Nature, March 12 1891

Surface Tension

I shall be obliged if you can find space for the accompanying translation of an interesting letter which I have received from a German lady, who with very homely appliances has arrived at valuable results respecting the behaviour of contaminated water surfaces. The earlier part of Miss Pockel's letter cover nearly the same ground as some of my own recent work, and in the main harmonize with it. The later sections seem to me very suggestive, raising, if they do not fully answer, many important questions. I hope soon to find opportunity for repeating some of Miss Pockels' experiments. --- RAYLEIGH, March 2.

Brunswick, January 10.

My lord, Will you kindly excuse my venturing to trouble you with a German letter on a scientific subject? Having heard of the fruitful researches carried on by you last year on the hitherto little understood properties of water surfaces, I thought it might interest you to know of my own observations on the subject. For various reason I am not in a position to publish them in scientific periodicals, and I therefore adopt this means of communicating to you the most important of them.

First, I will describe a simple method, which I have employed for several years, for increasing or diminishing the surface of a liquid in any proportion, by which its purity may be altered at pleasure.

A rectangular tin trough, 70 cm. long, 5 cm. wide, 2 cm. high, is filled with water to the brim, and a strip of tin about 1 1/2 cm. laid across it perpendicular to its length, so that the underside of the strip is in contact with the surface of the water, and divides it into two halves. By shifting thi partition to the right or the left, the surface on either side can be lengthened or shortened in any proportion, and the amount of the displacement may be read off on a scale held along the front of the trough.

No doubt this apparatus suffers, as I shall point out presently, from a certain imperfection, for the partition never completely shuts off the two separate surfaces from each other. If there is a great difference of tension between the two sides, a return current often breaks through between the partition and the edge of the trough (particularly at the time of shifting). The apparatus, however, answers for attaining any condition of tension which i at all possible, and in experiments with very clean surfaces there is little to be feared in the way of currents breaking through.

I always measured the surface tension in any part of the trough by the weight necessary to separate from it a small disk (6 mm. in diameter), for which I used a light balance, with unequal arms and a sliding weight

I will now put together the most important results obtained with thi apparatus, most of which, though perhaps not all, must be known to you.

I. Behaviour of the surface tension of water.-The surface tension of a strongly contaminated water surface is variable; that is, it varies with the size of the surface. The minimum of the separating weight attained by diminishing the surface is to the maximum, according to my balance, in the ratio of 52: 100.

If the surface is further extended, after the maximum tension is attained, the separating weight remains constant, as with oil, spirits of wine, and other normal liquids. It begins, however, to diminish again, directly the partition is pushed back to the point of the scale at which the increase of tension ceased.

The water surface can thus exist in two sharply contrasted conditions; the normal condition, in which the displacement of the partition makes no impression on the tension, and the anomalous condition, in which every increase or decrease alters the tension.

II. Mobility.-Upon the purity of the surface depends its mobility, and in consequence the persistence of a wave once set in motion. So long, however, as the water surface is in its anomalous condition, the damping of the waves is constant,1 and just at the degree of purity at which the tension ceases to alter the decrease of the damping begins.

If the balance is loaded with just the maximum weight which the surface tension can hold, and the normal surface is contracted till the weight break away, a measure is obtained of the relative amount of contamination by the ratio of the length of the surface before and after contraction, for the purer the surface, the smaller must be the fraction to which it is reduced before it begins to enter the anomalous state. By counting, with different relative contaminations, how often a wave excited by a small rod at the end of the trough passed along the surface adjusted to a length of 30 cm. before it ceased to be visible, I obtained approximately the following values for the number of the passages:
Relative contamination 00051015202530
Number of visible wave passages 17171717120803

The numbers of the upper row indicate the length at which the surface becomes anomalous in 30ths of its whole length; those of the second row are, a may be imagined, rather uncertain, particularly the greater ones, although they are the mean of many observations.

A perfectly clean surface, whose tension remains constant, even under the greatest contraction, can be approximately produced with the adjustable trough, by placing the partition quite at the end, and pushing it from thence to the middle. The surface on one side is thus formed entirely afresh, from the interior of the liquid.

III. Effect on a water surface of contact with solid bodies.-Every solid body, however clean, which is brought in contact with a newly formed surface, contaminates it more or less decidedly, according to the substance of which the body consists. With many substances, such as camphor or flour, thi effect is so strong that the tension of the surface is lowered to a definite value; with others (glass, metals) it is only shown by the increase of relative contamination. The contaminating current which goes out from the circumference of a bodyfor example, of a floating fragment of tinfoilis easily made visible by dusting the water with Lycopodium or flowers of sulphur. I will call it, for the sake of brevity, " the solution current."

The solution current of a body which is introduced into a perfectly clean water surface lasts until the relative contamination produced by it has attained a definite value, which is different for every substance.

Thus the solution current for wax ceases at a relative contamination of 0.55, that of tinfoil at a still smaller one, but that of camphor, not until the surface has become decidedly anomalous, and the separating weight gone down to within 0.80 of the maximum. If, on the other hand, the surface surrounding a small piece of tinfoil be restored to its previous purity, the current begin again with renewed strength, and it appears that this process may be repeated a often as desired without the solution current ever quite disappearing.

From this effect of the contact of solid bodies, it follows that a perfectly pure surface cannot be maintained for long in any vessel, since every vessel will contaminate it. Whether the air and the matter contained in it have a share in the gradual increase of relative contamination which occurs on water left standing, I know not; but the influence of gases and vapours does not appear to me important in general. The contamination by the sides of the vessel does not, however, always go so far as to diminish the tension, which remain normal, for example, in a glass of water, after four days' standing.

