174 ’ Mr H. FN ewall, On the [Jan. 30, 
very short duration: but in most steel, especially hard steel, it is 
peculiarly marked. 
To test whether chemical action at the surface has anything 
to do with the phenomenon, I have heated wires of hard steel 
in nitrogen and in vacuo and have still observed recalescence 
undiminished. Hence we must conclude that recalescence is not 
to be attributed either to oxidation or to occlusion of gases. 
Expansion of steel during recalescence. 
Dr Gore stretched iron wires in a special expansion apparatus 
and observed that at a certain point in cooling from a bright 
red heat the wire instead of continuing to contract regularly, 
suddenly expanded, and then contracted continuously. He 
thought that the wire to shew this expansion must be under 
tension and attributed the phenomenon to alteration in cohesion. 
Prof. Barrett shewed that this tension was not an essential 
part and that it even prevented part of the phenomenon, namely 
a contraction during heating, from being seen. 
In my own experiments I have found it convenient to have 
the wires under a slight pressure, but have necessarily had to 
be careful that the pressure did not bring about curvature, a 
very small amount of which would easily absorb, so to speak, 
the expansion looked for. I have confirmed Prof. Barrett’s ob- 
servation of expansion simultaneous with recalescence and have 
further observed a contraction during heating occurring simul- 
taneously with darkening or diminution of luminosity in spite of 
constant application of heat. 
The expansion during recalescence, one is compelled to believe, 
must be very considerable: for the reglowing part may be made 
very small compared with the rest of the length of the rod without 
doing away with the balance in favour of expansion. 
If a steel rod is put into a brass tube which completely encloses 
it, and the brass tube is put into an expansion apparatus and 
heated to a bright red heat, the contraction of the brass tube 
in cooling is attended by the same peculiarities as that of the 
steel rod, shewing the accession of heat very markedly at a certain 
point of the cooling. 
Glass hardness of steel produced by sudden chilling. 
It is known that glass hardness can only be produced in steel 
if the temperature from which the steel is chilled is higher than a 
certain temperature about dull red heat. 
I observe that a bar of steel cannot be made glass hard 
unless its temperature is raised above the temperature at which 
‘darkening’ takes place. This is I believe a new observation, 
