It is plain that, if the values of x calculated by (I.) and (III.) agree better than those of y calculated by (II.) and (IV.), the change must be due to resistance.

The battery consisted of twelve similar Daniell's cells of large surface. They were connected with a commutator, by which, with a single movement, they could be connected up either in series or six and six parallel. In this way, by using the same elements in each measurement, the effect of any slight accidental difference in their electromotive forces would be, to a great extent, neutralized. No difference could be observed by the discharge of an accumulator; and therefore I assume they were in every respect equal. The currents were

measured by means of a reflecting galvanometer.

A selenium plate with six elements in the dark gave a current c=0.498 microweber. When the battery was doubled and the constant resistance r1 added, the current observed was c=0.508 microweber.

Diffused daylight admitted to the selenium, with the six double surface elements the current increased to d=0.860 microweber; with the twelve elements in series and resistance r1 the current was c2=0.643 microweber.

Putting E=6.72 volts, the values of x and y, calculated from the above data, are as follows:

Similar measurements were made with two other plates, which gave the following results :

It appears from the above, that the agreement between the calculated values representing the change, on the supposition that it is due to a decrement of resistance, is much greater than that between the values calculated on the supposition that it is due to an electromotive force set up in the selenium in the same direction as the battery current.

There is much experimental work to be done yet with selenium before any theory of its behaviour can be advanced with confidence. So far, the experiments seem to suggest the suspicion that light causes a modification of the surface tension of selenium, possibly an expansion of the crystalline surface. The superficial crystals expanding and pushing against each other, so as to improve the doubtful points of contact previously existing between them, may account for the observed increase of conductivity in the light. Such a superficial expansion would probably be occasioned by heat; and this heat on the extreme surface might account for the alteration, by light, of the potential of the selenium plate, when made up in the form of a galvanic cell, being in the same direction as the alteration of potential produced by the direct application of heat when the selenium is in the dark. It might also account for the decreasing sensibility of the selenium by continued exposure to light, the superficial heat penetrating into the interior and relieving the state of tension of the surface.

I apprehend that the superficial atoms of any body, which are bounded on one side only by similar atoms, and on the other side by the medium in which the body is immersed, may be capable of assuming vibrations of different periods to those which the atoms underneath the surface can assume. It may also be that the luminous rays striking upon the superficial molecules of selenium, impart a vibration to them of a slower period than those of the exciting waves, and which corresponds nearer to the period of the heat rays.

To return, however, to the object with which this inquiry was undertaken, viz. the production of constant resistances for measuring-purposes, it is evident that selenium is, from its peculiar nature, a very unsuitable material. In its amorphous state it is dielectric; and in its imperfectly crystalline state its character seems to be that of a dielectric more or less charged with conducting crystals.

This character it probably never entirely loses, even when crystallized as far as it can be ; and to this fact is probably due, in a great measure, its peculiar behaviour.

In the light it would of course be utterly useless for measuring-purposes, whilst in the dark the apparent resistance of its junctions with the conducting wires changes, not

only with the direction of the current, but likewise with its strength, and to some extent also with its duration. To construct a selenium resistance for exact measuring-purposes, coefficients for all these changes would have to be determined, at very considerable trouble; and when determined, the inconstancy of the material is such that they would probably soon be altogether inapplicable.

In preparing the apparatus and making the experiments, I have been greatly indebted to the efficient aid rendered me by Mr. McEniry.

LVIII. On the Edge-angle and Spread of Liquids on Solid Bodies. By G. QUINCKE.

[Concluded from p. 339.]

10. SPREAD of Liquids upon the Surface of Solid Bodies. -It is possible to form an opinion in another manner than by direct measurements, about the magnitude of the surface-tension at the boundary of a fluid and of a solid body, from the magnitude of the edge-angle which a fluid-surface forms with a solid body.

From equation (5) follows at once

if the edge-angle of the common bounding-surface of the fluids 2 and 3 with the solid body 1 is an acute angle for fluid 3.

For glass as the solid body, and water as fluid 3, and for bisulphide of carbon, chloroform, olive-oil, turpentine, petroleum, or mercury as fluid 2, this condition is fulfilled with the slight allowances already explained* for impurity of the solid surface of glass.

The water will therefore have a greater adhesion to glass than the fluids named.

Whenever free fluid-surfaces bounded by air are absent, none of the fluids 2 investigated drives the water away from the glass surface, however different in magnitude may be their capillary constants. The edge angle was only in rare cases 0°; consequently the water also usually did not drive the other fluids from the glass surface.

