THE CONSTITUTION OF THE COPPER-TIN SERIES OF ALLOYS. 
5 
form the “freezing-point curve” ABLCDEFGHIK. This has been very carefully 
investigated by Koberts-Austen and by us. It contains angles at G, D, G, H, and 
I, which divide it into six branches, each corresponding to the crystallisation of a 
different solid. Roberts-Auster^ and Stanseield have also shown that it is possible 
to draw a continuous curve through corresponding lower halts in the cooling curves 
of the alloys from Sn 15-5 to Su 27, and thus obtain the curve C'XD'ET, to which 
we, on other grounds, have added the branch I The whole of this “ transforma¬ 
tion ” curve, as the discoverers called it, has recently been traced by us from our own 
experiments, so that, as given in Plate 11, it is a confirmation but not a copy of that 
of Roberts-Austen and Stanseield. The transformation curve has some analogy 
with the freezing-point curve, inasmuch as when the temperature of an alloy falls to 
tliat of a point on either of the two curves crystallisation commences, but while at 
the upper curve the new solid crystallises out of a uniform liquid, at the lower curve 
the phenomenon is a recrystallisation out of a body already solid and crystalline. 
At the top of Plate 11 there are two scales, the upper gives the percentage by 
weight of tin in the alloy, and the scale of equal divisions slightly lower down gives 
the atomic percentage of tin. Compositions expressed in atomic per cents, have the 
advantage of being at once convertible into formulae. Thus 25 atomic per cents, of tin, 
or, as we write it, Sn 25, implies Cu 75, and, therefore, the formula CugSn, and so on.^ 
One way of looking at the diagram is expressed by the statement that a vertical 
line stands for an alloy of a particular composition, irrespective of its temperature, 
while if we follow such a vertical line from the top to the bottom of the diagram we 
are, diagrammatically, watching the alloy cool. 
The freezing-point curve ABLCDEFGHIK divides the area into two parts such 
that above the curve every alloy is a homogeneous liquid, wliile, immediately below, it 
is a mixture of solid and li(|uid. We shall in future abandon the term “ freezing- 
point curve ” and employ the term “ llquidus,” suggested for such a curve by 
Professor Roozeboom. 
The Solidus. 
Professor Roozeboom, in the paper already I'efei'red to, has defined another curve, 
the “ solidus,” which is the complement of the liquidus. Suppose an alloy, or any 
other mixture, caused to cool so slowly that all parts are, at every temperature, in 
true equilibrium with each other. Then when the temperature falls to the liquidus 
a little solid will be formed, and as the temperature continues to fall, more and more 
solid will form, until finally a temperature is reached at which the last drop of liquid 
solidifies and the mass becomes wholly solid. This temperature is a point on the 
“ solidus,” and the curve passes through the solidifying, as distinguished from tne 
freezing, points of all the alloys represented on the diagram. Just as (in the 
concentration-temperature diagram we are considering) all points immediately above 
* Ill calculating the atomic percentage from the percentage by weight, the atomic weights employed 
were, Cu = 6.3‘3, Sn = 118’1. 
