This genus includes the species _P. longissimus_, _P. inflatus_, _P. platygnathus_, _P. modicus_, _P. gracilis_, and _P. sehuensis_; _Brontornis_, which has a shorter and wider mandible and smaller but stouter metatarsi, possesses in _B. burmeisteri_ a form as large as _Aepyornis maximus_, while _Opisthodactylus_ and other proposed genera are too imperfectly known to deserve consideration in our limited space. Besides the above, Dr. Gadow classed with the Stereornithes, _Diatryma_ of New Mexico, known from a metatarsus; _Dasornis_ of the London Clay, described from fragments of a skull; _Remiornis_ from the neighbourhood of Rheims, of which several imperfect bones have been found; and _Gastornis_ of both England and France, of which a fair number of parts have been unearthed. All occur in the Eocene, but the question of their relationship is by no means settled, and some writers consider _Gastornis_ to be nearly allied to the Anseres. This form appears to have been of the size of an Ostrich, with long leg-bones and short weak wings, and was probably flightless. Three species have been propounded, _G. parisiensis_, _G. klaasseni_, and _G. edwardsi_. * * * * * (B) With regard to the difficult question of the position in the system of the NEORNITHES ODONTOLCAE, a few introductory words of explanation are necessary. In 1872 Professor Marsh bestowed upon two fossils from the Cretaceous deposits of Kansas the names of _Hesperornis_ and _Ichthyornis_, which he proposed in the following year[45] to comprise in a Sub-class _Odontornithes_, so called from the presence of teeth in the jaws. Subsequently[46] he divided this Sub-class into two Orders, _Odontolcae_ and _Odontotormae_, the former containing _Hesperornis_, with the teeth arranged in grooves, and the latter _Ichthyornis_, where they were placed in distinct sockets. His views have been controverted by many writers, but Mr. Lydekker–an authority of great weight in this connexion–while fully admitting the affinity of the first form to {46}the Divers, and the resemblance of the second to the Gull-tribe, proposed in 1891[47] to retain the term Odontornithes for a series of birds ancestral to the modern series of toothless Carinatae, for which he adopted the title _Euornithes_, used in a narrower sense by Dr. Stejneger. It has, however, been decided to follow Dr. Gadow on this point; while the marks of distinction given below make it seem at least probable that, whereas _Ichthyornis_ may be referred to the Carinate division, _Hesperornis_ should be placed in closer proximity to the Ratite forms. Our Neornithes Odontolcae consequently contain the HESPERORNITHES, the ENALIORNITHES, and _Baptornis_, all of which appear to be nearly related. [Illustration: FIG. 13.–Restoration of _Hesperornis_. (From Huxley, after Marsh.) × 1/13.]