*Compare the researches upon flat drops and bubbles, Pogg. Ann. cxxxix. pp. 18-20, 22 (1870); upon submerged capillary tubes, ibid. pp. 42-44; upon the ascent in capillary tubes of several superposed fluids, ibid. pp. 50-52. And in Phil. Mag. April, May, and June, 1871.

That the phenomena produced by the same liquids may change with the nature of the solid substance follows from the researches of Chevreul*, in which, when air was excluded, olive-oil was driven away by water in porous porcelain, while in porous whitelead water was driven away by olive-oil.

This circumstance is therefore worthy of notice, since it has often been said that fluids with a lesser capillary constant, or tension of the free surface, drove away from solid bodies the fluids with a greater capillary constant of the free surface. If the fluids are brought into contact with the air, besides being brought into contact with one another and with the solid substance, then indeed the water is driven away by most of the fluids 2 with which the researches recited above were conducted; and I have already previously completely discussed the cause of this phenomenon. Hence it follows that the contact of air with fluids 2 and 3, which are in contact with a solid body 1, may promote or initiate the driving away of fluid 3 by fluid 2 from the surface of a solid body, when once α<α, and in consequence the sum of the surfacetensions a minimum ‡.

This remark appears to me to be of importance for the comprehension of the influence of the air, or of gases generally, upon the processes of diffusion in the nourishment of plants and animals, and in the influence upon the digestion of drinks containing carbonic acid.

If two fluids, which are mutually miscible in all proportions, come into contact with the same solid body (with glass in the present case) simultaneously, without access of air, there can be no surface-tension at the surface of contact of the two fluids, and the fluid of the lesser surface-tension a12 of the common boundary with the glass must drive away from the solid body the fluid with the greater surface-tension.

According to the figures of Tables VIII. and IX., water must therefore drive away alcohol from a surface of glass. This is in harmony with experience, since burnt clay and quartz-sand, which behave similarly to glass, deprive aqueous alcohol of water, as Wagenmann and I have found §.

According to Table IX. turpentine must drive olive-oil

*Comptes Rendus, lxiii. p. 63 (1866).

† Pogg. Ann. cxxxix. p. 58 (1870); and Phil. Mag. June 1871, pp. 466-9.

Pogg. Ann. cxxxix. p. 61 (1870); and Phil. Mag. June 1871,

p. 471.

§ Pogg. Ann. cx. p. 61 (1860). Compare also W. Schmidt, Pogg. Ann. xcix. p. 370 (1856), and Duclaux, Ann. de Chim. et Phys. (4) xxv. p. 486 (1875).

away from a glass surface; the contrary must occur according to Table VIII.

In fact the first occurs, since, as I formerly proved by the change of height of capillary ascent, in a capillary tube filled with olive-oil and dipped in turpentine, the latter drives away the olive-oil, and finally spreads upon the surface of the olive-oil.

In the case of all aqueous saline solutions which are miscible with water in all proportions, the saline solution must drive away the water from the surface of the glass the more easily the more concentrated it is, since (a) increases † with augmented concentration, and, as is shown in detail in § 9, a12 is the smaller as (a)=a, cos is greater for the free surface of the liquid.

The same must also occur in the case of different saline solutions which are miscible in all proportions, and which exercise no chemical action upon one another (as giving precipitates &c.).

Since for all saline solutions the edge-angle against glass has nearly identical values, the saline solution with the greater cohesion (a) will drive away from the surface of the glass that with lesser cohesion.

Hence it follows, further, that from a dilute saline solution, as a mixture of water of less cohesion with concentrated saline solution of a greater, there must collect on the surface of the glass a concentrated solution. The most soluble substances, which in strong concentration exhibit the greatest cohesion, must also collect specially readily on the surface of glass, or will be specially strongly absorbed by the glass surface.

The phenomena of so-called selective absorption appear to confirm this. Quartz, porcelain, &c., which have for saline solutions an edge-angle similar to that of glass (compare § 7 above), appear to absorb carbonate of potash, chloride of calcium, chloride of magnesium, &c. specially strongly, and hence, therefore, those salts which with strong concentration exhibit the greatest cohesion (a).

This is in harmony with experience in the fact that new, unused porcelain cylinders (such as are usually employed in galvanic batteries), when they have once been in contact with a dilute saline solution, retain salt absorbed even after a longprotracted steeping in water.

The influence of the molecular nature of the solid substance exhibits itself specially clearly in the case of lightly and

Pogg. Ann. cxxxix. p. 55 (1870). † Ibid. clx. p. 371 (1877).

Phil. Mag. S. 5. Vol. 5. No. 33. June 1878.

2 E