was then led to suspect the existence of a heavier gas in the atmosphere. He set to work to isolate this substance, and succeeded in doing so by the method of Cavendish. In the meantime Prof. Ramsay, quite independently, isolated the gas by removing the nitrogen by means of red-hot magnesium, and the two investigators then combining their labours, followed up the subject, and have given us a memoir which will go down to posterity among the greatest achievements of an age renowned for its scientific activity.

3

The case in favour of argon being an element seems to be now settled by the discovery that the molecule of the gas is monatomic, as well as by the distinctness of its electric spark spectrum. The suggestion put forward soon after the discovery was announced, that the gas was an oxide of nitrogen, must have been made in complete ignorance of the methods by which it was prepared. The possibility of its being N has been considered by the discoverers and rejected on very good grounds. Moreover, Peratoner and Oddo have been recently making some experiments in the laboratory of the University of Palermo with the object of examining the products of the electrolysis of hydrazoic acid and its salts. They obtained only ordinary nitrogen, not argon, and have come to the conclusion that the anhydride N,.N, is incapable of existence, and that no allotropic form of nitrogen is given off. It has been urged that the physical evidence in support of the monatomic nature of the argon molecule, viz., the ratio of the specific heats, is capable of another interpretation-that argon is in fact an element of such extraordinary energy that its atoms cannot be separated, but are bound together as a rigid system which transmits the vibrational energy of a sound-wave as motion of translation only. If this be the state of affairs we must look to the physicists for more light. So far as chemistry is concerned, this conception introduces an entirely new set of ideas, and raises the question of the monatomic character of the mercury molecule which is in the same category with respect to the physical evidence. It seems unreasonable to invoke a special power of atomic linkage to explain the monatomic character of argon, and to refuse such a power in the case of other monatomic molecules, like mercury or cadmium. The chemical inertness of argon has been referred also to this same power of self-combination of its atoms. If this explanation be adopted it carries with it the admission that those elements of which the atoms composing the molecule are the more easily dissociated should be the more chemically active. The reverse appears to be the case if we bear in mind Victor Meyer's researches on the dissociation of the halogens, which prove that under the influence of heat the least active element, iodine, is the most easily dissociated. On the whole, the attempts to make out that argon is polyatomic by such forced hypotheses cannot at present be considered to have been successful, and the contention of the discoverers that its molecule is monatomic must be accepted as established.

In searching for a natural source of combined argon Professor Ramsay was led to examine the gases contained in certain uranium and other minerals, and by steps which are now well known he has been able to isolate helium, a gas which was discovered by means of the spectroscope in the solar chromosphere by Professor Norman Lockyer in 1868. In his address to the British Association in 18721 the late Dr. W. B. Carpenter said :

'But when Frankland and Lockyer, seeing in the spectrum of the yellow solar prominences a certain bright line not identifiable with that of any known terrestrial flame, attribute this to a hypothetical new substance which they propose to call helium, it is obvious that their assumption rests on a far less secure foundation, until it shall have received that verification which, in the case of Mr. Crookes' researches on thallium, was afforded by the actual discovery of the new metal, whose presence had been indicated to him by a line in the spectrum not attributable to any substance then known.'

It must be as gratifying to Professor Lockyer as it is to the chemical world at large to know that helium may now be removed from the category of solar myths and enrolled among the elements of terrestrial matter. The sources, mode

1 Reports, 1872, p. lxxiv.

of isolation, and properties of this gas have been described in the papers recently published by Professor Ramsay and his colleagues. Not the least interesting fact is the occurrence of helium and argon in meteoric iron from Virginia, as announced by Professor Ramsay in July. Like argon, helium is monatomic and chemically inert so far as the present evidence goes. The conditions under which this element exists in cleveite, uraninite, and the other minerals have yet to be determined.

