178 Mr H. F. Newall, On the [Jan. 30, 
so that the experimental sphere occupied that spot. Heating was 
done by means of a Bunsen burner flame. 
Beyond the observation of correspondence of the character 
of reglow with that of the appearance of magnetic susceptibility, 
my results are not yet of such a kind as to warrant publication. 
General remarks. 
Many experimenters have shewn that there are as marked 
differences between steel above and below a certain temperature 
about red heat as there are between steel and other metals recog- 
nised as different. 
We may represent the different properties and their variations 
with temperature graphically, and we evidently deal with one 
and the same substance in steel with rise of temperature until 
darkening occurs. Then in the graphical representation we pass 
from one line to another and for certain limits of temperature 
about red heat there are two values of the ordinate for each value 
of the temperature abscissa, the limits being those corresponding 
to the temperatures of reglow and darkening. We can trace 
the curve for cold steel in heating up to darkening, and that for 
hot steel in cooling down to reglow. 
There is an actual dimznution of luminosity, an actual fall 
of temperature, not only a pause in the increase, as heat is applied 
to steel from without. As steel is heated, a certain pomt is 
reached, when some action—it may be physical or it may be 
chemical—begins and continues with the “ violence of instability”. 
Hither it is a case similar to that of water heated above its boiling 
point and suddenly bursting into vapour with a fall of temperature 
down to the boiling point: or it is some chemical decomposition 
of unusual type, which should have begun at a lower temperature, 
but which having been delayed continues, when once started, like 
an explosion. I cannot find reference to any such chemical 
action. Prof. Livemg has pointed out several chemical actions 
accompanied apparently with a fall of temperature, but has ex- 
plamed these by shewing that the fall is due to physical, not 
chemical changes, volatile compounds being formed in certain 
cases at temperatures above the boiling points, so that when 
formed they boil away with reduction of temperature. 
The only apparent chances for explanation by reference to 
chemical action are not hopeful :—carbides of iron, and allotropic 
changes—I hesitate to name ferricum and ferrosum. 
Reglow in cooling is somewhat similar in aspect to a pheno- 
menon, to which Prof. Dewar kindly called my attention. In 
gold assaying it 1s often observed that a molten button of gold 
may be cooled below its solidification point without solidifying. 
