4 
MESSRS. C. T. HEYCOCK AND F. H. NEVILLE ON 
ti’acing a curve was generally about two hours, and a little carbon was throAvn on 
the surface of the ingot, to avoid oxidation, or, as an alternative, a little coal-gas was 
allowed to hum at the mouth of tire crucible. 
A detailed discussion of the various curves is not necessary here, as they will be 
referred to later on, but it may be worth Avhile to say a ferA^ AA'ords about their 
general character. If a suhstance remained liquid throughout the Avhole range of 
cooling and underAvent no chemical change we should expect its cooling curve to be a 
sloping line, prol)ably someAvhat steeper at high temperatures than at Ioav ones, but 
free from abrupt changes of direction. The short upper branches of all the curves, 
near to where they are numbered, correspond to this cooling of a uniform liquid. 
But Avhen solidification, or any other exothermic change, begins the rate of cooling 
Avill be aijruptly decreased and the cooling curve aauII tend to become horizontal. 
This is Avell seen in the ujjper-niost curve of Plate 10, the flat under the letters 
“ Cu ” being due to the isothermal solidification of the copper. The upper flat in the 
curve of Sn 4 is also due to the formation of solid, and in this case a slio-ht surfusion 
must have occurred, the temperature of the liquid falling below the freezing-point, and 
then, 1)y the evolution of the latent heat of solidification, being perceptibly raised. But 
the flat is shorter than Avith pure copper, and rapidly rounds off in consequence of the 
solidification not being an isothermal process. The solidification of Sn 4 is probahly 
completed at 850°, and lieloAv that point there is no marked singularity in the cooling 
curve. But if Ave consider the curves of Sn 6, 8, 10, 12, or 14, aa'c see that not only 
does the uppermost halt, the freezing-point, become loAver and loAAmr as the 
percentage of tin increases, but that some distance beloAv this halt, at a tenqDerature 
almost the same for all tire curves, there is a second evolution of heat Avhich becomes 
moi'e and more marked AAuth increasing content of tin. Moreover, in the curA^es of 
Sn 10 to Sn 15, there is a third evolution of heat at about 500°, and this, as Ave shall 
be able to prove, occurs in an absolutely solid alloy. Similarly Sn 17 is solid at 700°, 
l)ut its cooling curve sIioaaas AA’ell marked heat eAmlution at a temperature below 500°. 
Valuable though cooling curves are for indicating critical points in the cooling, they 
harm a defect; in order to make tlie Avhole period of tracing the curve a reasonable 
one, it is unavoidalde tiiat tlie cooling at the higher temperatures should be somewhat 
rapid, hence certain rather sluggish reactions do ]iot complete themselves at the 
equilibrium temperature, and may indeed never become complete. One must, in 
coiisequence, exercise some caution in attemiAting to infer the magnitude of tlie heat 
evolution due to a particular change from the length of the flat it i)roduces. 
The Temperature ConcoitratioR Diagram .—(Plate 11.) 
The informati(.>n supplied Ijy the cooling curves can be presented in another and for 
some purposes more convenient form by taking the composition of each alloy as tlm 
horizontal ordinate, and temperature as before for the Amrtical one. Our Plate 11 is 
constructed in this AAuy. Here all the first halts of tlie cooling curves are used to 
