as the height of the aurora. Polar light, however, has been often ob served at much less height. Farquharson, by means of observations made at two distinct places, a mile apart, found a height of 2,481 English feet; and Captain Parry saw, on one occasion, polar light produced between the place where he was stationed and a mountain only about 3,000 yards distant. The French observations made at Bossekop equally proved the existence of polar light between the place of observation and a neighboring mountain.*

Although M. Loomis and M. Bravais himself believe that observations which give so low a height to polar light are erroneous, and the result of an illusion, I cannot agree with them, and I can offer in support of my opinion the phenomenon I observed on the 18th of October, 1868, at the entrance of the Norwegian archipelago, when the whole. horizon was covered with rays, which were soon united around the magnetic pole, forming a regular crown. All the phenomena I have observed and described in regard to the illuminated edges of clouds, show very plainly that in these cases the polar light was produced in the region of clouds, and even lower. Moreover, we know by numerous observations that the number of storms accompanied by lightning and thunder diminishes considerably in proportion as we approach the polar regions, so that at 70° they no longer occur. Must we then conclude that in these regions the clouds are completely deprived of electricity? Certainly not, but only that the electrical discharges are made in some other way. I have on several occasions observed discharges, accompanied with electric light, proceeding from scattered clouds or banks of clouds such as produced true polar rays, and still more frequently I have seen the edges of clouds illuminated with a yellow light. But, however, in these high latitudes electricity is discharged not only by clouds, but also directly by damp air, as takes place in winter in the temperate zones. A great many direct observations prove the existence of slow discharges of this nature, and very remarkable confirmation is given of it by M. Angström, who on one occasion proved the presence of the yellow ray of polar light over almost the entire sky.

If it is well established that the phenomenon of polar light has its source in the electricity of the air, it follows that its appearance depends less upon terrestrial magnetism than has been hitherto supposed. This may exercise a direct action upon the discharge already produced, but cannot contribute to its production, which must depend upon certain conditions of the different strata of air. Although terrestrial magnetism has an influence upon the position of the luminous bow of polar light,

*M. A. W. Malin, intendant of the Museum of Gottenburg, relates, in a description of a journey made in 1842 in the Laplands of Sweden and Norway, that, during an excursion from Maunu to Lyngen on the night of the 16th of March, he observed, at a height of 3,000 feet, with the temperature at 40 degrees below zero, a polar light between himself and the neighboring mountains, and heard a crackling sound which accompanied it.

it is difficult to believe, with Hansteen and Bravais, that the position of this bow is determined solely by the magnetic pole.

The apex of the polar bow is rarely in the exact direction of the needle of declination. Of two hundred and twenty-six observations of the posi tion of the azimuth of the luminous polar bow, 36 out of 100 placed it 30° more to the west, 32, 10° to 20°, 7,0° to 10°, and 4, 0° to 26° to the east ; from which it is evident that the position of the bow varies over a space of from 25 to 30° and more. These variations are very great to be explained by accidental perturbations in terrestrial magnetism, particularly as the greatest deviation in magnetic declination-that is, from 60 to about 70-observed occasionally in polar regions is due to polar light itself.

We may then consider it certain that terrestrial magnetism relatively plays but a secondary part in the phenomenon of polar light; that this part essentially consists in a directive action upon the rays of this light, and in a rotary movement imparted to these rays-facts positively demonstrated by the experiments of M. de la Rive.

