retains its specific absorbent power when in combination. Two of the above-described cases, however, are slightly anomalousthe chromate of chromium, and the double iodide of platinum and potassium; and these are not the only, nor indeed the greatest exceptions, for the ferric ferrocyanide dissolved in oxalic acid transmits blue rays in great abundance, which are absorbed both by ordinary ferrocyanides and ferric salts. It must be borne in mind that the statement, "all the compounds of a particular base or acid, when dissolved in water, have the same effect on the rays of light," is not universally true, even when a coloured base is combined with a colourless acid; and hence we might anticipate, what actual observation has shown to be the case, that variations would sometimes occur when both the acid and base are coloured. It may be therefore laid down as a general, though not an invariable law, that when an acid and a base combine, each of which has a different influence on the rays of light, a solution of the resulting salt will transmit only those rays which are not absorbed by either, or in other words, which are transmitted by both.

XLIX. On the Effect of Heat on the Colour of Salts in Solution. By J. H. GLADSTONE, Ph.D., F.R.S. &c.*

[With a Plate.]

S a general rule, the solution of a salt has the same power of absorbing or transmitting the rays of light at all temperatures. I am not acquainted with any instance of a dissolved colourless salt which assumes a colour when the solution is either heated or cooled; nor does the converse scem ever to occur,a salt coloured at the ordinary temperature, which loses that colour when heat is applied. Nevertheless it is not rare to find coloured salts, which, when dissolved in water, vary in shade or in tint according to the temperature.

In some instances, heating the solution seems merely to intensify the colour. This is the case with the following red, orange, yellow, and green salts:

Meconate of iron-red.

Terbromide of gold-red.
Red nitrate of cerium.

Bichromate of potash-orange.

Ferrocyanide of potassium-yellow.

Molybdous chloride-green.

More frequently a change takes place in the character as well as in the intensity of the colour when the solution is heated. The following cases have been observed.

* Communicated by the Author.

Bichloride of platinum, while it becomes more intense in colour, assumes also a redder tint.

Protochloride of platinum, held in solution by hydrochloric acid, changes also in the same way.

Bichloride of palladium acts precisely as the platinum salt does under the influence of heat.

Ferridcyanide of potassium gives a greenish solution, which, when heated, alters in colour, and if not too dilute, assumes a distinctly red appearance.

Polysulphide of potassium passes from yellow to a most intense red.

Sesquichloride of iron passes from orange to a deep and almost pure red.

Chloride of nickel passes from a bluish to a yellowish green. Iodide of nickel, when dissolved in a little water, gives a clear green solution, which, on the application of heat, becomes of a nondescript shade, that appears distinctly red by gas-light.

Chloride of copper gives a green saturated solution, which on the addition of more water becomes blue. If this blue solution be heated (unless too dilute), the green colour is restored. On cooling, it again becomes blue.

Bromide of copper behaves in every respect like the chloride. Sulphocyanide of cobalt in a minimum of water, gives a magnificent bluish purple colour, but on dilution it changes to the ordinary pink tint of cobalt salts in solution. If this be heated, provided it is not too dilute, it will reassume the purple hue.

Chloride of cobalt dissolves in water always of a pink, and in absolute alcohol always of a blue colour, while in mixtures of alcohol and water it will assume an immediate tint. By arranging properly the proportions of the two solvents, a liquid may be obtained which will show all the changes of an aqueous solution of the sulphocyanide, passing from pink through purple to blue when it is heated, and conversely, from blue to pink when it is cooled.

In all these cases it is to be understood that the change of colour lasts only as long as the heat continues. No permanent chemical change is effected; and the original colour of the solution returns in every instance as it cools.

A glance at the above observations will suffice to show that where the colour is not materially altered in character, it invariably becomes more intense when heated, that is to say, fewer rays are transmitted; and when the light is analysed by a prism, it is found that not in these only, but in every one of the instances, rays are absorbed by the hot solution which were transmitted by the same solution when cold.

Yet a distinction must be made between two classes of action, both of which influence the above observations, and influence

them in the same direction. In the last seven instances it is more than probable that a temporary chemical change is effected in the liquid. In the preceding paper, I have shown that the chloride and bromide of copper, the chloride and iodide of nickel, the sesquichloride of iron, and one or two other analogous salts, give a saturated aqueous solution of a different colour to what is presented by the same solution when more water is added; and that this change is due to the fact, that in the saturated solution both the halogen and the metal exert their own absorbent power on the rays during their passage through the liquid, while in the dilute solution the metal alone influences the transmission of light. That this phænomenon results from some difference of arrangement among the elements of the salt and water, scarcely admits of a doubt; yet what that difference of arrangement may be is not so easily determined. This chemical change it is that is affected by temperature, an increase of heat having the opposite effect to an addition of water, and a diminution of heat having the same effect as dilution. Thus the light transmitted by a dilute solution of chloride of copper at the ordinary temperature, when analysed by a prism, shows the spectrum represented in Plate II. fig. 9, but on heating the solution, an absorption due to chlorine manifests itself, and the coloured band is reduced to the dimensions of fig. 21 and similarly, the bromide of copper, if dilute and cold, presents the ordinary prismatic appearance of copper salts, fig. 9; but if the same liquid be heated, it assumes the green colour and the modified spectrum of the saturated solution, fig. 17, in which the bromine exerts its absorbent power. In each case the influence of the halogen disappears as soon as the heat is withdrawn; and that this is not confined to the range of temperature above what we designate as ordinary, was proved by exposing a green solution of chloride of copper to a freezing mixture, when it assumed a distinctly bluish tint.

