( 802 ) 
Liquid oocygen I. 
§ 2. Method of the magnetic rise. As mentioned above, we have 
rendered the method of the magnetic rise employed by Quincke, 
du Bois and other observers suitable for use at low temperatures. 
One limb of a vertical O-shaped tube, the upper portion of which 
contains the gaseous, and the lower the liquid phase of the experi¬ 
mental liquefied gas was placed between the poles of a magnet 
whose field was horizontal. 
Let H be the field, {H f the field in the other limb is supposed to be 
so small that ( H'jR )* is negligible), g the acceleration due to gravity, 
z the difference in height of the levels of the liquid under the in¬ 
fluence of H, q and p 0 the densities of the liquid and of the gaseous 
phases respectively, K and K 0 their respective volume susceptibili¬ 
ties, then 
(K-K 0 )H* = 2z( 9 - 99 ) 9 .(1) 
or, by introducing the absolute specific susceptibility x 
(X9-X„C> 0 ) & = 2z {Q—9 o)9- 
If x = Xo then the equation becomes simply 
which is the formula we have used for our calculations. 1 ) 
So there are striking advantages offered by this method parti¬ 
cularly for an absolute measurement, on account of its applicability 
to the case of a liquid in equilibrium with its own vapour. There 
are only two magnitudes to be determined, the distance z, which 
can be measured very accurately with a cathetometer, and the field 
H; nor have we to know the density of the liquid in order to be 
able to find the specific susceptibility. 
Magnetic rise apparatus. It is a very easy matter to cause an 
ordinary liquid to ascend under the influence of magnetic attraction, 
but the experiment is attended by serious difficulties when one has 
to deal with a liquefied gas. Boiling must be completely avoided, 
and care must be taken that the vaporization is unnoticeable. The 
first precaution is necessary because the motion of the liquid or of 
its surface would render adjustment quite impossible, and the second 
!) In § 5 we shall give the reasons why we think that % — * 0 , and should it be 
possible that this is not the case there is still the greatest probability that 
Zo<L5^; in the most unfavourable case at the boiling-point the correction remains 
below 0.002 in value, while at lower temperatures it is quite [negligible on account 
of the small value of 0 . 
