the image must be indistinct, for the rays diverge from an entirely different distance. But gradually the lenses adjust themselves to the actual divergence, i. e., for rays diverging from the real object, while the optic axes remain adjusted for the distance of the image. The difficulty experienced in dissociating these two adjustments causes the interval of indistinctness. The perception of the difference between the image and a real object is the sense of this dissociation. Consensual movements have been, perhaps, brought about by the necessities of single and distinct vision; Helmholtz has shown that other consensual movements may be dissociated when the necessities of single vision require it; these experiments show that the consensual adjustments of the eye may be dissociated when the necessities of distinct vision require it.

I was now anxious to determine what part was taken by the pupil. Is the contraction of the pupil more intimately associated with the axial or the focal adjustment? This question has been discussed by E. H. Weber, Cramer and Donders.† Weber believes it is directly associated with the axial adjustment, Cramer and Donders with the focal adjustment. To test this question, while I was obtaining the image and making it clear, an assistant standing behind and a little to one side observed my pupil reflected in a small mirror conveniently placed. After gazing intently at the real object until the pupil was steady, as soon as I converged the optic axes so as to obtain the image No. 1, the pupil was observed to contract decidedly, but as the image became clear it again expanded to its original size. Again at the moment of obtaining the 2d image the pupil contracted still more strongly, but as soon as the image became clear it again expanded, nearly, if not entirely to its original size. The same phenomena were observed for each of the images, only that in the nearest images when the convergence of the optic axes was extreme and the first contraction very great, the pupil did not return entirely to its original dimensions.

I then made similar experiments on the image beyond the real object. As before, I looked intently first on the real object at the distance of twelve inches until the pupil became steady. So soon as I gazed beyond the object the pupil of course expanded; but as soon as the image became clear it again contracted to nearly its original size. In this last experiment the pupil is apt to be unsteady. This might have been expected; for as we have already said, it is much more difficult to obtain this image clear or to retain it when obtained.

There is no doubt of the fact, therefore, that the contraction *Proc. Roy. Soc., April, 1864.

+ Donders, accommodation and refraction of the eye, Trans., p. 574.

of the pupil is most intimately associated with the focal adjustment.

I believe that this principle of dissociation of consensual adjustments explains perfectly certain phenomena of the stereoscope. It is well known that many persons experience difficulty in seeing stereoscopic pictures distinctly even when the two pictures are brought into perfect coincidence and I believe all persons experience some fatigue to the eyes in looking at stereoscopic pictures for a considerable length of time. I have often felt both the difficulty and the fatigue, though to a much less degree than most persons. The explanation of this difficulty is as follows. We judge of distance, as is well known by the axial adjustment. If then the two pictures are so taken that, in order to bring them together, the visual lines must meet at a certain distance, say fifty yards, then the picture will be seen at that distance and of course very much enlarged. But in order to see the picture clearly, the rays must come to the eye as if they diverged from the same distance; for the eyes are adjusted for that distance. To fulfil this condition lenses are always used; but it is obvious that a given pair of lenses are suitable for one distance only. For all other distances, or degrees of optic convergence, there must be some degree of dissociation of the two adjustments, and this is both difficult and fatiguing to most persons.

I have found that observations upon the images of the ruled diagram are a most delicate means of determining both the rotations of the eye, and the position of the Horopter. I hope in my next communication to take up this most difficult subject. (To be continued.)

ART. VII.-On the Geology of Lower Louisiana and the Rocksalt Deposit of Petite Anse; by EUGENE W. HILGARD, PH.D., of Oxford, Miss. (Abstract.)

THE discovery in 1862, of a deposit of rock-salt on the coast of Louisiana, was a fact so unexpected to geologists, that at any other time a detailed investigation cf its geological relations would quickly have followed the first announcement. The pressing necessities of the blockaded section soon caused its exploitation on the large scale, though in a very irregular manner; for a considerable period, these mines supplied the whole of the southwest. In November, 1865, Prof. Richard Owen made a brief examination of the locality, the results of which he published in the Transactions of the St. Louis Academy. A year

later Dr. Charles A. Gössmann, under the auspices of the American Bureau of Mines, made an examination of the locality, mainly with a view to the exploitation of the deposit; his report, published by the Bureau, as well as the specimens which he courteously exhibited to me, confirm previous conjectures that the overlying strata were the equivalents of the formation I have described as the "Orange Sand" of Mississippi. I therefore gladly availed myself at the earliest possible moment, of the offer of the Smithsonian Institution to defray my expenses in making a detailed geological investigation of the region. The low stage of water prevailing at the time (December, 1867,) rendered it possible to observe to the best advantage the formations exhibited on the banks of the Mississippi; the examination of which, from Vicksburg to the Passes, was a needful preliminary step to the determination of the formations of the coast.

