190 PHENOMENA DEPENDENT ON MOLECULAR PATHS 78 



Otto Schumann, 1 by oscillation experiments after Max- 

 well's method, 



77 = 0-000168 ; 



Schneebeli, 2 by transpiration experiments, 



7? = 0-000171 ; 

 Tomlinson, 3 by oscillation experiments with cylinders, 



77 = 0-000172. 

 The mean value of all these determinations is 



r] = 0-000172. 



If we join with this value, which is referred to centi- 

 meters, the value of the mean speed at C., as calculated in 

 28, which in centimetres per second is 



ft = 44,700, 



and employ the value of the density of air at C. under 

 the barometric pressure of 76 cm. as given byRegnault, 

 viz. 



p = 0-0012932, 



in grams per cubic centimetre, we find for the mean free 

 path, measured in centimetres, of a particle of air at C. 

 and under the pressure of one atmosphere the value 

 L = 0-0000096. 



From this mean value of the path we obtain, by com- 

 parison of it with the length H travelled in a second, the 

 mean number of collisions that occur per second and of the 

 paths that are newly begun in each second, viz. 

 fl/L = 4,650,000,000. 



The two numerical values of the free path and collision- 

 frequency here established on an experimental basis prove, 

 the former by its smallness and the latter by its greatness, 

 how correct was the expectation which Clausius 4 expressed 

 in his refutation of the objections raised against his theory. 

 The molecular free path has indeed come out as a magni- 

 tude which is three times smaller than the smallest micro- 



1 Wied. Ann. 1884, xxiii. p. 353. 



2 Archives des Sci. Phys. et Nat. Geneve, 1885 [3] xiv. p. 197. 



3 Phil. Trans. 1886, clxxvii. p. 767. 4 Compare 62 and 72. 