The formation of the auroral crown, which takes place when polar light is very intense and its rays are united around the magnetic zenith, is generally supposed to be an effect of perspective. When a certain number of polar rays, parallel to the direction of the needle of inclination, are projected to a considerable height, they ought to appear to unite around the magnetic zenith; but the aspect they present should be rather that of a lengthened point or a funnel, according as the ob server is placed on one side or in the middle of the phenomenon. In polar regions it often happens that the polar rays start from all parts of the horizon, which is the case when the observer is within the ring. If then the crown was a phenomenon of perspective, the rays should appear to unite at a rather sharp angle. Now, this is by no means the case, for they form a vault, resembling very much the cupola of a church. Although my experience is not sufficient for me to make a positive assertion, I am strongly inclined to believe that, under the influence of terrestrial magnetism, and perhaps also through the effect of the conducting power of the medium, the rays of light undergo a flexion, the result of which is to unite them really, not merely in appearance, in the upper parts of the atmosphere. In proof I may cite the polar light I observed on the 18th of October, under the 17th degree of latitude. The rays starting from all parts of the horizon formed an immense ring, and united around the magnetic zenith, where the crown was formed in a perfectly regular manner, presenting the appearance of a flattened cupola. However, the experiments of M. de la Rive, which have demonstrated the influence of magnetism upon electric light, under circumstances almost identical with those presented by polar light, furnish no proof that the rays of this light are really united under this influence.

Polar light, considered as an electric discharge, produces the following results:*

1. An electric current, produced by the discharge itself, which takes place slowly.

2. Rays of light consisting of an infinite number of sparks, each spark giving rise to two currents of induction, proceeding in opposite directions.

3. An electric current, proceeding in an opposite direction from that of the charge, and originating in the electro-motive force, discovered by M. Edlund in the electric spark.

In order to develop these currents, a closed circuit is necessary. It is true that in the phenomenon of polar light this does not properly exist, but in that case the earth on one hand, and the rarefied air of the upper regions of the atmosphere on the other, are immense reservoirs of electricity, which produce the same effect as if the circuit was closed. According to the theory of M. de la Rive, the positive electricity of the air, discharging itself into the ground, produces a current I will call principal; this current is re-enforced by one of the currents of induction, which M. Edlund has shown accompany the production of the electric spark; that is to say, the one which, going in the opposite direction from the charge, can alone acquire a certain degree of intensity. But this principal current thus re-enforced is counterbalanced in part by the one which has the contrary direction, and which produces the electro-motive force of the spark. We see by the observations made with telegraph wires during the appearance of polar light that it is sometimes the one, sometimes the other of these two currents which gains possession of the wire, the first being generally predominant, since it has been observed that the current given by telegraphic wires is more frequently directed from the north to the south than from the south to the north.

Under the circumstances which accompany the production of polar light, the latter contains in itself all the conditions necessary in order that magnetism may act upon it; for a ray of this light constitutes a current, flexible throughout, and consequently obedient to the law discovered by Plucker, according to which such a current necessarily takes the form of a magnetic curve. It is here we should look for the cause of the formation of the crown, taking at the same time into consideration the variations in conductibility of the rarefied air of the upper regions of the atmosphere.

When an arch of polar light appears, we consider that it forms part of a radiant ring, whose center coincides very nearly with the magnetic pole. The rays of this ring are parallel to the direction of the

The author precedes this part of his work with the description of an experiment, in which he tried, but unsuccessfully, the action of a magnet upon a series of discharges produced by Holtz's machine under certain conditions. We omit this description, as it would be unintelligible without a figure to illustrate it, and it is not of great importance, since the results of the experiment were negative.

needle of inclination, and ought consequently to diverge from all sides; a circumstance unfavorable to their union in perspective at the magnetic zenith. But it is not proved that the radiation has always for center the magnetic pole; it may very well have another central point, as was the case during the observation I made on the 18th of October. On that occasion, in fact, the ring could not have been completely visible if it had had for center the magnetic pole. Besides, this variation of position is more in conformity with the manner in which the electric charges operate, and with the peculiarities observed in the local appearances of polar light.*

Returning to the lights observed around the elevated peaks of the Spitzbergen, I ought to say that this phenomenon has been noticed before. The learned philologist of Finland, Cashen, witnessed it in his journeys to Siberia, and his description of it exactly accords with what I have myself observed. Similar light has been seen in South America above the peaks of the Cordilleras, and in several other localities mentioned by M. Delleman. The Archives des Sciences physiques et naturelles (tome xxxi, p. 15) contain an article by M. H. de Saussure, in which are described a great number of phenomena, belonging, without doubt, to the same category. But it is in the arctic regions, above all others, that we find circumstances most favorable to observations of this kind, and it is much to be desired that future expeditions will undertake them.

