18 
MESSES. C. T. HEYCOCK AND E. H. NEVILLE ON 
A to D, attacking and dissolving the copper-rich parts. With more tin than 25 atomic 
per cents, the action of the acid is reversed, the tin-rich parts being dissolved and the p 
and H bodies not being touched until all the excess of tin has been removed. It is 
improved in its action by the addition of small quantities of an oxidising agent such 
as bromine or ferric chloride. But the latter re-agent often reverses the light and 
shade of the pattern by what appears to be an electrolytic deposit on the tin-rich 
material. Such complications are not, however, confusing if one examines the same 
alloy in succession in different ways. 
Tn the case of the alloys discussed in the present paper the different ways of 
l)ringfng out pattern confirm each other and leave not the slightest doubt as to which 
part of an alloy is tin-rich or copper-rich. We find that for photographic purposes 
an etch with a dilute solution of ferric chloride acidulated with hydrochloric acid has 
Ijeen most frequently successful. Its principal drawlDack is the conscientious way in 
whicli this solution develops every polishing scratch, however minute. 
Preparation of the Ingots. 
We prepared the little ingots containing from five to ten grammes of allo}^ in three 
ways. The simplest was to melt the weighed materials, or the previously made alloy, 
under charcoal and to allow the ingot to cool spontaneously in the crucible, which 
remained in the furnace until cold. Such a process of cooling is more rapid at high 
tenq^eratures than at low, but there is no abrupt chill or change in the rate of 
cooling. We call these “ unchilled ingots.” In our experiments the ingots generally 
took from two to three hours in cooling. The alloys prepared thus ma}^ be said to be 
in the ordinary state in which previous workers have examined them. We soon 
found, however, that the patterns in these unchilled ingots were too comj)licated to 
admit of a satisfactory interpretation, inasmuch as the records of successive changes 
were superposed on each other in the same ingot. We therefore adopted the j)lan of 
allowing a small ii]got to cool somewhat slowly to a selected temperature and then 
chilling it l^y withdrawal from the furnace and instantaneous immersion in water. 
We thus stereotyped the larger structures that had formed durino- the slow coolino- 
previous to the chill, and although we did not always prevent the changes that were 
due at lower temperatures, yet the very short period of time available for them 
caused the corresponding detail to be, in most cases, very small and readilv 
distinguishable from that formed before the chill. We call these “chilled inc-ots” or 
“ chills,” and speak, for example, of “ the chill at' 775°,” the temperature being that to 
which the alloy had fallen in the furnace before its extraction for chilling. 
These chilled alloys gave us much valuable information, but we found that some of 
the transformations, especially those of the C, D and H temperatures, required a 
period of at least some hours to complete themselves, and that the cooling antecedent 
to most of our chills had not in general been slow enough to allow of complete 
