092 Mr. J. H. Jeans on 
Gravitation and the Harth’s mag 
netism are by this specula- 
tion traced to the circulation of something (probably the 
ether), such circulation being maintained by the doublets. 
Free electricity not maintaining this particular circulation 
does not partake of its effects. The theory here given for 
the Earth’s magnetic field may be applied to calculate that 
oz any heavenly body whose mass, radius, and velocity of 
rotation round its axis are known. If the principles dis- 
cussed in this paper are sound, they confirm the principle 
that there is essentially but one form of inertia in the 
universe. 
Melbourne, August 1904. 
LXIX. The Determination of the Size of Molecules from the 
Kinetic Theory of Gases. By J. H. Juans, IA, 
Fellow of Trinity College, Camlridge *. 
Si, HERE are five ways in which it is possible, in theory, 
to determine the size of molecules. These are 
rom observations :— 
i.) On the viscosity of gases. 
Gi.) ,, conduction of heat by gases. 
) .. coefficient of diffusion of gases. 
(iv.) ,, deviations from Boyle’s law. 
(v.) ,, volume occupied by matter in the solid 
and liquid states. 
Of these the last can hardly be regarded as of any value 
except as fixing an upper limit to the size of the molecules. 
For we can make no definite inference from observations of 
this kind except on the supposition that the molecules are 
packed as closely as is physically possible, and this suppo- 
sition we know to be erroneous. At the same time, since we 
know in which direction the error lies, the observations in 
question are of interest as fixing a limit in one direction. 
The quantities which can be deduced from these five classes 
of observations depend, however, not only upon the size of 
the molecules but also upon the constant N, which measures 
the number of molecules per c.c. in a gas at normal tempe- 
rature and pressure. If, for the moment, we regard all 
molecules as spheres of diameter c, we find that any of the 
first three kinds of observations enable us to determine No’, 
while either of the last two give No® 
Clearly, if we take one set of obser vations of either class, 
* Communicated by the Author. Read before the British Association 
at Cambridge, August 23, 1904. 
