to our last written equation we get x=694 a.

Similarly, from D3, (n=20), I find x=64.5a; and from F, (n=10) x=111·1a.

The last value is evidently too large, the discrepancy probably arising from the fact that Mr. Pye's instrument having a prisin of the extra dense and very yellow flint-glass considerably reduced the intensity of the F line by absorption.

Thus far, we have neglected the inequalities of shading and illumination; but as both chromosphere and leucosphere decrease in brilliancy pretty regularly from the limb of the sun, and since the upper limit of the former is much the more definite, it is evident that the introduction of these considerations would increase the computed ratio. We may, I think, then safely assume that the angular area of the leucosphere is at least 70 times as great as the area of that portion of the chromosphere and prominences which was exposed to view during the totality. I fear it is not possible to estimate this area with much certainty, but think it could hardly have been equivalent to anything less than a ring of 9" or 10" wide surrounding the sun. If so, we

find the leucosphere to be equivalent to a ring of about 10' in width, and considering the irregularity of its outline, some of the angular prolongations would be likely to attain twice that distance.

I am, of course, aware that the numerical data of this calculation are very uncertain, being mere estimates, not measurements; but the principle is correct, and the results are not sensibly affected by any ordinary amount of atmospheric influence, since the different colored rays would not have the ratio of their brilliancies much disturbed under any reasonably fair conditions.

The result agrees very well also with the photographs, drawings, and estimates of skilled observers.

As for the long rays, both bright and dark, which are often, and perhaps usually, seen crossing the leucosphere and reaching far beyond, the question whether they are also solar appendages appears to be very different and far more doubtful.

In the first place, I doubt if they are always present. At any rate, in 1869, in a most exceptionally clear atmosphere, I saw nothing of them at Burlington, Iowa, and believe (but am not absolutely certain) that none of our party did; while some observers in Kentucky saw them well, and at Sioux City they were especially conspicuous, as is evident from the remarkable, and I have no doubt, accurate picture given by Mr. Gilman in the Eclipse Report of the Naval Observatory.

In December, so far as I have heard, every one saw them. I have no hesitation in affirming that the corona as it appeared to me then was a very different phenomenon from what I saw the year before, and far more complex.*

The photographs also taken by Lord Lindsay's and the American parties in Spain, appear to differ essentially from each other, and from those taken by Mr. Brothers, in Sicily; and that in such a way as to suggest at least, that this more extensive radiance is of a far less permanent character than the leucosphere, and possibly of a different origin.

Should it turn out to be visible only where the sky is hazy, it might then be regarded as an effect of our own atmosphere produced, of course, not by the ordinary photospheric sunlight (which, as has been abundantly shown, by many writers, cannot during totality, illuminate the air near the moon's place), but by the light from the prominences and the lower regions of the leucosphere.

If, on the contrary, as seems to be the case, this radiance is often seen under unexceptionable atmospheric conditions, we appear to be shut up to one of two theories; either on the one hand that of Prof. Norton and Mr. Proctor, whose views regarding these rays are nearly identical, and represent them to be streams of matter, similar to cometary substance or auroralt beams, driven off by some solar repulsion; or on the other hand that of Oudemans, who considers them to be purely optical effects produced in cosmical dust between us and the moon, by the sunlight streaming across the uneven and ragged edge of our satellite.

With reference to the former theory, it is probably sufficient * Mr. Lockyer, in "Nature" for Feb. 23d, quotes from a letter of mine written a year ago, to show that my opinions regarding the nature of the corona have been considerably modified since then; and this is true to some extent, though I think the present approximation of our views is owing quite as much to a change in his own ideas-as would be evident on referring to his papers of the same, and even somewhat later date. But I should still write "I am strongly disposed to believe that the whole phenomenon" (i. e., the Corona as I saw it in 1869) is purely solar.” Since my name has sometimes been referred to in connection with the socalled "Auroral Theory of the Corona," it is proper for me to state, that I make no claim to be its originator.

When I discovered (as I supposed), the identity of the bright line in the corona spectrum with a line in that of the aurora, and announced my belief in the substantial identity of the two phenomena, I considered myself as simply subscribing to a view already current, and bringing a new argument to its support.

So far as I know, Prof. Norton was the first to work out and publish a connected theory of the subject, basing his conclusions largely upon his discussions of Donati's comet, which were printed in this Journal some years ago. Prof. Winlock also informs me that he has held and published a very similar opinion, and so I think have more than one of the European astronomers, though I cannot now give references. My own father, more than twenty years ago, was accustomed to teach from the same chair of astronomy which I now occupy, an essentially similar doctrine. Thus the idea had long been familiar to me, and I presume more or less so to astronomers generally.

to refer to Mr. Proctor's recent work on the sun, and to a paper which Professor Norton has published in a late number of this Journal, wherein his views are fully explained. According to this view, the leucosphere and the rays are alike solar appendages, and of identical origin and material.

The theory of Oudemans is stated in a letter to "Nature," contained in the number for Nov. 10, 1870.

With good weather, it would seem possible to decide between these two hypotheses at the next eclipse.

A series of corona photographs taken as rapidly as possible on the plan pursued by Mr. Brothers (who has shown that an exposure of from 8 to 10 seconds is sufficient to produce a fine picture) would probably indicate whether the dark streaks are really related to mountains on the edge of the moon or not.

It is probable also, as Prof. Pickering has suggested, that important information may be obtained by observing them with a spectroscope of high dispersive power and widely opened slit, using the 1474 line just as the Ĉ line is used in examining the prominences.

