The Sub-class _NEORNITHES_ may be arranged, as above stated, in three divisions, (A) NEORNITHES RATITAE, (B) NEORNITHES ODONTOLCAE, and (C) NEORNITHES CARINATAE. The first of these contains the Ratite Birds proper and possibly part of the so-called _STEREORNITHES_ of Patagonia (p. 43), with several fossil forms of doubtful position from England, France, and New Mexico, as will be seen below; the second the _HESPERORNITHES_ of the Cretaceous Shales of Kansas, the _ENALIORNITHES_ of the Cambridge Upper Greensand, and _Baptornis_ of the American Chalk; the third the _ICHTHYORNITHES_ of the aforesaid Kansas deposits, and all other existing Birds, with various extinct species closely allied to them. Of the points of distinction between the Neornithes and the Archaeornithes the most important are that the metacarpals are fused together, the second digit being the longest, and the third more or less reduced; and that the number of caudal vertebrae does not, as far as is known, exceed thirteen, of which the last five or six combine together to form a pygostyle, except in the Hesperornithes, Ratitae, and Tinamidae, where such is seldom the case.[14] The _centra_ of the vertebrae also are concave on one side only, except in _Ichthyornis_, and perhaps in _Enaliornis_. The possession of teeth is, of course, exceptional, as is the remarkable loss of the keel of the sternum in the Ratitae. It is now generally, if not universally, agreed that Flightless Birds were developed from those that could fly. It does not, however, necessarily follow that the Neornithes are direct {26}descendants of the Archaeornithes, as each may be a separate offshoot from the same parent stem. All we can safely assert is, that the former were in existence about the end of the Jurassic times, that teeth were still retained in some cases during the Cretaceous Epoch, and that not only normal forms, but also flightless forms without keel or pygostyle,[15] had arisen by that date.