456 



NATURE 



{March 22, 1877 



such an extent that their mutual attraction is destroyed ; and in 

 this case gases comply exactly with the Boyle-Mariotte Law. On 

 this hypothesis the law becomes a limit towards which every 

 gas tends in proportion as the dii^tance between its molecules 

 increases, and in proportion as their vis viva and the rapidity of 

 their motion increase. That idea finds no support in facts. If it 

 were accurate, then, at a certain high temperature, gases, espe- 

 cially those whose density is not great and whose particles are 

 endowed with a very great rapidity of movement, ought to con- 

 form rigidly to the Boyle-Mariotte Law ; but this is inconceiv- 

 able, leadin?, in fact, to the paradox examined above. More- 

 over, the researches of M. Regnault on hydrogen, the lightest 

 of gases, have shown the veiy opposite, hydrogen being positively 

 compressible under pressures only a little higher than that ot the 

 atmosphere. It indubitably follows that in reference to the 

 diminution of a mass of a gas filling a certain volume and the 

 increase of the rate of movement of the molecules, we cannot 

 expect a rigorous compliance with the Boyle-Mariotte Law. But 

 we must make certain of finding positive errors, and the follow- 

 ing are the grounds which urged me to have recourse to new 

 experiments in reference to the application to gases of the law 

 with which we are dealing. 



These grounds were explained by me in 1872 ; since then 

 similar ideas have been published by many others. Never- 

 theless, so far as I know, these ideas are far from having been 

 generally adopted into science. It is evident that to combat an 

 established opinion, a single h priori conclusion is insufficient ; 

 new researches are necessary, all the more that doubt may arise 

 in the mind on the accuracy of the data of experiments, not only 

 such as those of Rumford and M. Natterer, but even those which 

 the celebrated experiments of M. Regnault have established. On 

 examining critically the processes of the last-named eminent 

 -experimenter, we may not be able to discover any cause to ex- 

 plain the positive deviations which he has obtained for hydrogen, 

 although it is possible to admit that the negative deviations depend 

 on some defects in experimenting. New experiments and re- 

 searches thus became indispensable ; above all, for the purpose of 

 verifying the data of MM. Regnault and Natterer, following 

 methods not less precise than those which these observers em- 

 ployed. Thus it was specially necessary to experiment on the 

 compressibility of gases under pressures less than that of the 

 atmosphere, seeing that until 1872 there had been no accurate 

 researches on this point. 



I distributed as follows the work undertaken by me in 1872 : — 

 I commenced with pressures less than the pressure of the atmo- 

 sphere, and I passed from that to pressures which exceeded 

 thoie employed by M. Regnault. For the latter purpose I 

 devised in 1872, and have now constructed, a compound 

 manometer, containing alternate columns of mercury and of 

 water, and permitting the measuiement of exceedingly great 

 pressures by means of a large number of very low columns 

 of mercury. However, at present I shall not dwell upon 

 this side of the researches, seeing that the experiments are still 

 being anied on; I shall only endeavour to explain the main 

 p lints of the practical processes, and the results obtained under 

 small pressures. The first experimental researches made by me 

 on the compressibility of air under pressures less than that of the 

 atmosphere were made by means of very simple apparatus. 

 Imagine a vessel A terminated above and below by tubes. The 

 upper tube is always in communication with a syphon baro- 

 meter, or, as I call it, a baro-manometer. In this apparatus the 

 height of the column of mercury measures the elasticity of the 

 gas in the vessel. It is easy to make the volume of gas in the 

 baro-manometer remain the same all the time, notwithstand- 

 ing the variety of pressures. To accomplish this it is only 

 necessary to arrange so that we may at pleasure increase 

 or diminibh the quantity ot mercury in the baro-manometer. 

 The lower tube of A serves to introduce and to withdraw 

 the requisite quantities of mercury. It is not necessary to see the 

 height of the mercury in the vessel ; the mercury here serves 

 only to measure the volumes, and consequently, if we close the 

 feeding-tube by means of a cock, and if by means of an empty- 

 ing tube we allow all the mercury in the vessel to escape, we 

 may ascertain the capacity of the whole reservoir ; by emptying 

 only a part of it we may ascertain the volume of the gas at each 

 moment. Thus the weight of the mercury, directly observed by 

 means of a balance, gives immediately, in all my researches, the 

 volume occupied by the gas in each particular case. The first 

 experiments made in Ib73 on air, showed me that air under 

 pressures lower than that ot the atmosphere possesses a positive 

 compressibility, aiid that the tmaller the pressure the more are 



the divergences presented by air from Boyle's Law increased. 

