1 62 NATURAL HISTORY. 



general law remains unaffected. Thus, so far as our 

 present knowledge goes, milk-glands are never found 

 except in association with a double joint between the 

 skull and the backbone, and a simple structure of the 

 lower jaw. But it is quite possible that we might some 

 day find mammary glands in association with a differently 

 constructed jaw, or with a single occipital articulation, 

 however unlikely this may seem at present. Even living 

 animals are constantly demonstrating to us the danger of 

 these empirical generalisations. Thus, until recently, it 

 might have been safely asserted that all Vertebrate animals 

 with a bony skeleton, which produced their young as eggs, 

 had also a lower jaw in which each half was composed of 

 more than one piece. We know now, however, by recent 

 investigations that the duck-mole and spiny ant-eater, both 

 of which have the lower jaw simple, are oviparous animals. 

 Extinct animals still more forcibly exemplify the necessity 

 for caution in reasoning from the presence of one structure 

 that another correlated structure was present Thus, until 

 a few years ago, it would have been unhesitatingly admitted 

 that the possession of a covering of feathers was correlated 

 with the possession of a horny covering to the margins of 

 the jaws, and therefore with the absence of teeth. It 

 would also have been admitted that feathers were corre- 

 lated with saddle-shaped faces to the bodies of the verte- 

 brae of the neck. Through the researches of Professor 

 Marsh, we are, however, now acquainted with extinct 

 birds which must, as birds, have possessed feathers, but 

 in which the jaws were furnished with teeth in sockets 

 (the Odontornithes'). The same distinguished observer has 

 also brought to light an extinct bird {Ichthyornis\ in 