Taking a general survey of the results thus far obtained, it seems that two representatives of a new group of monatomic elements characterised by chemical inertness have been brought to light. Their inertness obviously interposes great difficulties in the way of their further study from the chemical side; the future development of our knowledge of these elements may be looked for from the physicist and spectroscopist. Professor Ramsay has not yet succeeded in effecting a combination between argon or helium and any of the other chemical elements. M. Moissan finds that fluorine is without action on argon. M. Berthelot claims to have brought about a combination of argon with carbon disulphide and mercury, and with the elements of benzene, . . . with the help of mercury,' under the influence of the silent electric discharge. Some experiments which I made last spring with Mr. R. J. Strutt with argon and moist acetylene submitted to the electric discharge, both silent and disruptive, gave very little hope of a combination between argon and carbon being possible by this means. The coincidence of the helium yellow line with the D, line of the solar chromosphere has been challenged, but the recent accurate measurements of the wave-length of the chromospheric line by Prof. G. E. Hale, and of the line of terrestrial helium by Profs. Runge and Paschen, leave no doubt as to their identity. Both the solar and terrestrial lines have now been shown to be double. The isolation of helium has not only furnished another link proving community of matter, and, by inference, of origin between the earth and sun, but an extension of the work by Professor Norman Lockyer, M. Deslandres, and Mr. Crookes, has resulted in the most interesting discovery that a large number of the lines in the chromospheric spectrum, as well as in certain stellar spectra, which had up to the present time found no counterparts in the spectra of terrestrial elements can now be accounted for by the spectra of gases contained with helium in these rare minerals. The question now confronts us, Are these gases members of the same monatomic inert group as argon and helium? Whether, and by what mechanism, a monatomic gas can give a complicated spectrum is a physical question of supreme interest to chemists, and I hope that a discussion of this subject with our colleagues of Section A will be held during the present meeting. That mercury is capable under different conditions of giving a series of highly complex spectra can be seen from the memoir by J. M. Eder and E. Valenta, presented to the Imperial Academy of Sciences of Vienna in July 1894. With respect to the position of argon and helium in the periodic system of chemical elements, it is, as Professor Ramsay poin's out, premature to speculate until we are quite sure that these gases are homogeneous. It is possible that they may be mixtures of monatomic gases, and in fact the spectroscope has already given an indication that they contain some constituent in common. The question whether these gases are mixtures or not presses for an immediate answer. I will venture to suggest that an attack should be made by the method of diffusion. If argon or helium were allowed to diffuse fractionally through a long porous plug into an exhausted vessel there might be some separation into gases of different densities, and showing modifications in their spectra, on the assumption that we are dealing with mixtures composed of molecules of different weights.2

1 Nature, vol. lii. p. 224.

2 The above was written before the interesting work of Profs. Runge and Paschen had become known in this country. These authors communicated papers to the Prussian Academy of Sciences on June 20 and July 11, in which they showed by the method advocated that helium from cleveite consists of two different gases (Sitzungsberichte d. k. Preuss. Akad. d. Wissensch. z. Berlin, 1895, xxx. and xxxiv.; also Nature, vol. lii. p. 520). The results were also made known by Prof. Runge at the joint meeting of Sections A and B on September 13.

1. A New View of the Genesis of Dalton's Atomic Theory, derived from Original Manuscripts. By Sir H. E. ROSCOE, F.R.S., and ARTHUR HARDEN.

A number of previously unknown manuscript volumes in Dalton's writing have been found in the library of the Manchester Literary and Philosophical Society. These consist of laboratory note-books containing the record of Dalton's practical work from the year 1802 onwards, and the notes used by him for some of the lectures delivered at the Royal Institution, London, in 1810.

The examination of these volumes has cast an unexpected light on the genesis of the atomic theory, and the relation in which that theory stands to the law of combination in multiple proportions. Neither in Dalton's published papers, nor in the 'New System,' was any satisfactory account to be found of the genesis of his theories, and hence the question as to whether the atomic theory was founded on an experimental knowledge of the law of combination, or whether Dalton arrived at this law as a necessary consequence of the atomic theory of matter, was not to be gathered from his own writings. The balance of evidence derived from these newly discovered documents is strongly in favour of the statement made in London by Dalton himself, in 1810, that he was led to adopt the atomic theory of chemistry in the first instance by purely physical considerations, in opposition to the view, hitherto held by chemists, that the discovery by Dalton of the fact of combination in multiple proportions led him to devise the atomic theory as an explanation.

