owing to the minuteness of its particles, and consequent (relatively) enormous resistance of the air. This illustrious savant is disposed to think that the consideration of Cagniard de Latour's experiments on the vaporization of liquids under high pressure would incline us to regard the solar faculæ as large aggregations of bona fide solid matter of a high degree of fixity, and in masses like gigantic soot-flakes of any form and magnitude, which, when formed, settle down to such a level as corresponds to their density when they rest in æquilibrio in a gaseous fluid of their own specific gravity. We do not wish either to accept or to reject this hypothesis, but would frame the following statement, which also includes this view of the case: Solar faculæ consist of solid or liquid bodies of a greater or less magnitude, either slowly sinking or suspended in æquilibrio in a gaseous medium. 25. In connection with this part of our subject it will be well to investigate the relative position of spots and their accompanying faculæ; and this is done in the following table for all the Kew pictures available for this purpose.

[Table III is omitted.]

26. It appears from the result of table III, that out of 1137 cases 584 have their faculæ either entirely or mostly on the left, while 508 have it nearly equally on both sides, and only 45 mostly on the right. Hence we see that faculæ are on an average to the left of their accompanying spots. The most obvious explanation of this would be that the faculæ of a spot have been uplifted from the very area occupied by that spot, and have fallen behind to the left from being thrown up into a region of greater velocity of rotation. All this is quite in accordance with our hypothesis regarding the nature of faculæ. We would likewise here remind our readers that we know from the observations of Kirchhoff that the sun's atmosphere contains vapors of substances, such as iron, which are condensed into the liquid or solid state at a comparatively high temperature. Now is it not natural to suppose that in the sun's photosphere we do really see such vapors so condensed, and very unnatural to imagine that such vapors are seldom or never condensed, and that what we really see is an incandescent plain underlying these vapors?

27. Let us now attempt to answer the third question: Is a

AM. JOUR. SCI.-SECOND SERIES, VOL. XLIII, No. 128.-MARCH, 1867.

spot including both umbra and penumbra a phenomenon which takes place beneath the level of the sun's photosphere or above it? To decide this question, let us state that there are a good many instances in which a spot breaks up in the following manner. A bridge of luminous matter of the same apparent luminosity as the surrounding photosphere, and unaccompanied by any penumbra, appears to cross over the umbra or center of a spot. There is good reason to think that this bridge is really above the spot; for were the umbra an opaque cloud, and the penumbra a semiopaque cloud, both being above the sun's photosphere; it is unlikely that the spot would break up in such a manner that the terrestrial observer should not perceive some penumbra accompanying the luminosity. Again, detached portions of luminous matter appear to move across a spot without producing any permanent alteration. We are on these accounts disposed to think that a spot including both umbra and penumbra is a phenomenon which takes place beneath the level of the brighter part of the sun's photosphere.

28. Let us here recapitulate the answers we have given to our three questions.

(1.) The umbra of a spot is nearer the sun's center than its penumbra, or, in other words, it is at a lower level.

(2.) Solar faculæ, and probably also the whole photosphere, consist of solid or liquid bodies of greater or less magnitude, either slowly sinking or suspended in æquilibrio in a gaseous medium.

(3.) A spot including both umbra and penumbra is a phenomenon which takes place beneath the level of the sun's photosphere.

§ VII. Concluding remarks.

29. It would thus appear that the central part of a spot is nearer the sun's center than the penumbra, and that both the umbra and penumbra are probably beneath the general level of the surrounding photosphere. Now the umbra or lowest part of a spot is much less luminous than the general photosphere. But what does this probably imply, according to the laws with which we are acquainted? It implies that in a spot there is probably some matter of a lower temperature than the photosphere. For is it not now recognized as a law, that if a substance, or combination of substances, of indefinite thickness and surface of small reflecting power have all its particles at a certain fixed temperature, this substance will give out nearly all the rays of heat belonging to that temperature? Now the sun, even when we look into a spot, is certainly a substance of indefinite thickness; and since a spot appears much less luminous than the ordinary surface, ought we not to conclude either that we there view matter of a lower temperature than the ordinary

surface, or that the matter which appears within a spot has a very high reflecting power compared to the ordinary matter of the photosphere? This last supposition is an unlikely one, and the probability is that in a spot we view matter of a lower temperature than the photosphere.

30. Presuming this to be the case, it appears to imply one of three things.

(1.) Either the general body of the sun at the level of the bottom of a spot is of a lower temperature than the photosphere; (2.) Or the lower temperature is produced by some chemical or molecular process which takes place when a spot is formed; (3.) Or it is produced by matter coming from a colder region. The first of these suppositions will not be generally received unless we are fairly driven to accept it.

The second hypothesis has already been started to account for the lower temperature of a spot; but we think that, according to the laws by which we should be guided in receiving or rejecting an explanation in a case of this nature, this idea ought to be rejected.

