.40 of an inch longer), (2) shorter and stouter bill, (3) much less deeply forked tail, (4) very much paler coloration both above and below (tail above grayish brown instead of black; the back olive gray instead of olive brown; throat and upper breast white instead of deep gray; less olive across the breast and the yellow of abdomen many shades deeper). Series of T. m. couchi and T. melancholicus when compared appear widely different, and specimens of either can be recognized at sight without the slightest difficulty. In fact, so far as the coloration of the ventral surface is concerned, T. m. couchi presents a very close resemblance to T. albogularis. For comparison in respect to size I append measurements of 7 specimens from the Lower Rio Grande Valley in Texas (Lomita Ranch, near Brownsville, Coll. George B. Sennett), as follows: 4 Males: wing, 4.70–4.98, averaging 4.89; tail, 3.68-4.15, averaging 3.91; exposed culmen, .80-.85, averaging .82. 3 Females: wing, 4.50-4.92, averaging 4.72; tail, 3.74-3.84, averaging 3.78; exposed culmen, .80-.86, averaging .83. General average of the 7 specimens (Chapada averages of 18 specimens in parenthesis for comparison); wing, 4.81 (4.38); tail, 3.87 (3.65); bill, .82 (.86). Depth of forking of the tail, .25-35, averaging about .25 (.60). A series of 8 specimens from Yucatan, Guatemala, and Panama (mostly unmarked for sex) are considerably smaller than the Texas series, averaging as follows: wing, 4.48; tail, 3.70; bill, .77. They are thus intermediate in size, as they are also in other characters, between T. melancholicus and T. m. couchi. Specimens from Cayenne, Trinidad, Venezuela, Colombia, and Amazonia are still smaller, and in coloration (except perhaps the Venezuela and Colombia specimens) more nearly approach T. melancholicus. They form, with the Central American birds, a thoroughly connected series, but being on the whole quite unlike either of the extremes, may very well stand as T. m. satrapa, under which latter designation the form was at one time recognized by various authors. 132. Milvulus tyrannus (Linn.).—Represented by 13 specimens, taken as follows: August, 5; September, 5; October 2; December, 1. Probably only a summer visitor. (To be continued.) Article XVIII. CHARACTERS OF PROTOCERAS (MARSH), THE NEW ARTIODACTYL FROM THE LOWER MIOCENE. BY HENRY FAIRFIELD OSBORN and J. L. WORTMAN. Among the many interesting discoveries made by the American Museum Expedition of 1892 were the feet and portions of six skulls of a species of Artiodactyl which appeared to present entirely new characters. The finest specimen has proved to be a perfect skull with complete dentition; associated with another skull are the complete fore and hind feet. In writing from the field-camp Dr. Wortman described the skull as four-horned, but in the Museum, while the specimen was being worked out of its sandy matrix, we found six, eight and finally no less than ten bony protuberances upon different portions of the cranium! The chief pairs are on the parietals and maxillaries; prominent laterally projecting plates are also developed upon the supraorbital ridges of the frontals, and the frontals develop a second conical pair close to the nasal suture above the lachrymals. Besides the great vertical plates, the maxillaries present two lateral protuberances just above the third premolar on either side. The shape of these processes dismisses at once the idea that they were horn cores and indicates that they bore simply a dermal covering. Other features of the skull while less striking and novel are no less unique; among these are the deep cleft between the maxillary plates, the abbreviated nasals, the small vacuity between the nasals and frontals, the prominent ridge extending forward from the anterior margin of the orbit, and the prominent rugose sagittal crest. The grotesque appearance is heightened by the large canines which lend to the lateral aspect of the skull a decided suggestion of resemblance to that of Uintatherium. The edentulous premaxillaries and short-crowned selenodont molars have the true ruminant appearance, but the structure of the feet at first sight suggests the Tragulines. We find two large and two small toes in the fore foot, all of them entirely separate, while the hind foot is supported upon two elongated and closely conjoined digits which form an incipient cannon bone. Fig. 1. Lateral view of the male skull slightly oblique in position, natural size. Before looking for relatives of this remarkable animal, a second skull was uncovered, and it at first appeared to represent an entirely distinct species. The parietals were unfortunately lost. in this specimen, but the frontals are complete and display no traces of either of the protuberances. Further examination, however, indicated the bases of the maxillary plates and of feebly developed canines, which suggested the idea that this skull might represent the female type while the former skull represented the male type of the same species. At this point Marsh's description of Protoceras celer was carefully studied and finally the supposed female skull was taken to the Yale College Museum, and upon being placed side by side with the type of Protoceras it was at once evident that they belong to the same species. This enables us to fully characterize the male and female skulls of