118 
MR. A. E. OXLEY ON THE INFLUENCE OF MOLECULAR 
(d) Correction due to Dissymmetry of the Phial. 
The phials were nearly spherical and were placed with their centres as nearly as 
possible at the point P (fig. l) where the force on them is a maximum and error due 
to displacement a minimum. The present measurements are differential ones, and 
provided the phial is replaced in the same position, no appreciable error can creep in 
from this source. 
(e) Latent Heat and Thermal Conductivity. 
The possible errors due to latent heat of change of state and thermal conductivity 
will be considered in more detail when the preliminary experimental curves are 
discussed. 
(f) Effect of Air Dissolved in the Substance. 
The error due to the presence of dissolved air in the substance is reduced to a 
minimum by heating the substance strongly. In those experiments performed at air 
temperature, an inappreciable amount of air was evolved on fusion. The error 
arising from this cause is negligibly small; for if a quantity of air equal in volume to 
that of the phial were expelled on fusion, it would account for only one-third of the 
actual effect found. But a small trace of air dissolved in the substance will, in all 
probability, remain in the substance if a considerably higher temperature is not 
sufficient to remove it; and although such a trace of air would have a small influence 
on an absolute measurement of the susceptibility, it will not appreciably affect the 
difference of susceptibility between the liquid and crystalline states. 
(5) Method oe Experiment. 
Fig. 6 shows how the value of the force acting on the suspended system varies 
with x , (a) when the phial is filled with air, (b) when it is filled with water. In each 
figure the dotted circle shows the size of the phial, and OM is a distance equal to the 
radius of the pole piece on the same scale. It will be seen that if the phial moves a 
small fraction of 1 mm. either towards or away from O, the resultant deflecting force 
remains very nearly constant, because one-half of the substance and phial has a 
slightly larger force acting on it, while the other half has a slightly smaller force. 
The pole pieces were cylindrical, 8 cm. diameter and 4‘65 cm. apart, so that the 
liquid air tube could be inserted for low temperature work and a small electric 
furnace for high temperature work. The advantage of large pole pieces is that the 
size of the phial is of less consequence. The hysteresis effect was not appreciable and 
the deflections were very consistent; a small increase of the current was readily 
followed by a small increase of the deflection. A Grassot fluxmeter showed that the 
field intensity was 3000 gauss for a current of 8'6 amperes. 
Measurements were first made with the phial filled with air, water, and the 
