which may be called Coryphodon hamatus. It indicates an animal somewhat larger than a tapir. The more important characters derived from an examination of this cranium, and some other remains of the same genus, are as follows:

The skull is elongated, the facial portion being most produced. A basal line extending from the lower margin of the foramen magnum along the palate to the end of the premaxillaries is nearly straight. The zygomatic arches are much expanded, but the malar is comparatively slender, and joins the maxillary in front of the orbit. The latter is of moderate size, and confluent with the large temporal fossa. The general form of the skull is indicated in the cut given below, figure 2. The maxillaries are massive, and on the sides behind the canines are deeply indented, giving a marked constriction to this part of the skull. The lachrymal forms the anterior border of the orbit, and its foramen is inside the orbital margin. The nasals are quite slender in front, and broad posteriorly. The premaxillaries are expanded transversely, giving a wide anterior narial aperture. The anterior palatine foramina are small, (fig. ure 2.) The posterior nares extend forward between the last upper molars. The dental formula is as follows:

3

1 1

4

Incisors ; canines; premolars; ; molars

3

3

3

gi X2=44. The teeth agree in all generic characters with those figured as Coryphodon by Owen and Hébert. The occipital condyles are well separated, and there is a condylar foramen. Between the basisphenoid and the periotic, there is a large opening, partially due to the divergence downward of the inner faces of the latter bones. There is a strong paroccipital process, and a postglenoid process, which varies in size in different species. In Coryphodon hamatus, it is long, and curved forward, and to this the specific name refers. The skull as a whole presents strong Perissodactyle characters.

The brain cavity in Coryphodon is perhaps the most remarkable feature in the genus, and indicates that the brain itself was of a very inferior type. It was quite small, as in all Eocene mammals, but its most striking features were the small size of the hemispheres, and the large expanded cerebellum. The form and relative size of these are shown in the accompanying cuts, figures 1 and 2.

The olfactory lobes were large, and entirely in advance of the hemispheres. They were bounded in front by a well ossified cribriform plate, and partially separated by a vertical bony septum. The cerebral lobes were ovate in form, and very small, a transverse section exceeding but little that of the medullar opening. In shape and relative size, the hemispheres and olfac

Figure 1.-Cast of brain-cavity of Coryphodon hamatus Marsh; top view. One-half natural size.

Figure 2.-Outline of skull and brain-cavity of Coryphodon hamatus; top view. About one-fifth natural size.

tory lobes of this genus are somewhat similar to those of Dinoceras. The cerebellum was proportionally large, and widely expanded transversely. Its peculiar form is shown in figure 1, which is drawn from a cast of the brain-cavity of C. hamatus. This portion of the brain nearly or quite equaled the hemispheres in size, thus differing widely from any known mammal. There is a well marked pituitary fossa, but no clinoid process. The foramina for the exit of the optic nerves are small, but for the others very large. The brain as a whole was very low in grade, and precisely such as might be expected in a mammal from the oldest tertiary deposits.

The skeleton of Coryphodon (Bathmodon) presents many features of interest, but only a few can now be mentioned. The limbs were comparativly short, and the femur has a third trochanter. The feet are especially interesting, as they present a primitive or generalized type. The manus and pes had each five short digits.*

The various characters shown in the skull and limbs of Coryphodon indicate that the animals of this genus were essentially five-toed Perissodactyles. They evidently represent a distinct family which may be called Coryphodontidae. Their geological horizon in this country is near the base of the Eocene, in the deposits named by the Survey of the Fortieth Parallel, under Clarence King, the Vermillion Creek series. The remains of the family at present known in this country are from Utah, Wyoming, and New Mexico.

Yale College, New Haven, April 15th, 1876.

* Prof. Cope has recently published (Catalogue of Eocene Vertebrata from New Mexico, p. 28, 1875), a figure of the "Hind foot of Bathmodon," which is incorrect in several respects, the hallux, for example, having three phalanges!

THE

AMERICAN

JOURNAL OF SCIENCE AND ARTS.

[THIRD SERIES.]

ART. LIV.-Note on the Duplicity of the "1474" line in the Solar Spectrum; by Professor C. A. YOUNG.

THE line "1474" is the one which is reversed in the spectrum of the Solar Corona, and coincides with one of the short lines in the spectrum of iron. In the iron spectrum it is brought out, however, only by the Leyden jar spark, and not by the electric arc between carbon points. As seen in the solar spectrum, with ordinary, or even very powerful, spectroscopes, it appears like a fine, hard, black line.

In examining this portion of the spectrum recently with a diffraction spectroscope, armed with a silvered glass "gitterplatte" of 8640 lines to the inch, for which I am indebted to the kindness of Mr. Rutherfurd, I find this line to be unmistak ably double; the two components are separated by a distance of only about of a division of Ångstrom's scale, i. e. about

of the distance of the D lines. The more refrangible component is heavier than the other and slightly winged or hazy at the edges, while the other is narrower and better defined. The more refrangible line is undoubtedly the real corona line, and the other belongs to the spectrum of iron, the close coincidence being merely accidental.

As long ago as 1870 I suspected the bright 1474, as seen on the limb of the sun, to be very slightly more refrangible than its dark analogue (the position of which with insufficient dispersive power would apparently correspond to the mean of the two components); and the suspicion has recurred from time to time on many occasions since then, while there has not been a single instance in which the bright line appeared to fall below AM. JOUR. SCI.-THIRD SERIES, VOL. XI, No. 66.-JUNE, 1876.

the dark one. Still there has never seemed to be sufficient evidence to warrant a positive assertion. Examination with a speculum metal "gitterplatte" of 6480 lines to the inch, some two years ago, suggested the idea that the dark line might be closely double, but the definition of the grating was not sufficiently good to decide the question. With the new one, however, there remains no doubt. Another grating with 17280 lines to the inch, which is temporarily in my possession, shows it nearly, though not quite as well.

The accompanying map of this region of the spectrum gives a good idea of the appearance of things, and the amount of dispersive power obtained. The scale numbers at the top are those of Ångstrom, those at the bottom are Kirchoff's.

The observations are best made in the spectrum of the 8th order (with the coarser grating-in the spectrum of the 5th order with the grating of 17280 lines to the inch) although the duplicity of the line is clearly visible in the spectra of the 6th, 7th, and 9th order, with proper precautions. The arrangement employed is that indicated in fig. 2. C is the collimator, the

beam of light, thrown into the room by a heliostat, being concentrated by the lens L which forms an image of the sun on the