68 
MESSRS. C. T. HEYCOCK AND F. H. NEVILLE ON 
ol)taiiied from Sn 70 in the sulphur-boiled series, \ye do not think that the earlier 
very high values from Sn 65 are to be trusted ; if we neglect these we get an 
average content of slightly under 39 per cent, of tin in the r], with practically no 
difference between the r] at 550° and 450°. It must be remembered that no method 
of chilling will prevent the presence of a trace of H on the plates of rj. 
The percentage recjuired by the formula CugSn is 38 "35, so that the analyses are 
consistent with the view that the p is this compound with a little of a more tin-rich 
body in solid solution ; but the excess of tin is so small and so easily accounted for 
that we have drawn the solidus from to Eg as a vertical straight line. 
The samples of H were prepared in various ways and isolated from excess of tin in 
the manner above descril^ed. The H prepared by the use of the copper stirrer, that 
is the H of fig. 100, contained 61'2 and 6r4 per cent, of tin ; after re-ciystallisation 
in tin in a vacuous tube and separation from excess of tin, it contained 61’5 and 
61'85 per cent. Sn 45, after 21 days in mercury vapour, gave 6()'42 per cent. 
Sn 60, similarly treated, gave 61T5, and Sn 90, after the same treatment, gave 60‘8 
and 61‘2 per cent. All these, with the exception of the re-crystalli.sed H. had been 
maintained for a long time at 350°. The mean value, neglecting that of the 
re-crystallisation H. is 61T per cent. A sample of Sn 90 was incubated for three 
days and nights at 250°, after a previous jDeriod at 350° : the H extracted from it 
contained 6D91 and 61'73 per cent, of tin. These two values suggest, though they 
can hardly be said to pro^e, an increase in the tin content of the H at lower 
temperatures. The percentage of 61T at 350° gives to the H an atomic percentage 
of 45’5, and agrees extremely well with the results of the microscopic examination of 
the alloys boiled in mercury. We have fixed the point H' in the diagram in 
accordance with these results. 
There is a reasonable certainty that the H is not pure CuSn, though it may very 
well be that body containing a little of the higher compound in solid solution. There 
is very little justification from these results for the slope we have given to the branch 
H'H" of the solidus, and we have therefore drawn it as a dotted line to indicate the 
uncertainty. We think a further examination of such alloys as Sn 47 after a 
prolonged heating at 250° may settle the position of H". At present, the only 
argument for placing H" at Sn 50 is to be found in the numerical value of the 
depression of the freezing-point of tin by small additions of copper, that is, from the 
slope of the branch IK of the liquidus. Tliis appears to prove that, whatever the 
molecule in solution in the liquid tin, it can contain only one atom of copper.* 
We hope that the experimental evidence given in the preceding pages will be 
considered to justify the general conclusions of Section I. 
We are much indebted to Miss D. Marshall, B.Sc., Lecturer at Girton, and to 
Mr. W. Fearnsides, B.A., of Sidney College, for the help they have given us in the 
‘British Association,’ 1900—“Report on Alloys,” by F. H. Neville. 