With a rising temperature the contamination from all substances seems to increase considerably; but I have not yet investigated this in detail.

IV. Currents between surfaces of equal tension.-Between, two normal surfaces, which are unequally contaminated by one and the same substance, a current sets in from the more to the less contaminated when the partition i removed; much weaker, indeed, than that exhibited in the anomalous condition by differences of tension, but, all the same, distinctly perceptible. With equal relative contamination by the same substance, no current of course sets in. It is otherwise when the contamination is produced by different substances.

I contaminated the surface on one side of the partition by repeated immersion of a metal plate, on the other by immersion of a glass plate, which had both been previously carefully cleaned and repeatedly immersed in fresh water surfaces. I then made the relative contamination on the two sides equal (i.e.=1/2) by pushing in the outer partitions by which the surfaces were inclosed. After the water had been dusted with Lycopodium, the middle partition was removed. I repeated this experiment eight times, with different change devised as checks.

On the removal of the partition a decided current set in each time, from the surface contaminated by glass to that contaminated by metal, and when I replaced the partition after the current had ceased, and investigated the contamination on both sides, I always found it greater on the metal than on the glass side.

Thus equal relative contamination by different substances does not indicate equality of that (osmotic?) pressure which is the cause of the current between surfaces of equal tension.

For further proof of this result I have made experiments with other substances; for example, with a floating piece of tinfoil on one side, and of wax on the other, when, after they had been acting for a long time, and then the relative contaminations had been equalized, a current resulted from the wax to the tinfoil; and again, with camphor on the one, and small pieces of wood and wax on the other side, which showed a current from the wax and wood to the camphor.

Since, therefore, the water surface assumes dissimilar qualities from contact with different substances, the conviction is forced upon me that it i these bodies themselves (glass, metal, wax, &c.) which are dissolved, though only feebly in the surface, and thereby render it capable under sufficient contraction of becoming anomalous.

V. Further observations on solution currents.-The following fact agree with this view. If a newly formed water surface be contaminated by small floating slices of wax until the latter cease to give solution currents, the relative contamination amounts to 0.55. If now another fresh surface is brought to the same relative contamination by tinfoil and a corresponding con, traction, and then a slice of wax from the first surface be introduced, it will develop a considerable solution current. This therefore depends on the substance with which the surrounding surface was previously contaminated.

Substances which are properly soluble in water, such as sugar and soda, exhibit a similar behaviour when immersed in the surface, only they continue to act in the anomalous condition.

A crystal of sugar placed in a normal but not perfectly pure surface produces a great fall of tension. If the surface be then made normal again by immersing and withdrawing strips of paper, and if this process be repeated several times, a normal surface is at last attained, which is contaminated by sugar only, and on the tension of this the sugar produces no further effect. A piece of soda held in the surface containing sugar greatly lowers the tension, and on the other hand on a surface rendered repeatedly anomalous by soda, soda acts but slightly, and sugar powerfully.

[In this experiment the sugar and soda crystals being instantly wetted, they do not really act by solution-currents, for the latter can only be produced by a dry body. The action here is an indirect one by intervention of the deeper layers.-February 26.]

VI. Behaviour of the surfaces of solutions.-The effect of soluble matter on the surface tension has absolutely nothing to do with the change which the cohesion of the water undergoes, through matter dissolved in the body of the liquid, for both sugar and soda solutions have a higher maximum tension than pure water, and yet these same substances introduced into the surface produce a fall in the separating weight.

In order to investigate the behaviour of the surfaces of solutions more closely, I introduced a saturated solution of common salt into the adjustable trough. The freshly formed surface of the solution of salt maintained it normal separating weight, (1.154 of that of water) even when most contracted, though it must necessarily have contained as much salt as the interior of the liquid. The entrance of the anomalous condition, then, does not depend on the absolute quantity of the contaminating substance contained in the surface, but when I placed some salt in contact with the normal surface of the saturated solution, it gave a solution current and lowered the tension, as in the case of pure water. I obtained similar results with a solution of sugar. From these experiments I concluded (a) that the surface layer of water can take up more of soluble substances than the internal liquid; (b) that the surface of a solution is capable of becoming anomalous under contraction, always and only, when it contains more of the dissolved substance than the interior of the liquid.

That the surface layer really possesses a higher dissolving power is further shown by the experiment, which is well known to you, in which a thin disk of camphor, so hung that it is half immersed in the cleanest possible water surface, is cut through in the course of a few hours. I will add by the way, that a newly formed surface of a saturated solution of camphor is normal according to my observations, i.e. that its tension remains nearly constant under contraction, and that small pieces of camphor floating on it still give solution streams and have slight motions. The solution stream seems in thi case to cease just when the surface begins to be anomalous.

What I have further observed regarding solutions in the surface and the like, seems to me less remarkable, and part of it still very uncertain. I therefore confine myself to these short indications, but I believe that much might be discovered in this field if it were thoroughly investigated. I thought I ought not to withhold from you these facts which I have observed, although I am not a professional physicist; and again begging you to excuse my boldness, I remain, with sincere respect,

Yours faithfully,


1 This is not quite exact. I found the number of visible passages constant = 3 in the anomalous state; but the velocity of transmission varying in some degree with the tension, the time required for the vanishing of the wave must really become a little longer when the tension i lowered. - February 26.