der Waals, Lord Rayleigh had written a year or two earlier: "To fix ideas, we may begin by supposing that the molecules are equal hard elastic spheres. —and then, at a later stage, shows how the theory breaks down if it is even assumed that the molecules are unequal. In the present case the same hypothesis served the same useful purpose.

It is obvious that Lord Rayleigh was not altogether satisfied in accepting the view that argon was monatomic on the basis only of the value of the ratio of the specific heats, and it seems that he was not much influenced by the fact that the periodic law gave support to this conclusion. At a later date he attempted to apply a further test, also based upon the kinetic theory of gases, according to which the rate of change of viscosity should be different in the case of a diatomic and a monatomic gas. He determined the temperature coefficients of change of viscosity of argon, helium, and other common gases, but the results threw no light upon the problem of the constitution of the inactive gases.

Dr. Johnstone Stoney who was very much interested in the subject, and made helpful suggestions in other directions, in a letter to the Chemical News in February, 1895, made an attempt to develop an explanation of the experimental value of the ratio of the specific heats, based upon the idea that the argon molecule consisted of a massive immobile atom, and a light a mobile atom, possibly of one of the elements which he suggested might fill the spaces between hydrogen and lithium. However, it is hardly worth while following up his argument. I should say amongst physicists generally that the deduction. as to the monatomicity of argon was generally accepted. A considerable number of chemists held to the view that argon was really N,, but without clear or definite reason (pp. 49, 54). Considering the position historically it must be remembered that chemistry was passing through a phase in which an attempt was being made to dispense with the atomic theory altogether, and to interpret chemical phenomena entirely by means of thermodynamics. When the movement was really in its decline in the year 1900, for it was of an ephemeral character, Professor Ostwald of Leipzig, the prophet of the movement, wrote: "For the representation of the simple and comprehensive laws to which weight and volume ratios of chemical compounds are subject, a hypothetical conception. has been in use since the time when these laws were first discovered, which affords a very convenient picture of the actual relations, and possesses therefore great value for the purpose of investigation or instruction. For this reason the hypothesis (atomic hypothesis) has been the basis of the language and modes of representation throughout the whole of chemistry, so that the results of investigation are almost exclusively communicated in that language. For

this reason alone a knowledge of the hypothesis is necessary.' The italics are mine. In such a frame of mind it must have been very difficult to attach any meaning to the value of the ratio of the specific heats at all, except to make use of the analogy of mercury, to identify the numerical values of the density with the combining weight. This school of thought has given us the word mole, introduced with a view to eliminating the idea of the molecule, and, I think, nothing more.

IV

AN UNPUBLISHED REPLY TO CRITICS

T

1895

HE paper describing the investigation entitled, "Argon, a New Constituent of the Atmosphere," was read at a meeting of the Royal Society on January 31st, 1927, and to meet the large demand for admission to the meeting it was held in the large theatre of the London University, then housed at the back of Somerset House. I was present at the meeting, with a group of juniors from University College. Ramsay gave an account of the work, and was in very good form, thoroughly enjoying himself, as he always did on such occasions. He and Lord Rayleigh were received most enthusiastically by the meeting, but much less so by the representatives of scientific bodies who spoke more or less ex officio. I am sure, however, that we of the audience were much more upset by the attitude of his scientific colleagues than was Ramsay, who had a very keen sense of humour, and could always turn the tables on a critic who was not too careful as to the remarks to which he might commit himself. He wrote in the notebook later in the week: "Thursday. Wrote notes on address. Spent mg. with Ld R. over transactions paper. Gave paper in Theatre of L. U. in afternoon to crowded audience. Ld K. in the chair. Present, Balfour, Bryce, Matthews, Playfair, and everybody chem. or phys. except Dewar (chem., phys., or neuter?). Congrats. after meeting." Lord Rayleigh appears to have expressed disappointment at the tone of the meeting, but his lucid and convincing reply to the discussion must have satisfied all reasonable critics. It certainly made it clear that remarks, such as those of Professor Armstrong, that the suggestions that the value obtained for the ratio of the specific heats pointed to the monatomicity of argon was "wildly speculative" were quite unjustified, and that the term might be applied more justly to the alternatives proposed by himself and other speakers.

