PROFESSOR J. A. EWING AND MR. YT. ROSENHAIN 
2:98 
There is one deduction from this solution theory of re-crystallisation that lends 
itself to experimental investigation. If eutectics play an essential part in re-crystal¬ 
lisation, these phenomena should disappear in the total absence of impurities capable 
of forming eutectics. In a perfectly pure metal, re-crystallisation by annealing after 
severe strain should not occur; but it is almost hopeless to obtain a specimen of such 
purity as to justify the conclusion experimentally in this form. The degree of purity 
required can be roughly gauged from the fact that OT per cent, of carbon in iron is 
easily visible as “pearlite” under the microscope; so that, probably, an impurity of 
one part in one million would vitiate the experiment. 
But our conclusion can be narrowed down to more practicable limits by putting it 
in this way :—That if the presence of a eutectic is essential to crystalline growth, 
then a crystalline boundary free from eutectic should be a barrier to all such growth. 
This condition can be approximately realised by means of a weld between two clean- 
cut surfaces of a metal. Lead lends itself particularly well to such experiments, as it 
welds readily under pressure without the aid of heat. In this way we have obtained 
a striking experimental verification of the conclusion deduced from the above theory. 
The welds were made in various ways ; generally two lead discs about IT inch in 
diameter had their surfaces scraped clean with a clean,sharp knife, the two surfaces Being 
put into contact immediately after scraping. They were then subjected to a pressure 
of 5 tons steadily applied in a testing-machine ; in some cases pressures up to 
50 tons were used. The behaviour of the specimens under pressure depended upon 
the previous preparation of the lead discs ; as a rule, and in order to obtain the 
metal in a condition where its crystals would grow rapidly, these discs were prepared 
by crushing a cast cylinder an inch long by f inch diameter. In other cases the discs 
were obtained by casting, and were then only strained when the welding pressure 
came upon them; in these cases there was considerable “flow” while the two discs 
were in contact, but their ultimate behaviour was the same in all cases. 
The welded discs so obtained were found to be firmly united and could be sawn 
into sections as desired ; they were cut up into sections suitable for microscopic 
examination, sometimes before, but generally after “annealing.” The annealing was 
done by exposing the specimens to a temperature of 200° 0. for a considerable time- 
varying from 24 hours to over a month, and the crystals in all cases grew vigorously. 
Sections at right angles to the plane of the weld were then cut smooth and etched 
for examination. On the freshly-cut surface the -line of the weld could never be 
distinguished. The etching had to be carried to a' considerable depth, because we 
found that the cut surface was covered by a thin layer of very minute crystals— 
evidently the result of the violent strains set up by the cutting-tool. 
Microscopic examination, generally at 80 diameters, of these etched sections showed 
that, although the crystals on either side of the weld had grown vigorously, none of 
them crossed the line of the weld, which was clearly visible as an inter-crystalline 
boundary continuous along the whole specimen ; in many cases aggressive individual 
