358 



SCIENCE. 



FN. S. Vol. VIII. No. 194. 



of the pelvic arch occur in the monotremes, 

 and there is a suggestion of their presence 

 in the anomodonts. In the case of the 

 skull the articulation of the lower jaw in 

 some anomodonts approximates to that of 

 the monotremes, while in others they re- 

 semble the marsupials and higher mammals; 

 further the supratemporal and quadrate 

 jugal of Labyrinthodonts may also be repre- 

 sented in Ornithorhynchus, as they certainly 

 are in Pariasaurus. The question is com- 

 plicated by the fact that the anomodonts 

 show resemblances to more than one mam- 

 malian type. For example, the teeth of 

 Diademodont resemble those of the lemurs 

 and of the rodents ; and the Theriodont and 

 and Dicynodont groups of the anomodonts 

 show affinities in the two chief divisions of 

 the mammals. Hence Professor Seeley 

 concludes that, though the points of resem- 

 blance between the mammalian and anomo- 

 dontian skeletons show the affinity of the 

 groups, they do not render it probable that 

 the anomodonts are the direct ancestors of 

 the mammals, but only form a collateral 

 line. For the common ancestor of both we 

 must go back to the Devonian or even to 

 the Silurian periods, and the interval be- 

 tween the mammals and the anomodont 

 reptiles is now so small that there is a 

 reasonable probability that it will be com- 

 pletely bridged by the discovery of further 

 specimens. 



Professor Osborn, of Columbia University, 

 said that certain general principles were 

 useful guides as to the probable nature of 

 the ancestral mammal; in the present im- 

 perfect state of the paleontological record 

 he preferred to commence by working back- 

 ward from the well known comparatively 

 recent forms. In the first place, mammals 

 possess the power of rapid adaptation to 

 their conditions of life. There have been 

 four main centers of adaptive radiation, 

 of which the best case is that of Australia, 

 where the marsupials have acquired forms 



which among placental mammals are di- 

 vided between different orders. The start- 

 ing point of each adaptive radiation has 

 been a small, unspecialized land mammal. 

 Finally, it is probable that the ancestral 

 mammal was omnivorous. Remembering 

 these principles we can trace the line of 

 mammalian descent backward ; it leads us 

 to the Jurassic, when the mammals were 

 all small and belonged to three groups— the 

 primitive insectivores, which have been 

 regarded as marsupials, although there is 

 no evidence to support that view ; second, 

 the multituberculata, which are probably 

 early monotremes ; third, the marsupials. 

 Reversing the order of inquiry. Professor 

 Osborn then referred to the fact that in the 

 Permian there are three groups of rep- 

 tiles, one of which is surprisingly mam- 

 malian in some of its characters, and tempts 

 us to connect the herbivorous section of 

 anomodonts with the monotremes. He 

 thought, however, that the many striking 

 points of resemblance between these reptiles 

 and mammals were due to parallelism, 

 similar characters having been independ- 

 ently acquired. He agreed with Professor 

 Seeley that the anomodonts are not the 

 direct ancestors of mammals, but are a 

 collateral line. He disagreed with Professor 

 Seeley when the latter sought for a much 

 earlier common ancestor of the mammals 

 and the anomodonts, as the speaker believes 

 that an undiscovered and less specialized 

 third subgroup of anomodonts will be 

 found to be the true ancestor of the mam- 

 mals. The Chairman, however, has shown 

 that the mammalian egg is amphibian rather 

 than reptilian in character; and if much 

 weight is to be laid on this point, then the 

 mammals may have descended from some 

 reptile which retained certain amphibian 

 characters. 



Professor Marsh expressed his belief that 

 the solution of this problem is still in the 

 future. He referred to his discussions of 



