166 ANNUAL RECORD OF SCIENCE AND INDUSTRY. 



of elastic spheres ; using, however, the methods of his paper 

 published in 1866, rather than those of his first paper, pub- 

 lished in 1860, which latter are more difficult of application, 

 and which he thinks led him into some confusion in treating^ 

 of the diffusion of gases. He finds evidence of considerable 

 importance in favor of his present hypothesis, namely, that 

 published in 1866, in the result of the experimental investi- 

 gations of Loschmidt, who has made a number of determina- 

 tions of the interdiffusion of gases in the cases of hydrogen, 

 oxygen, carbonic oxide, and carbonic acid, to which we have 

 alluded in a previous communication. Proof of a higher or- 

 der, however, is furnished by a comparison between the re- 

 sults of experiments of entirely different kinds ; as, for in- 

 stance, the coefficients of diffusion and those of viscosity. 

 He compares the results of some computations based on the 

 experimental investigations of Loschmidt, Meyer, and him- 

 self, and he finds a certain aofreement such that the ratio 

 of the above-mentioned coefficients agrees within the limits 

 of the possible errors of the determinations. In the theory of 

 the constitution of gases propounded by himself, an important 

 part is played by a quantity called the mean length of the iin- 

 iiiterriipted 2^cUh of a molecule, or, more concisely, the mean 

 path. He calculates the value of this mean path, in hydro- 

 gen, at the temperature zero degrees Centigrade, and baro- 

 metric pressure 760 millimeters, to be 965 millionths of a mil- 

 limeter, and for carbonic acid gas 430 of the same units. In 

 the case of hydrogen, the wave length of the ray of light 

 called the F ray is five times as great as the above -given 

 mean path of the molecule. He then proceeds to enter upon 

 a more hazardous question, namely, the actual size of the 

 molecule ; and making use of such approximate data as are at 

 present only available for this computation, he finds for tlie 

 diameter of the molecule of hydrogen about six millionths 

 of a millimeter, and for the molecule of carbonic acid about 

 nine such units. 12 A, 1873, 300. 



DIFFUSION OF CAEBONIC ACID. 



It has generally been supposed that in an inclosure, pro- 

 tected against the wind, a layer of carbonic acid gas will re- 

 main for a long time, with a layer of atmospheric air above 

 it, without any considerable mixture of the two gases taking 



