Dec. 25, 1879J 



NATURE 



191 



can then be ascertained. The determination of -the coefficient of 

 absorption presupposes, moreover, that the Henry-Dalton law of 

 absorption holds good for caoutchouc as well. That this law 

 is valid is proved by the experiments which I made several 

 years ago on the passage of gases through membranes of caout- 

 chouc, 1 and by means of which I have shown that, at the various 

 differences of pressure between 74 cm. and 2 cm. mercury, the 

 quantity of gas which passes through is proportional to the actual 

 pressure of gas upon the membrane. This relation between the 

 quantity of gas passing through and the pressure is only possible 

 in the case of the coefficient of saturation being proportional to 

 the pressure, or, in other words, when the Henry-Dalton law 

 holds good for caoutchouc within the given limits. 



The absorptiometer which I have constructed for the determi- 

 nation of the coefficients of absorption, consists of glass through- 

 out, a is a tube which is divided into tenths of cubic centi- 

 metres, and from which even hundredths of cubic centimetres may 

 be read off; b, c, and d are glass stopcocks ; e is a space which 

 serves as a receptacle for the caoutchouc, and is closed from 

 beneath by a glass stopper which renders it air-tight when shut. 

 The apparatus stands in a glass trough,,-;, of mercury, and is held 



in a vertical position by the holder /■. Its use is very simple. 

 The membranes of caoutchouc upon which our experiment is to 

 be made, and whose specific gravity has been previously ascer- 

 tained, is cut into strips of about 10 centimetres in length, and 

 I "5 centimetres in breadth, dried, weighed, and introduced into 

 the space e. The apparatus is first of all put in communication with 

 the Jolly quicksilver air-pump by means of the stopcock c, and 

 is pumped empty. Then both the stopcocks d and e are shut, the 

 apparatus is separated from the pump, a drop of water is intro- 

 duced at the bottom of the tube i above the stopcock c, and the 

 gas to be examined enters from above into the space inclosed by 

 the stopcocks b, c, and d. The further working of the apparatus 

 explains itself. If the volume of gas which has been allowed to 

 enter has been measured, and also the pressure under which it is 

 the stopcock d is opened, and after the lapse of from three to 

 twelve hours, the volume of gas and the pressure is again ascer- 

 tained. The calculation of the coefficients of absorption is made 

 according to the known formula. 



I will here remark that the pressure of the gas which remains 

 in the caoutchouc after the apparatus has been pumped free there- 

 from can only be measured by the hundredth part of a millimetre 

 of mercury, which at the same time is the limit of the power of 

 action of the Jolly pump. 



For the experiments, red vulcanised caoutchouc of about one- 

 third of a millimetre in thickness was employed. Its specific 

 gravity at 15° C. was 1-02685. 



The coefficient of absorption of the four following gases wa s 



1 Wietlcvski in Poggcndorjf s Atmalen, clviii., 539-568. 



ascertained : nitrous oxide (N.O), carbonic acid (CO s ), hydrogen 

 and atmospheric air. 



It was shown that the coefficient of absorption of caoutchouc 

 for gases, within the limits of the examination, are linear func- 

 tions of the temperature, and that they diminish with an increase 

 of temperature in the case of nitrous oxide and carbonic acid. 

 The coefficient of absorption of hydrogen, on the other hand, 

 grows larger with increase of temperature, and atmospheric air 

 shows a similar tendency. The coefficient of absorption is as 

 follows : — 



At 5 C. At :o°. At 15°. 



For N 2 ... 1-8229 ... 1-6896 ... 15564 

 ,, C0 2 ... 1-1991 ... 1-1203 ... 1-0416 

 ,, H ... — ... 006121 ... 0-08157 



„ Air ... — ... 0-09832 ... 0-11710 



With the assistance of these values the constant D can now 

 be ascertained. For the description of the diffusiometer which 

 I have constructed for that purpose and for the method of obser- 

 vation, I must refer the reader to my paper in Wiedemann's 

 Annalcn, vol. viii. pp. 29-52. 



The experiments showed that the constant D amounted to — 

 At 120 C. At 14° C. 



ForN s O ... 56 ... 62 j IO _ 8 cm* 



. „ CO ; ... 54 ... 61 1 X ,0 iic" 



Nitrous oxide and carbonic acid have thus almost equal con- 

 stants, a somewhat greater value being accorded to nitrous oxide 

 (being the somewhat specifically lighter gas). The constant 

 for these two gases increases with the temperature, and is at 10° 

 50 times smaller thanZ> for carbonic acid in water, 1 and 300,000 

 times smaller than the constant of free diffusion for carbonic 

 acid and air at the same temperature and the same pressure. 



If the great difference in the coefficients of absorption of 

 caoutchouc for both gases is taken into account, it is at once seen 

 that the constant D depends neither upon the chemical nature of 

 the gas nor upon the value of the coefficients of absorption. It 

 can, in this case, depend only upon the physical properties of 

 the gases, and since specific gravity is the principal property in 

 which gases differ from each other in physical respects, the con • 

 stant D must depend upon the specific gravity of the gases. 

 Proof of this is afforded by the determination of the constant D 



for hydrogen gas : it comes to 353 X 10 - 8 



The constants 



for these three gases is thus nearly in inverse ratio to the square 

 root of the specific gravity of the gases. 



If the behaviour of the nitrous oxide is held as normal, it is 

 found that D is about 27 per cent, greater for hydrogen than it 

 would be if the constant under consideration were exactly in 

 inverse ratio to the square root of the specific gravity of the gas. 

 The same variation here appears which Graham has observed in 

 the diffusion of gases through plates of graphite. Hydrogen 

 diffused itself through a plate of 0-05 centimetres in thickness — 

 supposing the air to show its normal behaviour — about 9 per cent, 

 quicker than is prescribed by the above relation. A similar 

 variation was observed when hydrogen diffused itself in oxygen or 

 carbonic acid instead of air. Granted that this deviation is in 

 inverse ratio to the specific gravity of the gas, it would, 

 in the case of the aforesaid graphite, amount to about 23 per 

 cent, for hydrogen in comparison with nitrous oxide. The 

 deviation is thus with such heterogeneous bodies as vulcanised 

 caoutchouc and compressed graphite, not only of the same 

 direction, but also of the same order, hence there is no ground 

 for supposing that the gas, in its passage through a non-absorbent 

 porous partition like a plate of graphite, should change its 

 aggregate condition, and since the dependence of the constant 

 D of a gas upon its specific gravity can only be considered a 

 sign of the gaseous form of the aggregate condition of the 

 diffusing body, it follows, then, that gases cannot possibly exist in 

 caoutchouc in a fluid form, and they retain also during their 

 absorption by caoutchouc all the properties which belong to them as 

 gases. Graham's hypothesis oj the nature of absorption of gases 

 must certainly, therefore, be regarded as false, and a greater or less 

 degree of penetrability of the layer for one or other of the gases 



1 The constant D for CO- in water is, according to my experiment, about 

 0*000025 — . It depends neither upon the coefficient of absorption nor upon 



the coefficient of saturation. On the other hand it depends upon the vis- 

 cosity of the fluid. If any body, eg., a crystalloid or a colloid, is dissolved in 

 water, and a more viscous fluid is thereby obtained, the constant D decreases. 

 This constant, however, as is shown by my experiments with glycerine in 

 water, cannot be diminished at will by increasing progressively the viscosity 

 of the medium in which the diffusion of the gas takes place. .(See Wiede- 

 mann's Annalen. vol. iv. pp. 263-277, and vol. vii. pp. 11-23.) 



