This last experiment therefore confirms the preceding. Hence it cannot be the induction-current opposite in direction to the discharge which most readily traverses the spark, but must be the other.

As it is now certain that the induction-current which has the same direction as the discharge most readily traverses the spark, another conclusion can be drawn from the preceding experiments; for they show that the diminution in the deflection of the magnetic needle which the induction-current causes is greater when that current traverses the spark from the disk to the point than when its direction is opposite. An induction-current which traverses a spark does so most readily when it can go from the disk to the point. This result, which holds for the case in which the spark passes in a space filled with air, is essentially the same as that which Professor Riess found for the spark in rarefied air.

6. We now proceed to the case in which the electrical discharge traverses a spiral, and thereby produces induction in an adjacent spiral. When the latter spiral is connected with a galvanometer and its ends are in metallic connexion, no deflection is obtained, because the two induction-currents are equal and in opposite directions. If, on the contrary, the spiral is opened so wide that formation of sparks takes place, the magnetic needle makes a deflection which indicates that the inducing current is in the same direction as the discharge-current. When the spark is formed in a space containing air, this holds, under ordinary circumstances, whatever be the shape of the poles. We really have not less than four currents-that is, two induction-currents and two disjunction-currents. When the galvanometer is inserted in the conduction and not provided with a suitable bridge, the system of currents is still more complex. Both inductioncurrents have the same electromotive force; and if the spark opposed the same resistance to each, there would be no action upon the galvanometer. As regards the disjunction-currents, their electromotive forces can by no means be equal. first induction-current, or that which is in the opposite direction to the discharge, must break through a dense layer of air; and as this cannot occur without a more considerable tension of the electricity, a powerful disintegration of the polar surfaces thereby ensues. The second induction-current, or that which goes in the same direction as the discharge, instantaneously follows the first, strikes therefore, in the spark, air already rarefied, and the disintegration is less. On this account the first induction-current must produce the most powerful disjunctioncurrent. This latter current, which goes in the same direction as the second induction-current, produces the deflection of the magnetic needle. Hence the capacity of the first induction-cur

The

rent to produce the strongest disjunction-current need not be ascribed to any special property of it, as it is sufficiently explained by the fact that the formation of sparks commences with this current. When the poles are in an enclosed space, from which the air can be exhausted, the electromotive force of the disjunction diminishes in proportion as the air is rarefied. Finally the electromotive force of induction, which does not depend on the density of the layer of air traversed by the spark, begins to be greater than the former, and then the deflection of the magnetic needle mainly depends upon the induction

currents.

Professor Riess has shown, by means of the electric valve which he has devised, that when this is inserted in the path of an induction-current, the following relations take place when the density of the air and the position of the valve are altered. When the spark is formed under a pressure of one atmosphere, there is obtained on a galvanometer inserted in the circuit a deflection in the same direction as that which would be obtained with the second induction-current. Here, as regards the direction of the deflection, it is immaterial whether the second induction-current goes from the disk to the point, or vice versa. When the current in question goes from the disk to the point and the air is gradually exhausted from the valve, the deflections of the magnetic needle are always in the same direction, but their magnitude gradually diminishes at first, increasing again on subsequent rarefaction. When, on the contrary, the valve is so applied that the second induction-current goes from the point to the disk, the deflection diminishes more rapidly with the rarefaction, and afterwards changes to a deflection towards the opposite side, which increases when the rarefaction is increased.

These details could not well have been sufficiently explained before the discovery of disjunction-currents; but now the explanation follows spontaneously. The deflection obtained when the valve was full of air did not arise, as has been hitherto assumed, from the second induction-current, but from the disjunction-current, which is caused by the first induction-current. When the air is rarefied, the disjunction-current becomes feebler, and the induction-currents begin to have more and more effect; at last they determine the direction of the deflection. Now, from the results of experiments 27 and 28, we know that the inductioncurrent can traverse the spark most readily when it goes from the disk to the point. If, therefore, the valve is so applied that the second induction-current goes from the disk to the point, the direction of the deflection must remain unaltered when the air is *Abhandlungen über die Lehre von der Reibungs-Electricität. Berlin, 1867, p. 316. Pogg. Ann, vol. cxx. p. 513.

gradually rarefied. But the deflection is not caused during the whole time by the same current: in the unrarefied air the direction of the deflection is chiefly determined by the disjunctioncurrent, and in the rarefied air by the second induction-current. If, on the contrary, the valve is applied so that the first induction-current goes from the disk to the point, this current acquires the upper hand and determines the direction of the discharge when the air is rarefied; hence in this case the deflection must alter its direction when the air is gradually rarefied. In the space filled with air the disjunction-current has the upper hand; in rarefied air the first induction-current is the more powerful.