Having previously examined and described the sections exhibited at Vicksburg, Grand Gulf and Fort Adams, I merely landed at some intermediate points to verify the conclusion previously reached, viz., that below Vicksburg, no marine formation crops out on the river banks, reports to the contrary notwithstanding; and that the profiles at Natchez, Rodney and other points are essentially similar to that at Fort Adams, where we find the strata of the (fresh-water) "Grand Gulf group " in a position nearly or quite horizontal; overlaid, first by the materials of the "Orange Sand," which in its turn is capped by the stratum of the "Loess" or Bluff formation, covered by a thin deposit of "Yellow Loam."

Facing southward from the "Blockhouse hill" at Fort Adams, we observe a wilderness of the characteristic sharp ridges of the Loess region, often fore-shortened into veritable peaks, elevated between 300 and 400 feet above the river. In this region, the Grand Gulf strata have been traced southward by Dr. George Little, the present State geologist of Mississippi, as far as the head waters of Thompson's Creek, northwest of Clinton, La.

The Orange Sand proper is visible, near the river, as far south as Jackson, La., but farther inland extends to a lower latitude. As for the Loess, it appears in full force and characteristically developed for some distance south of Fort Adams. But (according to Dr. Little's observations) these features become gradually modified as we advance southward. The Loess deposit thins out, its materials become poorer in lime and fossils, and assume more and more the character of a common fine grained "hardpan;" the transition being by insensible de

* See Report on the Geology and Agriculture of Miss., 1860.

grees, while the two extremes are very obviously distinct. At the same time, the clayey substrata which, farther above appear only in patches (as at Nevitt's bluff, two miles above Natchez, as well as at the latter place itself) are seen more frequently and continuously, until, at Port Hudson, they become predominant.

The exposure at Port Hudson, previously examined in part by Bartram, Carpenter and Lyell, is about three miles in extent, from the mouth of Sandy Creek above the town, to Fontania Landing, 1 miles below. Its lower half is washed, and continually encroached upon, by the river; its upper portion is now inland of an extensive sand bar. The strata are disposed horizontally or basin fashion, and vary a good deal both in thickness and materials, as shown in the subjoined profiles, situated about a mile apart; the correspondence of strata is ascertained by actual tracing of the stratification lines.

At the stage of extreme low water prevailing at the time, the stump-stratum No. 1 was visible to the thickness of 10 ft. at its highest point; showing several generations of stumps above one another, also the remnants of many successive falls of leaves and overflows. The wood is in a good state of preservation; no prostrate trunks to be seen at present.

The main clay deposit, No. 2, varies but little in general character; although very solid, its tendency to cleave into prismatic forms renders it very liable to "cave" into the river. The upper portion of the stratum, especially near its southern end, contains strings of calcareous nodules, on stratification lines eight to twelve inches apart. No fossils save rare impressions of leaves.

No. 3 is exceedingly variable. At the northern end of the outcrop, it is a narrow band of swamp deposit; at the first of the profiles given, it bears the character of a sandbar; lower down, it returns to that of a swamp deposit; still below, it is represented by a fine white silt, without a trace of vegetable remains. Lower down again, a lignitic layer appears at its base, with leaves and fruit of living species of lowland trees; while near Fontania, it is again a sandbar, with an abundance of prostrate trunks of driftwood, coarse sand and pebbles.

The green clay stratum No. 4 varies little, either in thickness or composition, and like the stump-stratum No. 1, forms a convenient level of reference.

The hardpan stratum No. 5 I conceive to be the more immediate representative of the Loess proper, with which it is connected by gradual transition, though at times greatly resembling some of the materials of the Orange Sand. It is void of fossils.

The present profile differs in many respects from those given by previous observers, which lay some distance farther west, where the river now flows. The strata are accordingly as variable in an east and west, as in a north and south direction, and with the exception of Nos. 1 and 2, are such as are now shown in ditches cut into the modern river-bottom deposits.

The stump-stratum No. 1, however, as appears from numerous data collected by myself or contained in Humphreys and Abbot's Report on the Mississippi river, exists at about the same level (i. e., near that of tide-water) not only over all the socalled Delta-plain of the Mississippi, but also higher up, perhaps as far as Memphis, and all along the gulf coast, at least from Mobile on the east to the Sabine river. Wherever circumstances allow, the overlying clay stratum No. 2, is also observed. These facts indicate the wide spread prevalence, during the epoch succeeding the drift, of quiet, shallow freshwater lagoons and swamps of slightly varying elevation; through which the continental waters may for some time have found an outlet without a definite channel representing the Mississippi of to-day. The Port Hudson profile appears to be typical, its features being reproduced wherever denudation has not removed these deposits down to the level of the stumpstratum, as is mostly the case.

The Five Islands.

The chain of five islands rising partly from the sea, partly from the coast marsh, between the mouth of the Atchafalaya and Vermilion river, have been described by Mr. Thomassy,* * Géologie pratique de la Louisiane; New Orleans, 1860.