The hind leg is much like that of Mastodon, both in femur and tibia, though the fibula is placed rather more on the outside of the tibia, to which it joins by a small round articular facet.

We have many foot bones, with which, and Mr. Marsh's plates, I can restore his elephantine feet.

To clothe this skeleton with flesh requires, of course, some imagination. As for the so-called "horn-cores," I see no reason to suppose that they ever carried horns, and the convenient term "Dinocerata" is probably a misnomer; neither they nor the upper surface of the skull show any large channels for blood vessels, or foramina for nerves, and hence it is probable that they had a covering of thick skin only.

The enormous basin between the last pair and the occipital crest is also a problem. I cannot conceive of it as empty, I cannot imagine a filling for it.

Rejecting the proboscis I can only suggest a pair of large flexible lips, and it is probable that the throat was much like that of Rhinoceros.

Certain rugosities on the temporal bone suggest to Mr. Cope that the ear was large and elephantine.

The legs were undoubtedly large, and the tail broad and flat.

TRANSFORMATIONS OF PLANORBIS AT STEINHEIM, WITH REMARKS

ON THE EFFECTS OF GRAVITY UPON THE FORMS OF SHELLS AND

ANIMALS. By ALPHEUS HYATT, of Boston, Mass.

[ABSTRACT.]

THE limits of this abstract do not permit me to give an analysis of the work done by the author at Steinheim, or yet of Hilgendorf's remarkable essay which led to these investigations. We both agree on the theoretical presentation of the results, that is,

1 Steinheim is a village on the eastern declivity of the mountains in the eastern part of Wurtemburg in southern Germany, and has been noted for many years for the peculiar series of fossil shells which occur in great numbers in the strata of sand and limestone in the pits close to the village. For further information see Genesis of Planorbis at Steinheim, by the author. Anniversary Memoirs of Boston Society of Natural History, 1880.

2 Ueber Planorbis multiformis in Steinheimer Süsswasserkalk, Monatsberichte, K. P. Akademie, Berlin, July, 1855, p. 474, 1 plate.

in the opinion, that in this small lake in Tertiary times a large number of distinct forms of very different aspects were evolved from an older Tertiary species, the Planorbis levis.

Our disagreements though numerous and interesting involve minor points, except in one important particular.

Hilgendorf, whose pamphlet on these shells is said by some Germans to be the only perfect demonstration of the doctrine of evolution of organisms ever printed, found all the conditions of the problem in order.

He collected the parent forms in the lowest formation, and all the distinct descendant forms of the series in the successive formations lying one above another in their proper genetic order and sequence. Thus, not only did he trace several series of species, one into another, through the formations beginning with the lowest and earliest deposited, but was able to state that their relations in time or succession precisely accorded with their apparent genetic relations.

The evidence, as he first published it, and which he still asserts, after renewed investigations, to be true, is more perfect than that of any genealogical tree.

It would be similar to the genealogy of a family, if we could trace the members back into geologic times, and show not only that the existing descendants were derived from a common stock, but that these were vastly different from each other now, and from their ancestors.

My own observations refuse to be reconciled with his so far as the perfection of details is concerned, and in my opinion, though the evidence is exceptionally perfect, there are in this place, as in others, the usual gaps resulting from the imperfections and the necessarily fragmentary nature of geological history. While, therefore, I cannot join him in believing in the perfection of the record, my results after all are not theoretically contradictory, and I can only account for the appearances by the same theory of the descent of numerous specifically distinct forms from one species. The results are roughly exhibited on Plate I, and may be described briefly as follows. Figs. 1, 8, 12 and 16, are the ancestors, varieties of Planorbis levis from the older Tertiaries of another locality, identified, named, and kindly sent to me with

For a fuller account see Memoir above quoted.

eleven other specimens of this species by Prof. G. Sandberger,4 who opposes the evolutionary conclusions of both Hilgendorf and myself.

From these four varieties spring four distinct lines of descent. Fig. 1 begins the series from 2-7 in which of course numbers of the connecting forms are not figured. Fig. 8 begins the series from 9-11 much shorter and containing fewer forms than in series 2-7. Fig. 12 also gives rise to a short series with only few forms. Fig. 16, however, is the starting-point for a compound series, or one composed of at least three sub-series, 19-20, 21-24, and 25-28.

The intermediate forms by which the gap between the four ancestors and the four first forms of each series, viz.: 2 and 9 and 17, which occur in the Steinheim basin, is very complete, but necessarily left out in this plate.

Numbering the series from left to right we see that Series I has three sub-series. Two of these show a tendency to uncoil, to become distorted and smaller than the ancestor, fig. 16, while the third decreases in size, but has a form, fig. 20, which is turreted like figs. 11 and 6.

Series II maintains size about the same throughout, but becomes flatter than the originating form fig. 12. Series III grows sensibly larger, and 10-11 are turreted shells with a more ridged and sub-angular form of whorl than the primal form of fig. 8.

