693 
more quickly. The instruments are however sufficiently rapid to 
be also serviceable in case of more quickly crystallizing substances, 
though the accuracy will then, of course, be less. 
2. Salol can be supercooled very easily. Spontaneous crys- 
tallisation, without solid substance being purposely added to the 
supercooled liquid, can be almost entirely excluded. The spontaneous 
setting in of crystallisation only becomes troublesome with very 
strong supercooling (more than 40°). This is, however, only the 
case when the substance is kept sufficiently dry. A slight quantity 
of water immediately causes the formation of ‘seeds’ of the solid 
substance in different parts of the liquid. Observation is rendered 
impossible by this. 
3. The melting-point of salol is very conveniently situated (42° C.), 
so that the observations can be made in the neighbourhood of 
room temperature. 
The observation is now carried out in the following way. The 
crystallisation is started on the upper side’) of the liquid, which 
is in one leg of a U-shaped tube, by the introduction of a small 
quantity of the solid substance. For this purpose the tube is placed 
in a well-stirred thermostat filled with water. In this way the joints 
of the extremities of the thermo-element and the copper wires con- 
ducting the current to the galvanometer, are kept at the constant 
temperature of the water in the thermostat. Accordingly the galva- 
nometer indicates the difference of temperature of the joint in the 
axis of the cylindrical tube and the thermostat. The deflection of 
the galvanometer was photographically registered, so that the regis- 
tered curve enables us to see and measure at a glance how the 
temperature has changed at a definite point of the axis in course 
of time. 
The curve obtained, has however, still another meaning. During 
solidification, the boundary plane of the solid and the liquid 
phase moves with constant velocity and retains its form. This ren- 
ders it probable, that the distribution of the temperature in the solid 
substance and in the liquid will also move unchanged with this 
velocity. A theory of the process of solidification confirms this sup- 
position °). The temperature which is registered at a definite point as 
1) To avoid convection currents in the liquid, the crystallisation must proceed 
from above downward, and not inversely. The liquid has, indeed, the highest 
temperature at the surface of the solid phase, where the heat of fusion is liberated. 
If the hottest place of the liquid is at the top, convection currents through diffe- 
rence of temperature cannot occur. 
5 H. C. Bureer, These Proc. XXIII 1920, p. 616, further cited as loc. cit. 
45* 