Protoceras celer Marsh, and to define the family Protoceratida, which Marsh was unable to do from lack of sufficient material. We may now (1) define the family, (2) give a new definition of the genus based upon the discovery of the sexual characters and differences between the male and female, and upon the characters of the feet, (3) define the species. (4) We will then expand Marsh's description of the female skull; (5) in comparison with the male skull; (6) the structure of the fore and hind feet together with the cranial characters throws some light upon (7) the affinities of Protoceras. Order ARTIODACTYLA. Family PROTOCERATIDÆ Marsh. No Molars brachy-selenodont. Upper and lower canines in both sexes. upper incisors. Lower canines and incisors forming a single series. Male skull with bony protuberances upon parietals and frontals and vertical plates upon frontals and maxillaries. No true horns. Females with small parietal protuberances (maxillary plates unknown). Orbits posterior in position, prominent, widely separated. Optic foramina not confluent. Lachrymal duct with single orifice within rim of orbit. No lachrymal vacuity. Lachrymals articulating with nasals. Nasals extremely abbreviated. Maxillaries with large, free, superior border, produced (in the males) into a broad thickened plate rising above vertex of skull. Fore feet with trapezium, trapezoid and magnum developed and distinct. Four complete, separate and functional metapodials, carpo-metacarpal articulation ‘inadaptive.' Lunar resting equally upon unciform and magnum. Hind feet with two functional metapodials, lateral toes (II and V) incomplete. All elements of the pes separate in the young; tendency to form a cannon bone (III and IV) in adult stage. Ectocuneiform and navicular tending to combine (not with each other) with cuboid. Fibula reduced to a malleolar bone tending to coösify with tibia. Ulna well developed, tending to coösify distally with radius. This family is at present only known to include the genus Protoceras from the upper part of the White River Beds (Lower Miocene) of North America. Genus Protoceras Marsh.2 Dentition: I, C, P, M. First upper and lower premolars simple, bifanged, in diastema midway between canine and second premolar. Third and second upper premolars with strong internal cingula. Fourth upper premolar with single external and internal crescents. Lower incisors and canines with narrow spatulate crowns. Posterior nares open between second molars. Tympanic bulla not inflated. A strong lateral maxillary ridge. Species Protoceras celer Marsh." Male parietal protuberances large, laterally compressed, close together. Type: a female skull in the Yale College Museum. "A Horned Artiodactyle (Protoceras celer) from the Miocene," American Journal of Science, January, 1891, pp. 81, 82. THE SKULL. The Female Skull of Protoceras. History.-Marsh's type specimen is a skull with the posterior portion in fair preservation and the anterior portion broken off in a line just behind the anterior extremity of the nasals passing down just in front of the second premolar. In the following abstract of the author's description we omit some of the details, such as the absence of the first premolar, which are found to be incorrect: "In general form and proportions this skull is of the ruminant type. Its most striking feature is a pair of small horn-cores situated, not on the frontals, but on the parietals immediately behind the frontal suture.... The horn-cores are well separated from each other, and point upward, outward and backward, overhanging somewhat the temporal fossæ. They are conical in form with obtuse summits. . . . The occiput is very narrow, indicating a small cerebellum, and the occipital crest is very weak. The occipital surface slopes backwards.... The facial region of the skull is narrow and elongate. On the outer surface of the maxillary just above the antorbital foramen, there is a deep depression which probably contained a gland. The usual ruminant fossa in front of the orbit appears to be wanting. The orbit is large, and completely closed behind by a strong bar of bone....The paroccipital processes were well developed, but there were apparently no auditory bullæ....As the animal represented by this skull is very distinct from any hitherto described, the genus may be named Protoceras in allusion to the early appearance of horns in this group. The species may be named Protoceras celer. The characters now known suggest affinities with the giraffes, but indicate a distinct family which may be called the Protoceratida." Measurements of Type.-Distance between orbits across frontals, 75 mm., about 3 inches. Distance between summits of horncores, 32 mm., about 14 inches. Width of palate between true molars, 32 mm., about 14 inches. Length of skull, estimated at 200 mm., about 8 inches. The female skull in the American Museum collection is in fair preservation; it has the cerebral hemispheres exposed, and entirely lacks the parietals and the occipital ring; the nasals are complete to the tip; the maxillaries have lost the superior border; the premaxillaries are complete. It is thus impossible to determine whether the maxillaries bore the large vertical plates which constitute so striking a feature of the male skull. Three features indicate that these plates were absent; first, the upper broken |