If we seek for the reason why the clouds of the upper latitudes discharge themselves under the form of polar light, and not that of thunder and lightning, we find it in the permanent humidity of the air. The hygrometric observations, made during the expedition of the Sophie, show that the air is constantly saturated with aqueous vapors, which condense frequently into clouds, more rarely into rain. It is clear that this stratum of humidity, a good conductor of electricity, determines a slow discharge. If between the poles of an electrical machine, not sufficiently near together to produce a discharge, we project, by means of an atomizer, some water in spray, we see the discharge under the form of brilliant rays. It is the same in a glass cylinder in which the air has been rarified by several strokes of a piston, a rarefaction sufficient to produce a mist. The discharge, which at first appears as a

*The author remarks here that the appearance of polar light is always accompanied by a dark segment, through which the stars are visible. He mentions in this connection the experiment of M. de la Rive, who, in transmitting the discharge of an induction coil through very rarefied air, proved the existence of a very remarkable dark band near the negative electrode. The author cites an experiment made by himself in the presence of M. Edlund, during which he obtained by means of an electrophorus, in a tube of very rarefied air, a bluish light, followed by a dark band strongly marked around the negative electrode, and a kind of crown of rays around the positive electrode. In the phenomenon of polar light, the earth constitutes the negative electrode, the rarefied air of the higher regions of the atmosphere the positive electrode, and the dark segment bears a strong resemblance to the dark band of the preceding experiment.

spark, is gradually transformed into a luminous current, exhibiting the ordinary colors observed in polar light.

The more the relative degree of the humidity of the air increases an augmentation, which at the surface of the earth proceeds from the equator to the poles, the more easily the electric discharge acts under the form of polar light. But there may be a limit beyond which the humidity is so great that the discharge takes place without the accompaniment of light. Such would seem to be indicated by the table prepared by M. Loomis, of the geographical extension of polar light; according to which there must be a zone, comprised between the 68th and 76th degree of latitude in Europe, and between the 50th and 64th in America, in which the greatest number of phenomena of polar light are produced. This very interesting peculiarity has been proved by observations made during our expedition; at Spitzbergen the polar light always appeared to the south, while at a lower latitude, the 69th degree, it appeared at the zenith or to the north.

I now return to the question of spectral analysis; and for the moment notice only the mobility, so marked, of the light of the yellow ray, which seems to indicate a discontinuous luminous source which is evidently formed by an infinite number of sparks, succeeding each other in rapid succession.

As to the crackling or rattling noise which accompanies the appearance of polar light, I cannot say anything positively, since, on the occasions when I made my observations, the combined noise of the sea and the wind was of such a nature as to drown the faint crepitation of an electric discharge. It is very probable that such a noise can be heard under certain circumstances, for instance when the discharge takes place at a minimum height, and also when it is made between small particles. of ice, which produce longer, and consequently stronger, sparks than those formed between particles of water. As these circumstances which are necessary for the production of this sound rarely occur, we can understand why observers do not agree in regard to the fact of its exist

ence.

Some remarks upon the memoir of M. Lemström, by Professor De la Rive.

I find in the observations made by M. Lemström, in the polar regions, such a complete confirmation of the views I have expressed on several occasions in regard to the cause and explanation of polar auroras, that I cannot refrain from noticing very briefly some of the points upon which observation and theory completely agree. I have generally found myself in accord with observers, whether such as Parry, Franklin and Ross, or Bravais and Martins; it is rather between the theorists and myself that there has been occasionally some difference. I ought to thank M. Lemström for the pains he has taken, on every occasion, to mention my experiments, and the consequences I have deduced from them, the accuracy of which he has confirmed.