A current of coal gas or of hydrogen was passed through a pipe-stem into the crucible; and this gas, burning over the surface of the molten metal, proved a perfect protection against oxidation. The metal was stirred by a plunging stirrer of graphite. The alloys were made by adding weighed quantities of the second metal in succession to what was originally a weighed quantity of the first metal, and taking the freezing point after each addition.
Section III.
This section contains tables divided into parts and into series. The tables give the freezing point and the composition of each alloy, expressed in percentage weights of one of the constituent metals, and also in atomic percentages. By atomic percentage we understand the number of atomic weights of one metal contained in every 100 atomic weights o'f the two metals in the alloy.
Section IV.
The complete freezing-point curves given in the paper are for the following pairs of metals—Ag-Cu, Ag-Pb, Ag-Sn, Pb-Cu, Sn-Cu, Ag-Sb. But incomplete curves are also given, showing the freezing points of dilute solutions of Bi, Au, Ni, Fe, Al, in copper, and of Bi, Pt, Au, Al, and Tl, in silver.
It has not been our aim to make a special study of very dilute solutions, but the results we have obtained when utilised in the equations given in the paper give as the latent heat of fusion of a gram of copper the number 50 calories, and as the corresponding latent heat of silver 27 calories. This latter number is considerably greater than the 21 calories given by Person, and both numbers can only be regarded as provisional.
The silver-copper curve shows no indication of chemical combination, unless it be the unexpected fact that the eutectic alloy occurs exactly at the composition Ag3Cu2. The comparatively small value of the two atomic falls makes it improbable that the two metals form monatomic molecules in this alloy.
In the silver-lead and silver-tin curves, which have a good deal of likeness to each other, the eutectic alloy contains so little silver that the curve consists almost wholly of the branch starting from pure silver. For the first 20 atoms of added metal the lead curve agrees very well, and the tin curve fairly, with the ideal curve of equation (2) ; but with more lead or tin the total depression becomes much less than that of the ideal curve at the same concentration. We are disposed to see in this,.not an evidence of chemical combination, but rather an aggregation of the lead or tin atoms into larger