2. Report on the Teaching of Science in Elementary Schools.
See Reports, p. 228.

3. The Action of Nitric Oxide on some Metallic Salts.
By H. A. AUDEN, B.Sc., and G. J. FOWLER, M.Sc.

The experiments here recorded are part of a systematic investigation into the conditions of stability of the oxides of nitrogen. They are by no means complete, but the results so far obtained appear to be of sufficient interest to warrant a preliminary notice.

The reactions of nitric oxide have so far alone been studied. The gas was prepared by Emich's method-viz., the interaction of sodium nitrite, strong sulphuric acid, and mercury. The mixture was kept in continual agitation by a specially contrived stirrer, worked by a turbine. In this way a regular stream of gas is obtained, which analysis showed to be of a high state of purity.

In order to study the action of nitric oxide upon the salts selected a weighed amount of the salt was placed in a boat contained in a Lothar Meyer constant temperature furnace. By means of a thermostat, also devised by Lothar Meyer, the temperature can be kept constant to within one degree. Temperatures above the range of an ordinary instrument were measured by means of a high temperature thermometer, constructed by Max Kaehler and Martini, of Berlin, which would give accurate readings to over 400°.

The salt was heated gradually in a stream of nitric oxide, and the phenomena noted as the temperature rose. The salt was weighed at different intervals of temperature and time. Thus it was possible to tell at what temperature reaction began, and at what point it attained a maximum velocity.

So far oxy-salts have been chiefly studied. It was thought that by comparing their behaviour under the above conditions some light might be thrown on their stability, and thence on their constitution.

One or two oxides were first examined, the results agreeing with those of Sabatier and Senderens; e.g., PbO, forms a basic nitrate of lead: when heated in

nitric oxide the action begins at 15°, but does not attain its maximum till over 130°.

MnO, behaves similarly, but the change is not so rapid. It attains a maximum at 216°. In neither case is any but a trace of a nitrite formed.

Silver oxide, if containing traces of moisture, yields a mixture of almost equivalent parts of silver nitrite and metallic silver at the ordinary temperature. At higher temperatures, with the dry oxide, nitrate and metallic silver are formed almost entirely.

Silver permanganate behaves, when treated with nitric oxide, very much as a compound of oxide of silver and a higher oxide of manganese might be supposed to do. It begins to be attached at the ordinary temperature, and at 80° the alteration is very rapid. The residue was found to consist of metallic silver, silver oxide, silver nitrate, and manganese dioxide. Very little, if any, manganese nitrate was formed.

Potassium permanganate is much more stable than the silver salt. It is not appreciably attacked till a temperature of over 100° is reached, and the increase in weight becomes rapid at 190°.

The residue on moistening was not alkaline, and no manganese could be dissolved out. The potassium is converted into nitrate, and the manganese into oxide.

Interesting differences were noted in the behaviour of other silver and potassium salts, notably, the chlorate and iodate.

Potassium chlorate is attacked by nitric oxide at the ordinary temperature, chlorine being evolved in considerable quantity, and nitric peroxide being formed. The gaseous product was condensed in a tube immersed in a freezing mixture, and the percentage of chlorine in the brown liquid obtained was determined. It was found to be much in defect of that required to form nitrosyl or nitroxyl chloride. So that the reaction does not consist simply in the formation of an oxychloride of nitrogen. On analysis of the residue in the boat, no chloride of potassium was found to be present. Nitrate was formed, and also a trace of perchlorate. This seems to be direct proof that in potassium chlorate the potassium and chlorine are separated.

With barium chlorate a similar reaction takes places.

With silver chlorate (prepared according to Stas's method from silver oxide) chlorine was given off, but a considerable amount of silver chloride was also formed, nearly one-third of the silver present being found as chloride. This may be due to a difference in constitution between the chlorates of silver and of potassium, or to a difference in the stability of the salts and the products of reaction.

That some difference of constitution exists between the silver and potassium salts appears to derive confirmation from the behaviour of their iodates when treated with nitric oxide.

