1888. ] Recalescence of Steel, and allied Phenomena. 173 
bright patch in the middle of the hottest part of the plate and 
finally disappears altogether. These appearances I have observed 
in plates } in. thick, but most markedly in the very thin plates 
used in the manufacture of pens. Messrs Gillott of Birmingham 
were kind enough to present me with several of these thin 
sheets. 
Barrett, who first records observations on recalescence, failed 
to observe it in wires of small diameter or in bars of large 
diameter. JI have observed it in steel plate ‘1 mm. thick, and 
in steel spindles over 13 mm. in diameter. I have no reason to 
believe that the phenomenon does not occur in the largest ingots 
of steel. 
Barrett does not seem to have satisfied himself completely 
of the existence of a corresponding effect in the heating. I have 
described above the appearance of the dark patch and expanding 
ring, and have no difficulty in observing this ‘darkening’ even 
in wires ‘5 mm. diameter, though it is most markedly seen in 
the sheet steel. 
Prof. Barrett shewed by enclosing a wire in a kind of air ther- 
mometer that at the moment of recalescence in cooling there was 
an increase of pressure in the air surrounding the wire. I have 
shewn by putting a thermo-electric couple into a small hole in a 
thick bar of steel, that there is a rise of temperature throughout 
the mass of the steel. 
Prof. G. Forbes (Proc. R. S. Edin. vitt. 363, 1876) has at- 
tempted an explanation of the reglow, roughly as follows. Iron 
is a bad conductor af heat—worse above than below dull red 
heat (Tait). Radiation from the surface carries off heat more 
rapidly at first than it is supplied from within by conduction. 
As cooling proceeds, the radiation diminishes, conductivity in- 
creases. At a certain point in the cooling the supply from 
within is greater than is disposed of by the surface by radiation. 
Hence the surface rises in temperature and will even reglow. 
Prof. Forbes also gave reasons why the effect was not observed 
in very small or in very large pieces of metal. There must be 
some thickness that the difference of temperature within and 
without may be great enough: this is not attained in thin wires; 
there must be surface enough relatively to mass, and convection 
currents enough, to bring about the cooling of the radiating 
surface: this is not attained in large masses of metal. 
I have observed recalescence and darkening in steel whose 
thickness is less than ‘1mm. Hence we must believe that there 
can be very considerable difference of temperature in ‘(05 mm. of 
iron, or else give up the ingenious theory suggested by Forbes. 
I find that some soft iron does not shew reglow at all: that 
in some hard iron, it is very difficult to observe it because of its 
138—2 