The telescope, to which the spectroscope employed for the purpose is attached, should be of wide angular aperture, so as to give abundant light, but of very short focus, forming an image of the sun not much more than of an inch in diameter, in order that the rays may be so little magnified that their outline, if they are really of leucospheric origin and give the green monochromatic light, may be readily observed through the slit.

There remains still unsettled another interesting series of questions concerning the nature of the substance composing the leucosphere, and the relation of this envelope to the sun; whether it be a true atmosphere, or a mere cloud of transient particles a meteor-flock, as Mr. Proctor imagines.

Apart from the difficulty of supposing such a multitudinous. and continual supply of meteoric matter as this theory would require, and neglecting all consideration of the peculiar form assumed by this envelope, which seems to be deepest precisely over those solar latitudes where the spots and prominences are most numerous, and even to be governed in the minutiae of its outline by the position of its prominences, I find what seems to me an almost insuperable objection to it in the powerful winds and cyclones which prevail in the region above the chromosphere at elevations of from 50,000 to 100,000 miles.

These winds, by which the tops of the solar flames are whirled and driven, present, so far as observations now go, every characteristic of true aerial currents in a continuous medium; and the whole appearance and behavior of the solar protuberances, except at the moment of eruption, is that of clouds floating in an air.

But if we then consider the leucosphere as a true solar atmosphere, how can we reconcile its enormous extent with the known smallness of the pressure at its base, determined by the experiments of Lockyer, Frankland, and Wullner? In either of two ways; first, and perhaps on the whole most probably, this atmosphere may consist of some new kind of matter whose density is far below that of even hydrogen; or it may be composed of matter whose specific gravity (not density) is diminished, annihilated, or even rendered negative by some such solar repulsion as appears to be operative in the formation of a com

et's tail.

There is no doubt that the line which characterizes its spectrum coincides with one of the lines of the Iron spectrum within the limits of any present means of observation: and so close a coincidence can hardly be accidental. And yet in the spectrum of iron, this line is only a faint and unimportant oneone of the last to make its appearance under the stimulus of the electric spark, and so little comparable in intensity with many others in its immediate neighborhood, that Mr. Huggins failed to map it on his spectral chart.

It is certainly difficult to understand how, if this line be really of the same origin as its fellows, it should remain the sole survivor of changes which have been able to exterminate from the spectrum all its more conspicuous associates; and accordingly, from this point of view, it becomes natural to suppose, as I suggested in 1869, that when the line appears in the spectrum of iron it may be due not to the iron itself, but to some associated substance (possibly standing in relation to the peculiar magnetic properties of this remarkable metal)—some occluded gas, which can also exist free in a state of inconceivable tenuity, as we have it in the leucosphere, and probably also in the streamers of the aurora, and in the tails of comets*. -a near relative, so far as gravity is concerned, to the luminiferous ether and to the Urstoff of German speculators.

The view of Mr. Lockyer agrees with this in supposing the leucosphere to consist of some new form of matter.

On the other hand, it is to be noted that alterations of pressure and temperature do produce in known spectra great changes, somewhat such as would be required in order to reduce the complicated spectrum of iron to this one line. (But I do not know of any case where one of the unimportant lines is the last to disappear.)

* It is perhaps worthy of remark that in the spectrum of Brorsen's comet, as given by Mr. Huggins, the position of one of its three bands appears to coincide with 1474, so far as can be judged from the figure and the measurements, which, however, do not permit auy very rigid testing of the matter. It seems somewhat probable that this same line will be observed in the spectrum of the tail of the first bright comet offered to observation.

Furthermore, those other lines of iron, which are often seen in the chromosphere spectrum, are nearly all of about the same order of prominence as 1474. The more conspicuous iron lines seldom if ever appear reversed, while in the case of other substances their strongest spectral lines are always the first to turn bright.

If then we admit a sufficient repulsive force, it seems still possible to suppose that the leucosphere may consist of iron in the state of vapor and fog; and the well known wide diffusion of this metal in meteoric matter makes it comprehensible how its lines should occur in the spectrum of our own terrestrial aurora, and in any other places where they may be found.

Possibly future researches in the laboratory may throw clearer light upon the subject.

It is hardly necessary to add that our own terrestrial atmosphere, when clear, appears to me to play only a very subordinate part in the phenomenon. Some influence it must, of course, have; but remembering how much the inner portion of the coronal ring exceeds in brilliance the outer, it would seem that the illumination of the lunar disc must give us an exaggerated measure of the true atmospheric effect. This illumination makes the edge of the moon only enough brighter than the center to give it the appearance of a globe, but of almost inky blackness.

With the subjective element, the case is very different. In untrained observers especially, it may be so influential that two intelligent persons standing side by side will describe and even sketch upon paper appearances most grotesquely different from each other and from the truth.

Even skilled observers are greatly affected by the strangeness and peculiar nature of the phenomenon, and the excitement of the occasion.

Hanover, March 23d, 1871.

ART. XLVII.-On the supposed Legs of the Trilobite, Asaphus platycephalus; by JAMES D. DANA.

AT the request of Mr. E. Billings of Montreal, I have recently examined the specimen of Asaphus platycephalus belonging to the Canadian Geological Museum, which has been supposed to show remains of legs. Mr. Billings, while he has suspected the organs to be legs so far as to publish on the subject, has done so with reserve, saying, in his paper, "that the first and all-important point to be decided, is whether or not

* Q. J. Geol. Soc., No. 104, p. 479, 1870, with a plate giving a full-sized view of the under surface of the trilobite, a species that was over four inches in length.