1. The narrow and elongated tubular form of the corolla quite effectually excludes those which are large enough to set the lever in motion; 2. If such an insect, for instance a hive bee or small humble bee, should force its way into the tube, by the time its head had reached and elevated the sterile end of the lever, the tip of its abdomen would have passed the lowering polliniferous end, so that no pollen would reach the insect, and the object of the motion would be defeated. Bees might, to be sure, visit the flowers solely for their pollen, and I have no doubt that they occasionally do so, in which case they must often render some service in their fertilization, as is the case in so many flowers visited for pollen alone. Bees being excluded for the reasons above given, we turn to Lepidoptera, which sometimes visit the flowers, their long and slender proboscides enabling them to reach the nectar with little exertion; but it remains to be shown that these organs are sufficiently large or rigid to set the stamens in motion. Even if it should be shown that the large nocturnal moths do move the levers, which I am far from believing to be the case, the brilliant scarlet color is one ill adapted to rendering the flowers conspicuous in the twilight or night, and, so far as I know, one which is never possessed by flowers especially dependent upon these insects for their fertilization; beside which, we do not find that close constriction of the mouth or anterior part of the corolla bespeaking adaptation to the Lepidoptera. It appears, then, that when these insects visit the flowers of our sage, they may be of some use in transferring pollen, since their heads may encounter stigma and anthers, but they do this without rendering the motility of the latter of any value.

The only alternative, then, is birds, which, to be of the highest use in this connection, must be found in the native habitat of the plant, must visit flowers frequently for nectar, small insects attracted by the latter, or for both, and finally, must have slender and elongated beaks capable of insertion into the tubular flowers. All of these conditions are fulfilled by many of the humming birds, which reach their greatest number in both species and individuals in Equatorial America. The color of this Salvia is one of the most attractive to humming birds, and a glance at Figure 4 will show that one of those with an elongated beak cannot fail to operate the lever in the most perfect manner; its extensible tongue, however, rendering it by no means necessary for its beak to equal the corolla in length.

In a brief note published in the Botanische Zeitung for 1870, p. 275, Fritz Müller states that in Brazil the scarlet Salvias are frequently visited by birds, and although no species are named, there is little reason to doubt that the one under consideration was among those observed. Our single species of Trochilus, the ruby throat, possesses a beak rather short for the most efficient working of the staminal lever, yet from the statements of friends and from personal observation I can vouch for its frequently rendered service, and the greater part of the capsules of this plant which mature in our borders are to be credited to this active little creature.

Although this paper is confined to a single species of Salvia, it by no means follows that others may not offer examples of equal or even of greater adaptation to birds, and several such might be mentioned. The conclusion seems, on the whole, warranted, that several tropical American Salvias are as perfectly adapted to profit by the visits of birds as many other species of the genus are to profit by the visits of bees.

-:0:

ON THE ORIGIN OF THE FOOT STRUCTURES OF THE UNGULATES.

THE

BY E. D. COPE.

HE following considerations have been suggested by a study of the primitive types of the odd and even-toed ungulates. I first, in 1874, recorded the opinion that the Mammalia with a reduced number of digits were derived from pentadactyle plantigrade types. The ungulate order which fulfills this requirement is the Amblypoda, and from them, I doubt not, both the Perissodactyla and Artiodactyla have arisen, although not from any of the genera now known. Both of these great orders display a regular diminution in the number of the digits; in the former, by reduction and extinction on both sides of the third digit; in the latter, by reduction and extinction on each side of the third and fourth digits. Mr. John A. Ryder2 has pointed out that reduction in digits is probably directly related to strains and impacts. He reminds us that the anterior digits are reduced in Mammalia of unusual scansorial or fossorial powers; while in forms which display 1 Journal Academy Philadelphia, March, 1874.

2 AMERICAN NATURALIST, October, 1877.

powers of running, the reduction is seen first in the posterior feet, which propel the body much more than the fore feet. This view is well illustrated in the Perissodactyle families, the majority of which have the digital formula 4-3.