Series IV exhibits not only greater increase in size, but vastly greater differences in form and in other characteristics of the shell from fig. 1, with which it started.

We can, therefore, without fear of error call series IV a highly progressive series; Series II a persistent series; sub-series 3 of Serics I a partly retrogressive series; sub-series 2 of Series I, a purely retrogressive, and sub-series 1 of Series I also partly retrogressive; since, though it decreases in size and becomes deformed and uncoiled, it also has a tendency to produce a new characteristic, the transverse ribs, and also increases in size its more closely coiled forms, as in fig. 26.

There is also other testimony going to show that this classifiIcation is correct. Semper's researches on Lymnæus stagnalis show, that under the most favorable physical conditions, this species increases to a maximum of size and has larger whorls, while

The author of the most complete Memoir on Tertiary shells in existence, Conchylien d. Vorwelt.

under less favorable conditions with relation to food and temperature, the size is very much decreased.

The immediate results of weakness, produced by wounds, are also important in this connection.

Pl. 2, figs. 21, 21a and 22, a diseased Pl. oxystomus, var. revertens HILG., is a diseased specimen of the same species as fig. 9, Pl. 1. Compare this diseased, partly uncoiled, shell with the species, fig. 23, Pl. 1, Pl. denudatus and minutus.

The weakness consequent upon old age is equally significant and has a similar meaning.

Pl. 2, fig. 22, represents the effect of old age in distorting the growth of the outer whorl of Pl. oxystomus. Compare with fig. 22, Pl. 1. These are true cases of disease of comparatively rare occurrence in Pl. oxystomus. I have in my collection many similar cases and have lately found some diseased shells of Pl. trochiformis, fig. 7, Pl. 1. These are dwarfed, and show a tendency to unwind the spiral, so that they look remotely like the beginning of a series of transition forms from Pl. trochiformis to Pl. denudatus, fig. 24, Pl. 1.

This statement is proved in my Memoir by numerous instances, and also by the Magnon Planorbis figured by M. Piret. Many of these acknowledged to be distorted are precisely similar to fig. 24. These distortions are produced at Magnon, France, undoubtedly, as in these other instances, by the unfavorable physical surroundings. I have also found near St. Johns, New Brunswick, in company with a complete series of uncoiling Valvata, like those of Planorbis on Pl. 1, Series I, sub-series 2, a number of distorted Planorbis, the counterparts of many of M. Piret's figures.

Thus we can assume without fear of error, that the increase in size and the progression of Series IV were due to the favorable conditions of the Steinheim basin, for that series; and that of Series I to the unfavorable conditions, and the mingling of characteristics or of progression and retrogression in sub-series 3 of Series I; and Series II and III to the different extent which these series are affected, either favorably or unfavorably by the same surroundings.

The effect of heredity is seen in the fact, that however great the modifications, as in fig. 7, the young are smooth and like fig. 6, though smaller of course, while the young of fig. 6 is similar to a small copy of 5, and the young of fig. 5, to a miniature fig. 4.

The young of fig. 4, however, may be said to be like fig. 2, first, and then like fig. 3 before it becomes angular as in fig. 4. There is also a very early stage in all these shells,5 which is a mere bag and is called the ovi-shell or embryonal shell. The first turn, which the young makes in growing around this embryonal shell, has a cylindrical mouth whorl like miniature fig. 8, Pl. 1.

This stage is very prolonged in fig. 3, much shortened in fig. 4, and still more restricted to the internal whorls in fig. 5, though in some specimens very much more prolonged than in others. In figs. 6-7 it remains about stationary, the size and aspect of the young being similar to fig. 8 or 9 or 2, during the earlier stages, until the shell is between one-hundredth and four-hundredths of an inch in diameter.

This is heredity with acceleration, or the inheritance of the characteristics of the original species, fig. 8, at earlier and earlier stages in successive species.

It is the same history with the characteristics which first appear in the adults of the series. The angular form of the outer whorl of fig. 4 is often in Pl. tenuis (fig. 4), confined to the outer whorl or whorls built during the adult period of growth. This becomes confined to the younger stages or inner whorls after a certain number of intervening intermediate generations in fig. 5, and are also left out in many specimens of figs. 6 and 7, in whose young they are replaced by the stouter form and heavier ribs and ridges of fig. 5.

The same is true of transverse ridges of fig. 25.

These with few exceptions first appear upon portions of the outer whorl.

Then in successive generations they appear at earlier stages, that is upon the internal whorls, and grow larger in the adults.

The uncoiling tends to appear in the same way first, upon the outer whorls, and it must be noticed, that in no case is the innermost of all the whorls uncoiled, or the bag-like form of the embryo changed.

Such facts show that heredity has the power, even in spite of the forces which changed the adults of the Series IV so completely in this locality, to reproduce or continually rejuvenate the ances

The centre' fig. 21a, and fig. Sa, Pl. 2.

For enlarged figure of this early stage see Pl. 1, figs. 29-30.