bearing on our knowledge of ornithology will be most conveniently treated here.
In November 1870 O. C. Marsh, by finding the imperfect fossilized tibia of a bird in the middle cretaceous shale of Kansas, began a series of wonderful discoveries of great importance to ornithology. Subsequent visits to the same part of North America, often performed under circumstances of discomfort and occasionally of danger, brought to Marsh. this intrepid and energetic explorer the reward he had so fully earned. Brief notices of his spoils appeared from time to time in various volumes of the American Journal of Science and Arts (Silliman’s), but it is unnecessary here to refer to more than a few of them. In that Journal for May 1872 (ser. 3, iii. p. 360) the remains of a large swimming bird (nearly 6 ft. in length, as afterwards appeared) having some affinity, it was thought, to the Colymbidae were described under the name of Hesperornis regalis, and a few months later (iv. p. 344) a second fossil bird from the same locality was indicated as Ichthyornis dispar—from the fish-like, biconcave form of its vertebrae. Further examination of the enormous collections gathered by the author, and preserved in the Museum of Yale University at New Haven, Connecticut, showed him that this last bird, and another to which he gave the name of Apatornis, had possessed well-developed teeth implanted in sockets in both jaws, and induced him to establish (v. pp. 161, 162) for their reception a “subclass” Odontornithes and an order Ichthyornithes. Two years more and the originally found Hesperornis was discovered also to have teeth, but these were inserted in a groove. It was accordingly regarded as the type of a distinct order Odontolcae (x. pp. 403-408), to which were assigned as other characters vertebrae of a saddle-shape and not biconcave, a keelless sternum, and wings consisting only of the humerus. In 1880 Marsh brought out Odontornithes, a monograph of the extinct toothed birds of North America. Herein remains, attributed to no fewer than a score of species, which were referred to eight different genera, are fully described and sufficiently illustrated, and, instead of the ordinal name Ichthyornithes previously used, that of Odontotormae was proposed. In the author’s concluding summary he remarks on the fact that, while the Odontolcae, as exhibited in Hesperornis, had teeth inserted in a continuous groove—a low and generalized character as shown by reptiles, they had, however, the strongly differentiated saddle-shaped vertebrae such as all modern birds possess. On the other hand the Odontotormae, as exemplified in Ichthyornis, having the primitive biconcave vertebrae, yet possessed the highly specialized feature of teeth in distinct sockets. Hesperornis too, with its keelless sternum, had aborted wings but strong legs and feet adapted for swimming, while Ichthyornis had a keeled sternum and powerful wings, but diminutive legs and feet. These and other characters separate the two forms so widely as quite to justify the establishment of as many orders for their reception. Marsh states that he had fully satisfied himself that Archaeopteryx belonged to the Odontornithes, which he thought it advisable for the present to regard as a subclass, separated into three orders—Odontolcae, Odontotormae and Saururae—all well marked, but evidently not of equal rank, the last being clearly much more widely distinguished from the first two than they are from one another. But that these three oldest-known forms of birds should differ so greatly from each other unmistakably points to a great antiquity for the class.
The former efforts at classification made by Sundevall have already several times been mentioned, and a return to their consideration was promised. In 1872 and 1873 he brought out at Stockholm a Methodi naturalis avium disponendarum tentamen, two portions of which (those relating to the diurnal birds-of-prey and the Cichlomorphae, or forms related Sundevall. to the thrushes) he found himself under the necessity of revising and modifying in the course of 1874, in as many communications to the Swedish Academy of Sciences (K. V.-Ak. Förhandlingar, 1874, No. 2, pp. 21-30; No. 3, pp. 27-30). This Tentamen, containing his complete method of classifying birds in general, naturally received much attention, the more so perhaps, since, with its appendices, it was nearly the last labour of its respected author, whose industrious life came to an end in the course of the following year. From what has before been said of his works it may be gathered that, while professedly basing his systematic arrangement of the groups of birds on their external features, he had hitherto striven to make his schemes harmonize if possible with the dictates of internal structure as evinced by the science of anatomy, though he uniformly and persistently protested against the inside being better than the outside. In thus acting he proved himself a true follower of his great countryman Linnaeus; but, without disparagement of his efforts in this respect, it must be said that when internal and external characters appeared to be in conflict he gave, perhaps with unconscious bias, a preference to the latter, for he belonged to a school of zoologists whose natural instinct was to believe that such a conflict always existed. Hence his efforts, praiseworthy as they were from several points of view, and particularly so in regard to some details, failed to satisfy the philosophic taxonomer when generalizations and deeper principles were concerned, and in his practice in respect of certain technicalities of classification he was, in the eyes of the orthodox, a transgressor. Thus instead of contenting himself with terms that had met with pretty general approval, such as class, subclass, order, suborder, family, subfamily, and so on, he introduced into his final scheme other designations, “agmen,” “cohors,” “phalanx,” and the like, which to the ordinary student of ornithology convey an indefinite meaning, if any meaning at all. He also carried to a very extreme limit his views of nomenclature, which were certainly not in accordance with those held by most zoologists, though this is a matter so trifling as to need no details in illustration. His Tentamen was translated into English by F. Nicholson in 1889, and had a considerable influence on later writers, especially in the arrangement of the smaller groups. In the main it was an artificial system. Birds were divided into Gymnopaedes and Dasypaedes, according as the young were hatched naked or clothed. The Gymnopaedes are divided into two “orders”—Oscines and Volucres—the former intended to be identical with the group of the same name established by older authors, and, in accordance with the observations of Keyserling and Blasius already mentioned, divided into two “series”—Laminiplantares, having the hinder part of the “tarsus” covered with two horny plates, and Scutelliplantares, in which the same part is scutellated. These Laminiplantares are composed of six cohorts as follows:—
Cohors 1. Cichlomorphae.
Cohors 2. Conirostres.
Cohors 3. Coliomorphae.
Cohors 4. Certhiomorphae.—3 families: tree-creepers, nut-hatches.
Cohors 5. Cinnyrimorphae.—5 families: sun-birds, honey-suckers.
Cohors 6. Chelidonomorphae.—1 family: swallows.
The Scutelliplantares include a much smaller number of forms, and, with the exception of the first “cohort” and a few groups of the fourth and fifth, all are peculiar to America.
Cohors 1. Holaspideae.
Cohors 2. Endaspideae.
Cohors 3. Exaspideae.
Cohors 4. Pycnaspideae.
Cohors 5. Taxaspideae.
We then arrive at the second order Volucres, which is divided into two “series.” Of these the first is made to contain, under the name Zygodactyli,
Cohors 1. Psittaci.
Cohors 2. Pici.
Cohors 3. Coccyges.
Cohors 4. Coenomorphae.
Cohors 5. Ampligulares.
Cohors 6. Longilingues or Mellisugae.
Cohors 7. Syndactylae.
Cohors 8. Peristeroideae.
The Dasypaedes of Sundevall are separated into six “orders”; but these will occupy us but a short while. The first of them, Accipitres, comprehending all the birds-of-prey, were separated into 4 “cohorts” in his original work, but these were reduced in his appendix to two—Nyctharpages or owls with 4 families divided into 2 series, and Hemeroharpages containing all the rest, and comprising 10 families (the last of which is the seriema, Dicholophus) divided into 2 groups as Rapaces and Saprophagi—the latter including the vultures. Next stands the order Gallinae with 4 “cohorts”;
