1240 
strictly defined and reproducible values are not known up to the 
present; they must be determined using the pure a-, f-, y,.... 
modifications of the metals. 
b. Latent heat of fusion. 
10. If a molten metal solidifies at temperatures more or less 
below its melting point, there may be formed different modifications 
which remain simultaneously present after solidification. From this 
we may conclude a priori that different observers will find different 
values for the latent heat of fusion of that metal. These values will 
be a function of the thermal history of the specimen experimented 
with, this history depending on the special conditions of the experiment. 
14. In other words if a definite quantity of a molten metal 
solidifies. the heat effect will be a function of the quantities of the 
a-, B-, y-,... modifications, which are formed during solidification. 
As the heats of transformation are of the same order as the latent 
heats of fusion, we may expect large differences in the case of different 
observers, owing to the varying circumstances of the experiments. 
For instance the latent heat of fusion of cadmium is said to be 
13,7 gram calories per gram’), whilst we found’) the heat of 
transformation of y-cadmium into @-cadmium to be 6.6 gram calories 
per gram at 18° C. 
12. Reviewing the literature, we have found indeed (comp.’ 
Table I p. 1241) that there exist very marked differences. 
While they reach 20 per cent in the case of lead, they are as 
high as 80 per cent in the case of sodium. 
13. Without doubt these deviations are partly to be attributed 
to the presence of impurities in the metals experimented with, while 
the influence of errors of thermometry is also to be taken into 
account. These errors cannot however give rise to the differences 
shown in the table. We hope to report shortly on this question in 
detail. 
14. From what has been said above it is evident, that the latent 
heats of fusion of the metals known hitherto as well as their heats 
of reacton and other data calculated from these values must be 
redetermined with the pure a-, 8-, y-,... modifications of the metals. 
1) Ann. de chim. et de phys. (3) 24, 274 (1848). 
2) Proc. 17, 1050 (1915). 