No reason has ever been suggested, so far as I am aware, in explanation of the fact that one series of ungulates has retained. two digits, and the other only one; that is, why there should have been two kinds of digital reduction instead of one kind. In seeking for an explanation, we will remember that the tarsus in the odd or single-toed line, is bound together by fixed articula

NI

FIG. 1.-Right posterior foot of a species of Coryphodon from New Mexico, onehalf nat. size. From Report Expl. W. of 100th Mer. G. M. Wheeler, Iv. Pl. LIX. FIG. 2.-Right posterior foot of Aphelops megalodus Cope, from Colorado, onehalf natural size. From Report U. S. Geol. Surv. Terrs. F. V. Hayden, Iv, Pl. cxxx. tions, while in the cloven footed line it is interrupted by the hinge between the first (astragalus), and second rows of bones. The hinge-joint being more liable to luxation than the fixed articulation, requires a wider basis of support, such as would be furnished by two divergent digits, rather than by a single central one.

In the early types, where the median digits are slender, the mechanical advantage in favor of the bidigital over the unidigital arrangement is much more obvious than in modern genera. Late in time, the horse developed the middle digit to such a width as to form almost as good a support as the bidigital structure. In

the Eocene genera, the slender median digit could not have sustained the weight on a hinge, without great risk of dislocation. This explanation it can be said, applies only to the posterior foot. The posterior foot has, however, led the way in the evolution of Ungulata, and the fore foot may have followed in accordance with the law of antero-posterior symmetry in growth. A curiously malformed deer from Mendocino county, Cal., throws some

FIG. 4.

FIG. 3.

FIG. 3.-Right posterior foot of Protohippus sejunctus Cope, from Colorado, about one-half natural size.

From Report U. S. Geol. Surv. Terrs. F. V. Hayden, IV.

FIG. 4.--Right posterior foot of Poëbrotherium labiatum Cope, from Colorado, three-fifths nat. size. From Hayden's Report, IV, Pl. cxv.

light on this subject. It has apparently a single functional digit

on each foot.

Examination shows that the posterior foot is

bidigital, but that the phalanges are fused; while the anterior foot

tal! Similar evidence is furnished by the genus Eurytherium of the French Eocene. Its posterior foot is modified artiodactyle, while the anterior is modified perissodactyle. We may assume from these facts, that the posterior foot is more subject to the influences which tend to produce the bidigital structure than is the anterior limb.

The

I suspect that the production of a ginglymus in the middle. of the tarsus, has been due to the use of the posterior limb in soft swampy ground. In the absence of this condition, as in a life on harder ground than swamp, no ginglymus would be formed. The action of an ungulate in walking through deep mud is very suggestive. The posterior foot is bent on the leg, and the antero-posterior strain of the weight or propulsive force, is transverse to its long axis. In progression on dry land, the impact is in the direction of the length or axis of the foot. obvious effect of a cross strain is to produce by degrees greater and greater mobility of some articulation. The one which has yielded is that between the two tarsal rows. Another effect of walking in swampy ground is to spread the digits apart. As the first digit of both feet is always of reduced size, there are practically but four digits to be considered. The weight falling nearly medially on these, would tend to spread them equally, two on each side. Thus the same cause may have been effective in producing both FIG. 5-Left fore the artiodactyle structures. The perissodactyle foot with part of radius of Poëbro- structure, so soon as the lateral digits are much therium vilsoni Lei- reduced, ceases to be favorable for progression

dy, from Colorado,

three-fifths nat size. in soft ground, owing to the liability of the lateral From Hayden's Re- digits to injury, in following the principal one into the yielding material, filled with sticks and

port, IV, Pl. cxv.

other hard débris.

The lowest existing forms of the Artiodactyla, the Omnivora, are universally swamp lovers and livers. So we are told are the lower existing Perissodactyla, the tapirs and rhinoceroses. The higher types of both orders are dwellers on plains and in forests. We do not know the habits of the Eocene Perissodactyla, but I doubt their having inhabited muddy ground to the same extent as