By the foregoing investigations we have obtained a simple means of experimentally proving whether a given deflection of the magnetic needle is caused by a disjunction- or by an induction-current. Experiments 22 to 24 show that when the spark is formed between the disk and the point, the disjunctioncurrent is most powerful when the discharge goes from the disk to the point, or, what is the same thing, when the disjunctioncurrent goes from the point to the disk. Experiments 27 and 28, on the contrary, have shown that when an inductioncurrent produces a deflection, this is greatest when the induction-current goes from the disk to the point. If, therefore, the current which produces the deflection first goes from the disk to the point, and thereupon by reversing the valve a greater deflection is obtained, it is a case of a disjunction-current; but if the deflection is smaller when the valve is reversed, an induction-current is the cause. This holds, without exception, when the deflection is produced either by a disjunction- or by an induction-current only. When both currents act simultaneously and to the same extent, this rule, as may be easily seen, may, under certain circumstances, be misleading.

M. Riess has already found that, when the valve was full of air, the greatest deflection was obtained when the second induction-current (which, in his opinion, determines the direction of the deflection) went from the point to the disk. This, however, as we have seen, is a proof that the deflection was caused by a disjunction-current.

The subsequent series of experiments confirm this observation, and prove afresh that the deflection in the case in question was caused by a disjunction-current. The expression "Disk -positive," signifies that the second induction-current went from the disk to the point; and the expression the "Disk negative" signifies the contrary.

We can hereby explain the peculiarity that, in that position of the valve which gives the smallest deflection in a space filled with air, the deflection remains unaltered towards the same side when the air in the valve is rarefied.

On a superficial consideration, it may seem absurd to suppose that the disjunction-current can produce upon the magnetic needle an action many times as powerful as the discharge by which it is caused. It might be thought that the direct action of the discharge upon the magnetic needle must be just as great as when this current first produces a disjunction-current which subsequently exerts magnetic action. Yet it may easily be seen that this discrepancy is only apparent. That electricity consists of motion is indubitable; but this presupposes something which moves, whether it is the smallest particles of a body, the æther, or any other body. Now, if the mass set in motion in the electrical discharge be called M, and its velocity V, MV2 is its vis viva in the discharge. If, in like manner, m denotes the mass in motion in the disjunction-current, and v its velocity, mv2 is the vis viva of the disjunction-current. This latter quantity cannot be greater than the former, but smaller; for the entire vis viva of the discharge never passes to the disjunctioncurrent. If, now, the deflection of the magnetic needle were proportional to the vis viva of the current which acts upon it, the deflection which the disjunction-current causes could not possibly be greater than that which the discharge could directly produce; but the action upon the magnetic needle is not proportional to the vis viva, but to the intensity; that is, proportional to mv ; and this quantity may readily be many times as great as MV, although my must always be less than, or at most equal to MV2. If, for instance, M=1 and V=100, MV2=10000; if m=10000 and v=1, MV2=mv2, but mv=100MV. In the electrical discharge the mass moved is inconsiderable, but its velocity is large; in the disjunction-current the reverse is the case. By the mechanical work which the discharge performs in the spark, one of these forms of motion is changed into the other.

I will remark, in conclusion, that in my opinion it would be desirable to revise the electrical investigations which were instituted before the discovery of the electrical disjunction-current, and in which electrical sparks and an enclosed circuit occurred. However trustworthy the observations, the explanations could scarcely be either correct or complete when this mode of development of electricity was unknown.

IV. On a possible Cause of the Bright Line observed by M.
Angström in the Spectrum of the Aurora Borealis. By A. S.
DAVIS, B.A., Mathematical Master, Leeds Grammar School.

To the Editors of the Philosophical Magazine and Journal.
GENTLEMEN,

IN

N a letter upon "A Theory of Nebula and Comets," published in your Magazine for last month, I endeavoured to show that there are grounds for believing that the solid or liquid bodies composing a meteoric band are surrounded by very rare and extensive atmospheres. I assumed that the meteoric bodies are so numerous and the atmosphere of each so extensive, that in the neighbourhood of the sun, and as far from it as the tails of comets are formed, the atmospheres of the different bodies encroach upon one another, and so form a continuous envelope of gaseous matter about the sun. As the tails of comets are known to extend to a much greater distance from the sun than the distance of the earth, we must admit that the earth is moving through this gaseous envelope. Though the matter which forms it is exceedingly rare, yet it must be much condensed in the neighbourhood of a large attracting body like the earth. The question then arises, is there any evidence of the existence of such gaseous matter in our atmosphere ?

I think that the spectrum of the aurora borealis indicates the existence of this matter in the higher regions of the atmosphere. The spectrum of an aurora observed by M. Angström consists mainly of one bright line not belonging to any known terrestrial substance, besides several very faint atmospheric lines and some faint bands of light. This proves that there exists in the upper regions of the atmosphere a kind of matter not known to exist in an appreciable quantity in the lower strata. Moreover, M. Angström has found the same bright line in the spectrum of the zodiacal light, which shows that this matter is of the same kind as that which exists in the sun's envelope.

From the great superiority in the brightness of this line in the auroral spectrum, compared with the atmospheric lines, we might be led to suppose that the matter to which it is due either exists in larger quantities than the elements of the atmosphere in those regions from which the light comes, or else that the electrical currents render it more luminous than the other matter present.

This, however, is not necessarily the case; for the light due * Poggendorff's Annalen, May 1869, and Phil. Mag. September 1869. Phil. Mag. S. 4. Vol. 40. No. 264. July 1870.

D