 The first apparatus was constructed in a very simple manner ; 

 some errors might be suspected, and this is why I am not con- 

 fident of the results obtained. I therefore constructed a second 

 and a third apparatus, modifying successively not only the dimen- 

 sions, but also the very construction of the details of the 

 apparatus. I then arranged a fourth apparatus, by means of 

 which, with Michel KirpiichefT, whose death is a sensible loss 

 to Russian science, I made numerous observations. The report 

 of these experiments was given by me in 1874 to the Russian 

 Chemical Society, and printed in the Bulletin of the St. 

 Petersburg Academy of Sciences. The experiments themselves 

 are described in considerable detail in vol. i. chap. 9, of my 

 work " On the Elasticity of Gases." It is impossible to describe 

 in an article the great number of particulars which belong to 

 these researches ; we are limited to the more important details 

 which I have introduced into this inquiry, as also into my sub- 

 sequent investigations. 



The normal metre and normal kilogramme which I employed 

 were compared with the Paris standards at the Conservatoire des 

 Arts et Metiers in concert with M. Tresca ; their sub-divisions 

 were then carefully verified. 



I had to work a long time at the construction of the baro- 

 meter, and I found out a new process for constructing this appa- 

 ratus, which consists specially in terminating the end of the 

 barometric chamber by a capillary tube bent downwards. By 

 means of this tube it is possible to expel the las.t traces of gas 

 which remain in the vacuum, and thus to show the experiment 

 of obtaining an absolute vacuum, i.e., to construct a barometer 

 Such that with the diminution of the volume of the chamber the 

 indications do not vary. The construction of two barometers 

 with a common chamber and a single descending capillary tube 

 affords an easy means ot obtaining with the greatest precision 

 the determination of the feeblest tensions in the barometric 

 chamber. It is only necessary to direct the telescope of the 

 cathetometer to the top of the column of mercury in one of the 

 barometers when the mercury in the other is at its maximum 

 height, and when the volume of the vacuum is very small ; then 

 pouring out the mercury contained in the other barometer, and 

 thus diminishing the pressure which acts upon the vacuum, we 

 may increase its capacity. Then ihe least quantity of gas con- 

 tained in the vacuum will give an increase of height in the baro- 

 meter observed. By constructing the barometer with the 

 greatest care, and filling it with mercury distilled according to 

 Weinhold's process, it is possible, as our numerous researches 

 with M. Hemilian have proved, to obtain a perfect barometer 

 requiring no correction for the tension of air which may re- 

 main in the vacuum. This result is obtained solely by means 

 of the capillary tube relerred to above. The process shows, 

 moreover, the possibility of constructing barometers without 

 boiling the mercury and without removing them from the posi- 

 tion which they are ultimately to occupy. Here then is an un- 

 doubted improvement in the construction of an apparatus so 

 important as the barometer, in a great number of physical 

 researches. 



Next a very long time and a great number of trials were 

 necessary in order to attain the desired accuracy in measuring 

 heights. I always employ the comparative method, consisting 

 in placing beside the height to be measured a standard metre, 

 weU tried beforehand in all connections. My normal measuies 

 are generally in the form of tubes, in the inside of which is in- 

 troduced water which enables me to appreciate at each moment 

 the temperature of the measure, and if necessary even to change 

 it The telescopes of all my cathetometers are fitted with micro- 

 meter eye-pieces, caiefully constructed by our engineer, M. 

 Brauer, justly noted for his long residence at Pulkova, and for 

 the construction of a great number of astronomical and magnetic 

 apparatus. Besides the central cross-wire the micrometer eye-piece 

 is fitted with one movable wire, or still better, with two movable 

 wires. The fixed wires passing by the optical and geometricall 

 centre of the telescope fitted with a level sensible to about 2-3" are | 

 directed towards a point of the object whose height is to be deter- 

 mined. Then both telescopes fastened to the same cathetometer, .^ 

 or better to two small separate cathetometers, are directed to thej 

 normal measure arranged at the side, and the double movablej 

 wire serves to determine the distance of the fixed wire from tha 

 nearest lines of the normal measure. This last is placed at such 

 a distance from the column whose height is being determined,! 

 that the measure and the object may be distinctly visible without 

 changing the position of the eye-piece. Every variation in the 

 position of the eye-piece may derange the position of the opticalj 