There is an addendum to the original paper, in which are given details of further determinations of the density of argon, further measurements of the wave-length of sound in the gas, and comparative measurements with the argon and hydrogen thermometers, proving that the value of the gas constant for argon did not vary over a wide range of temperature. This addendum. is actually an extract from a second paper which Ramsay wrote but did not

publish, and which I recently found among his papers. In it he deals somewhat exhaustively with the various suggestions which had been put forward as to the nature of argon, and in view of the criticism of the method of determining the ratio of the specific heats, he describes an attempt to estimate the relative values for the ratios for air and for argon by another method.

This short paper is now reprinted as I found it. The figure is from the laboratory notebook. Its meaning was not clear till the unpublished paper was found.

"In a former communication it was shown, that the density of argon is approximately 20, and that the ratio of its specific heats closely approaches the theoretical ratio for a monatomic gas, and the actual ratio found by Kundt and Warburg for mercury-vapour. The conclusion for The conclusion for mercury-vapour corroborates the deduction from its chemical behaviour, namely, that it is a monatomic gas; and it appears equally justifiable to draw the same deduction in the case of argon. But, as was pointed out, if argon be a monatomic gas of density 20, its molecular weight must be 40, inasmuch as the molecular weight ascribed to the standard, hydrogen, is 2. And if, as the ratio of specific heats appeared to show, it be really a monatomic gas, then its atomic weight must be identical with its molecular weight, and must be 40.

"It is desirable, however, before coming to a final conclusion, to consider all possible hypotheses, and it may be well to state them here, so that the evidence to be adduced may be applied to one or other supposition. And it is not perhaps inconceivable that argon is a compound gas, the atoms of which are so firmly combined with each other that there is no possibility of internal motion of one atom relatively to the others. It may be a compound gas in the sense in which free hydrogen is a compound, i.e., it may consist of two or more atoms of the same element; or it may be a compound of two or more atoms of different elements. Or again, it may be a mixture of two or more elements; or of two or more compounds if such compounds are possible.

"Some of these suppositions can be easily disposed of, by a correct determination of its density; and for that reason, further determinations of the density of the gas were made, care being taken, before each determination, to circulate the gas over red-hot magnesium, copper, copper oxide, soda-lime, and phosphoric anhydride for several hours; in fact for long after all contraction had ceased. The results are in complete accordance with previous determinations of density; and for convenience of reference, the former numbers are included in the table which follows.

"The general mean is 19.901; and if Nos. (2) and (3) be rejected as possibly containing nitrogen, the mean of the remaining determinations is 19.941.

"The molecular weight may be taken as 39.9 without appreciable error. "The first supposition, that argon is a compound of two similar atoms, and that it possesses the formula A, would lead to the atomic weight 19.94. This supposition will be considered later. Were it a compound of the formula A,, the atomic weight would be 13.3, a supposition which need not be further considered.

"The second supposition, that argon is a compound of two or three atoms of known elements would be at once dismissed, were it not that it has been frequently suggested that it corresponds to an ozone form of nitrogen, N. Against this supposition, which naturally occurred to us as soon as an approximate density determination had been carried out, and long before the first notice of our discovery, the following facts must be urged.

"1. The density of such a compound must be 21.045-more than a unit higher than that found.

"2. The existence of N, is very unlikely for purely chemical reasons. The compound N,H is now known, and if N, were capable of existence, it would be the only molecule known, to which the addition of an atom of hydrogen would produce a molecule capable of existence.

"Moreover, evidence of the existence of N., i.e. N-N, has recently been obtained, by treating silver hydrazoate, AgN,, suspended in chloroform, with the theoretical amount of iodine. Very little gas is evolved, and the liquid contains a substance possessing the hydrazoic smell, and yet differing from hydrazoic acid. An unfortunate accident has prevented Mr. George McDonald, who began the research, from continuing it to a close. It is practically certain that N,, if existent, would be an explosive compound.

"3. The inert character usually attributed to nitrogen is a mistaken one. Nitrogen is really a very active element if presented under favourable conditions. to the attack of other elements. A list of the elements which combine directly with nitrogen has been given in the former memoir; and because they are rare,