Potassium iodate heated to 80° in nitric oxide begins to give off iodine, and the reaction becomes rapid at 110°, crystals of iodine condensing on the cool portion of the tube; no trace of iodide, however, is formed, as is shown by there being no liberation of iodine on acidifying a solution of the residue after adding some potassium iodate. The residue is not alkaline, the potassium being converted into nitrate, recognised by the evolution of ammonia when the residue is warmed with zinc dust and caustic soda.

Silver iodate, on the other hand, is stable up to a rather higher temperature than the potassium salt, and when heated above this temperature, about 110°, no trace of iodine is given off, but all the silver is converted into iodide, none being dissolved out by water, and the yellow residue being insoluble in dilute nitric acid.

The perchlorates and periodates which have been examined show themselves more stable than the corresponding chlorates and iodates.

Of the salts so far examined the chromates have shown themselves the most stable, being analogous in this respect to the sulphates.

Lead chromate was unaltered at temperatures exceeding 400°.

1895.

UU

Silver chromate did not suffer appreciable change till above 300°. Metallic silver was found to be present in the residue as well as silver nitrate. The chromium was all converted into the sesquioxide. Some amount of nitrite of silver was also formed.

Silver sulphate is only attacked at the highest temperature of the furnace.

It was found in certain cases-e.g., with lead nitrate-that the intermixture of a decomposable oxide-e.g., PbO, or MnO,—with the salt, caused the latter to be attacked at a temperature below that at which action begins with either the salt or oxide taken separately.

Experiments have also been in progress on the interaction of nitric oxide and various gases, but the results are not yet quite complete enough for publication.

4. On the Respirability of Air in which a Candle Flame has burnt until it is extinguished. By FRANK CLOWES, D.Sc.

At the last meeting of the British Association the author stated the composition of artificial mixtures of nitrogen and carbon dioxide with air, which were just able to extinguish various flames. It was found that the flames of ordinary candles and lamps were extinguished by mixtures which contained on the average about 16.5 per cent. of oxygen and 83.5 per cent. of the extinctive gases. A flame of coal-gas, however, required for its extinction a mixture still poorer in oxygen, and containing 11.3 per cent. of oxygen and 88.7 per cent. of the extinctive gases. These results have since been confirmed by a different method. The method consisted in allowing the flames to burn in air enclosed over mercury until they were extinguished; the remaining extinctive atmosphere was then subjected to analysis, and its composition was found to be practically identical with that previously obtained from the artificial mixtures. An analysis of air expired from the lungs proved that it was also of the same composition as that which extinguished the flame of an ordinary candle or lamp.

The average percentage composition of expired air and of air which extinguishes a candle flame is as follows:-oxygen 164, nitrogen 80-5, carbon dioxide, 3.1.

Now an atmosphere of this composition is undoubtedly respirable. Physiologists state that air may be breathed until its oxygen is reduced to 10 per cent. The maximum amount of carbon dioxide which may be present is open to question, but it is undoubtedly considerably higher than 3 per cent. Dr. Haldane maintains that the above atmosphere is not only respirable, but would be breathed by a healthy person without inconvenience of any kind; he further states that no permanent injury would result from breathing such an atmosphere for some time.

The conclusion to be drawn from these facts is that an atmosphere must not be considered to be dangerous and irrespirable because the flame of an ordinary candle or oil lamp is extinguished by it. The view is very generally advanced that a man must, on no account, venture into air which extinguishes the flame of a candle or of a bundle of shavings. It will be seen that this precaution may deter one from entering an atmosphere which is perfectly safe and respirable, and from doing duty of a humane or necessary character. An atmosphere which extinguishes a coalgas flame, however, appears to approach closely to the limit of respirability, as far as the proportion of oxygen which it contains is concerned. Hence the coal-gas flame appears to be a more trustworthy indicator of respirability than the flame of a candle or oil-lamp. Undoubtedly the candle and lamp flames should be discarded as tests of respirability of air.

5. The Action of Light upon the Soluble Metallic Iodides in presence of Cellulose. By Douglas J. P. BERRIDGE, B.A., Malvern College.

It was shown by Cook, in 1894, that whilst potassium iodide, purified by ordinary methods, is decomposed by light, the salt is not thus affected if purified by either fusion with charcoal or crystallisation from absolute alcohol. Although