MAMMALS. 375 



tachment of the ovum to the walls of the uterus, these walls rise 

 up over and enclose the egg, thus coming in contact with it on 

 all sides. From but one side, however, is the (discoidal) pla- 

 centa developed, and in this region the decidua is spoken of as 

 the decidua serotina, while that which covers the smooth or non- 

 villous portion of the chorion (chorion Iseve) forms the decidua 

 reflexa. The rest of the uterine walls, which do not connect 

 with the ovum, are also cast off at birth, and these form the 

 decidua vera. 



In older books the eutheria of the following pages are fre- 

 quently divided into the Implacentalia, including the marsupials, 

 and the Placentalia, including the remaining orders ; but the 

 recent discovery that at least one genus of marsupials (Pera- 

 ■meles) has a true allantoic placenta tends to break down this 

 line. Still the distinction is one of convenience, and has been 

 used in these pages, the term placentalia including all the orders 

 from edentates to primates, the implacentalia, the marsupials, 

 and frequently the monotremes, when these have not been 

 specially mentioned. 



There are two views as to the origin of the mammals ; the 

 one that they have descended from the theromorphous reptiles, 

 the other that they have sprung from the amphibia. The first 

 of these receives its chief support from paleontology. The 

 theromorphs have a heterodont dentition, a triple occipital 

 condyle from which the paired condyles of the mammals can 

 be derived by a suppression of the basioccipital portion, as well 

 as several features in the skeleton of the limbs. The advocates 

 of this view suppose that the quadrate has disappeared in the 

 region of the glenoid fossa. 



The amphibian view receives its support in the double occip- 

 ital condyle, the impossibility of deriving the mammalian ovum 

 from that of any known reptiles, and its easy homology with 

 those of amphibia, and in the relations of the ear bones. This 

 view recognizes the quadrate in the incus, and this articulates 

 with the stapes, a condition repeated in the urodeles, but not 

 derivable from anything known in the reptiles (see p. 1 59). 

 Another difficulty with the reptilian hypothesis is the impos- 

 sibility of deriving the mammalian hair from any exoskeletal 



