B. — CHEMISTRY. 57 



To apply this method to a selected alloy we first determined its 

 cooling curve by means of an automatic recorder, the curve usually 

 showing several halts or steps in it. The temperature of the highest 

 of these steps corresponded with a point on the liquidus, i.e., when 

 solid first separated out from the molten mass. To ascertain what 

 occurred at the subsequent halts, ingots of the melted alloy were slowly 

 cooled to within a few degrees above and below the halt and then 

 chilled, with the result just seen on the screen. 



The method of chilling also enabled us to fix, with some degree 

 of accuracy, the position of points on the solidus. If an alloy, chilled 

 when it is partly solid and partly liquid, is polished and etched, it 

 will be seen to consist of large primary combs embedded in a matrix 

 consisting of mother liquor, in which are disseminated numerous small 

 combs, which we called ' chilled primaay.' By repeating the process 

 at successively lower and lower temperatures we obtained a point at 

 which the chilled primary no longer formed, i.e., the upper limit of 

 the solidus. 



Although we made but few determinations of the physical properties 

 of the alloys, it is needless to say how much they vary with the 

 temperature and with the rapidity with which they are heated or cooled. 



From a consideration of the sing-ularities in the liquidus curve, 

 coupled with the microscopic examination of slowly cooled and chilled 

 alloys, we were able to divide the copper- tin alloys into certain groups 

 having special qualities. It would take far too long to discuss these 

 divisions. In interpreting our result we were greatly assisted not only 

 by the application of the phase rule, but also by the application of 

 Roozeboom's theory of solid solution (unfortunately Professor Rooze- 

 boom's letters w^ere destroyed by fire in June 1910) and by the advice 

 he kindly gave us. At the time the paper was published we expressly 

 stated that we did not regard all our results as final, as m\ich more 

 work was required to clear up points still obscure. Other workers — 

 Shepherd and Blough, Giolitti and Tavanti— have somewhat modified 

 the diagram. (Slides shown.) 



Neither Shepherd and Blough nor Hoyt have published the photo- 

 micrographs upon which their results are based, so that it is impos- 

 sible to criticise their conclusions. Giolitti and Tavanti have published 

 some microphotographs, from which it seems that they had not allowed 

 sufficient time for equilibrium to be established. In this connection I 

 must call attention to the excellent work of Haughton on the con- 

 strtution of the alloys of copper and tin [Journ. Institute of Metals, 

 March 1915). He investigated the alloys rich in tin, and illustrated 

 his conclusions by singularly beautiful microphotographs, and has done 

 much to clear up doubtful points in this region of the diagram. I 

 have dwelt at some length on this work, for copper-tin is probably the 

 first of the binary alloys on which an attempt had been made to 

 determine the changes which take place in passing from one pure 

 constituent to the other. I would again call attention to the fact that 

 without a working theory of solution the interpretation of the results 

 would have been impossible. 



Smce 1900, many complete equilibrium diagrams have been pub- 



