduced from the typical Cotylosaurian 2, 3, 4, 5, 3 to 2, 3, 3, 4, 3. 



Left manus of Scymnognathus ti- 

 griceps Broom and Haüghton, re- 

 duced. The carpus is drawn as 

 found, r.s. is possibly a part of a 

 rudimentary prepollex or radial 

 sesamoid. 