No doubt, if we knew of a case of the production of low temperature, and had at the same time an independent proof of some chemical or molecular process, such as evaporation, it would be quite allowable for us to associate the chemical or molecular process with the production of cold as at any rate the most likely hypothesis; but we do not advance in our explanation of the low temperature by attributing it to an imaginary process of the existence of which we have no proof, and which is equally mysterious with the phenomenon for which it is supposed to account. Rather let us see if this reduction of temperature can be explained by any other phenomenon of the exist ence of which we have independent evidence. This leads us to consider the third hypothesis, which supposes that the reduction is produced by matter coming from a colder region. Now, in the first place, we have such a region in the atmosphere above the photosphere, which (Art. 19) we have shown to be of a lower temperature than the photosphere itself. Again, the observations of Chacornac and Lockyer on the behavior of the matter surrounding a spot appear to suggest the existence of a downward current, which is therefore a current from the colder regions above.* On the other hand, the proper motion of spots observed by Carrington is in favor of this hypothesis, since a current coming from a region of greater to a region of less absolute velocity of rotation would be carried on forward, and most so nearest the equator; and this is precisely the motion of

Does not the observation by Lockyer of the facula "giving out" appear also to indicate that the lower regions of a spot are in reality hotter than the surface, leaving the inferior luminosity to be accounted for by the downrush of a cold atmosphere from above?

spots observed by Carrington. Again, we have seen (Art. 26) that the faculæ fall behind; so that we may imagine two currents to be engaged in the formation of a spot,-the one an ascending current carrying the hot matter behind, the other a descending current carrying the cold matter forward. One advantage of this explanation is that all the gradations of darkness, from the facula to the central umbra, are thus supposed to be due to the same cause-namely, the presence to a greater or less extent of a comparatively cold absorbing atmosphere.

31. It is but just to ourselves and to M. Faye, to mention that both have imagined the phenomenon of sun-spots to be due to ascending and descending currents. M. Faye's hypothesis was published a little before ours; but we shall readily be believed when we state that an idea of this kind presided over the construction of table III, in which we have proved that the faculæ are, on an average, to the left of their accompanying spots. It was not, however, until a short time before the publication of the abstract of this paper by the Royal Society, that, by discussing the subject together, we had matured our views so far as to connect the descending current, not only with Carrington's proper motion, but also with the presumed lower temperature of a spot. In this last respect our hypothesis differs entirely from that of M. Faye, who does not imagine that the inferior luminosity of a spot indicates the presence of matter at a lower temperature than the photosphere.

32. In conclusion, we would venture to suggest that if the photosphere of the sun be the plane of condensation of gaseous matter, this plane may be found to be subject to periodical elevations and depressions in the solar atmosphere. It may be that at the epoch of minimum spot-frequency this plane is uplifted very high in the solar atmosphere, so that there is comparatively little cold absorbing atmosphere above it, and therefore great difficulty in forming a spot. If this were the case we might expect a less atmospheric effect or gradation of luminosity from the center to the circumference at the epoch of minimum than that of maximum spot-frequency. Perhaps on some future occasion we may be able to produce evidence of this, and even of the unequal atmospherie effect of the two limbs of the sun at the same time; but in the meantime we shall content ourselves with suggesting this to the observers of our luminary as a simple inquiry that may possibly prove productive.

33. We are especially desirous of bringing under the early notice of the scientific world the accumulation of observations we are making, in order that others may put forth their own conjectures in elucidation of solar physics. In venturing the opinions we have stated, we do so with some reserve, and with the conviction that possibly they may hereafter require modifications.

ART. XX.-On the Subterranean Sources of the Waters of the Great Lakes; by GEORGE A. SHUFELDT.

IF we take down the map of North America, and follow around the borders of our chain of Great Lakes, we find that the tributaries for supplying the mighty torrent of water which pours in immense volumes over the Falls at Niagara, and thence through the St. Lawrence to the sea, are few in number and insignificant in effect. Lake Superior, the largest body of fresh water in the world, has an area of 32,000 square miles and a mean depth of one thousand feet. There are a few small streams, none worthy of the name of rivers, which find their outlet in this lake-the St. Louis and Ontonagon are the largest of these; but there is probably not water enough discharged into the lake to make up for the atmospheric absorption and evaporation. The entire State of Wisconsin, even from the very borders of Lake Superior, is drained by rivers, which flow into, and are tributaries of the Mississippi. These are, in chief, the Wisconsin River, the Black, Chippewa, Fox and Rock Rivers, the waters of which all flow southward, to the Gulf of Mexico. The whole State of Minnesota with its thousands of lakes and streams may be called the mother of the Father of Waters-for all of her waters which do not gather into the great Red River of the North are discharged into the Mississippi, and do not contribute to keep up the supply of Lake Superior; and on the northern shore of the lake, in the British possessions, there are no rivers which flow in this direction. Here the current is the other way and the streame find their way to Hudson's Bay and other more northerly seas. The outlet of Lake Superior is the River St. Mary's-a stream of considerable magnitude which discharges the surplus waters of the Lake in the direction of Lake Huron. Lake Superior is 627 feet above the sea level.

If we examine the surroundings of Lake Michigan we shall find the evidences of this theory still more striking. This lake has an area of 22,400 square miles, and a mean depth of 900 feet. It is above the sea level 578 feet or forty-nine feet below Lake Superior. It is also an immense body of water, whose sole apparent sources of supply are found in a few small streams which flow into it from the State of Michigan.

The largest of these are the Grand and Manistee rivers; from Wisconsin there is only one small stream, the Milwaukee river at Milwaukee. From Illinois there is only the Chicago river, a sluggish stream without a current; and indeed there is, at only ten miles distance from the banks of the lake south and west,