The case of the cobalt salts is evidently analogous: the dilute sulphocyanide of cobalt in water, or the chloride in aqueous alcohol, gives a prismatic appearance similar to, but not identical with, that represented in fig. 12; but when heated, additional dark spaces show themselves; the more refrangible red and less refrangible orange rays are wholly absorbed, and the more refrangible portion of the orange is allowed to penetrate but to a short distance, while a perfectly black line shows itself coincident with D, but somewhat broader. In fact the hot cobalt salt gives precisely the same prismatic appearance as that given by smalt-blue glass, or by an alcoholic solution of cobalt, as figured in the plate annexed to my paper "On the Use of the Prism in Qualitative Analysis," in the Quarterly Journal of the Chemical Society, April 1857,

No chemical change, however, appears to be concerned in the alteration of colour produced by heat in the other cases mentioned above. The elevation of temperature seems merely to heighten the absorbent power of the dissolved salt, so that the light absorbed by a certain quantity of the heated solution is the same as would have been absorbed by a larger quantity of the same solution, if cold. This will fully account for the changes, not merely in such salts as meconate of iron, where an increased intensity of colour is all that is observed, but for those changes which involve the character as well as the depth of the colour. These latter substances are in fact more or less dichromatic, that is, they present different colours according to the quantity of salt which the light has traversed before it reaches the eye; and the reason of this will be apparent on glancing at the spectrum of bichloride of platinum represented in fig. 20. It will be there seen that a thin stratum of the salt transmits all the rays from the extreme red to the fixed line F, and a little blue and violet beyond; but that as the thickness increases, the rays about band afterwards about E are absorbed. The general impression conveyed by the rays that traverse a thin stratum is yellow; but when the green rays are absorbed, it changes naturally to orange, becoming more and more red as the stratum increases. Now the effect of heat upon a thin stratum, or a weak solution, is solely to produce the same amount of absorption as would be produced at the ordinary temperature by a thicker stratum, or a stronger solution.

In reference, however, to the change that ensues when a solution of polysulphide of potassium is heated, I doubt whether any increased thickness of the same liquid would give so intense a red. Can it be of the same nature as the modification of colour that takes place in melted sulphur at a far higher temperature? This is rendered more probable by the fact, that the yellow colour of the potassium salt in solution is due to the sulphur; yet on the other hand, sulphur dissolved in naphtha shows no indication of redness when the liquid is boiled.

It is scarcely necessary to remind either physicists or chemists of the observation made by Sir David Brewster, that the absorption bands of peroxide of nitrogen are increased by heating the gas; or of the observations of Schoenbein and others, that several solid substances, such as oxide of zinc, or gallate of iron, absorb, when heated, rays that they transmit or reflect when cold. The above observations on dissolved salts give, therefore, results which are perfectly in harmony with the little that was known before about the effect of heat on the chromatic phænomena presented by other pure chemical substances.

L. A Demonstration of Sir W. R. Hamilton's Theorem of the Isochronism of the Circular Hodograph. By ARTHUR CAYLEY, Esq.*

IMAGI
ΙΜΑ

MAGINE a body moving in plano under the action of a central force, and let h denote, as usual, the double of the area described in a unit of time; let P be any point of the orbit, then measuring off on the perpendicular let fall from the centre of force O on the tangent at P to the orbit, a distance OQ equal or proportional to h into the reciprocal of the perpendicular on the tangent, the locus of Q is the hodograph, and the points P, Q are corresponding points of the orbit and hodograph.

It is easy to see that the hodograph is the polar reciprocal of the orbit with respect to a circle having O for its centre, and having its radius equal or proportional to h. And it follows. at once that Q is the pole with respect to this circle of the tangent at P to the orbit.

In the particular case where the force varies inversely as the square of the distance, the hodograph is a circle. And if we consider two elliptic orbits described about the same centre, under the action of the same central force, and such that the major axes are equal, then (as will be presently seen) the common chord or radical axis of the two hodographs passes through the centre of force.

Imagine an orthotomic circle of the two hodographs (the centre of this circle is of course on the common chord or radical axis of the two hodographs), and consider the arcs intercepted on the two hodographs respectively by the orthotomic circle; then the theorem of the isochronism of the circular hodograph is as follows, viz. the times of hodographic description of the intercepted arcs are equal; in other words, the times of description in the orbits, of the arcs which correspond to the intercepted arcs of the hodographs, are equal. It was remarked by Sir W. R. Hamilton, that the theorem is in fact equivalent to Lambert's theorem, that the time depends only on the chord of the described arc and the sum of the two radius vectors. And this remark suggests a mode of investigation of the theorem. Consider the intercepted arc of one of the hodographs: the tangents to the hodograph at the extremities of this arc are radii of the orthotomic circle; i. e. the corresponding arc of the orbit is the arc cut off by the polar (in respect to the directrix circle by which the hodograph is determined) of the centre of the orthotomic circle; the portion of this polar intercepted by the orbit is the elliptic chord, and this elliptic chord and the sum of the radius vectors at the two extremities of the elliptic chord, * Communicated by the Author.