188 P. B. LOOMIS— ORIGIN OF SOUTH AMERICAN FAUNAS 



grasses and succulent leaves, and even the deserts find types which can 

 grow under their unfavorable conditions. 



All animals depend on plants for their food supply. In the water it is 

 microscopic types like algse, diatoms, and desmids which are the ultimate 

 source of food for the teeming life of oceans, lakes, and rivers; but on 

 land it is the higher flowering plants which furnish the animal food. 

 While amphibians gained some of the advantages of land life, they were 

 not only unable to lay an egg which could develop in the air, and so have 

 to come back to the water for breeding, but they also were unable to 

 utilize land plants for food, as must be concluded on looking over the 

 range of their adaptations, from their universally carnivorous dentition. 

 These first land vertebrates, then, must have depended ultimately on the 

 microscopic plants of the water ; for, while some might have fed on other 

 amphibians, these had to depend on fishes, which in turn ultimately de- 

 pended on the water plants. In like manner the early reptiles, while they 

 advanced so that they could breed on land, are still all carnivorous; so 

 their food supplies must ultimately have been aquatic, this in spite of 

 the fact that forests of great trees covered the lowlands and swamps. 

 Just why ferns, sigillarians, calamites, gymnosperms, etcetera, did not 

 appeal to the early land animals, I do not know, but it is striking that 

 our early amphibians and reptiles are all carnivorous. In the Cretaceous 

 we find a part of the dinosaurs taking vegetable food of land character. 

 These forms flourished just when the first angiosperms, like birch, beech, 

 maple, plane, sweet-gum, and breadfruit-trees, were coming into domi- 

 nance; and, of course, these dinosaurs became the food supply for the 

 large carnivores. So, then, it is not until the Cretaceous that any of the 

 vertebrates became fully adapted to land life. 



Among these great reptiles, even back to Comanchean time, there are 

 mammals, small, and the first ones essentially carnivorous or perhaps 

 insectivorous. Matthew has suggested that the start of mammalian de- 

 velopment was in the trees,- and has given various anatomical reasons 

 for this belief. It seems to me probable, and at first the food of these 

 tiny forms was probably insects which got their food from land plants. 

 In the trees the early mammals were also where seeds were most accessi- 

 ble, especially when green and soft ; but it was not until herbaceous plants 

 with seeds rich in starch begin to appear in abundance that the mammals 

 increase in number, variety, and size. They are the first class of verte- 

 brates which turned directly to that great unutilized supply of food, and 

 this made suddenly possible a tremendous increase in the vertebrate life 

 on land. 



2W. D. Matthew: Am. Nat., 1904, vol. 38, p. 811. 



