176 Mr H. F. Newall, On the [Jan. 30, 
junction of the Pt Cu couple, if at all. The free ends were 
attached to a galvanometer which was most satisfactorily dead 
beat, a vane dipping into water being attached to the magnet 
and mirror. The galvanometer indications during heating de- 
noted a continual rise of EMF. in the circuit until the steel 
became red hot; then there came a slight fall of E.M.F. coin- 
cident with ‘darkening’ in the steel ; and then again a continuous 
rise until the temperature of the steel became steady. In cooling, 
the fall of E.M.F. in the circuit went on continuously until after 
diminishing in rapidity the fall changed into a rise coincident 
with the reglow in the steel tube. In making this experiment 
I met with considerable difficulties before success; my difficulties 
were due to my having at first taken too small a steel tube in 
which to enclose the Pt Cu couple, so that the thick wires of 
the couple conducted away the heat too quickly for any actual 
rise in the reglow to be observable—the pause was obvious 
enough. Care was taken to shew that the galvanometer indi- 
cations were not due to insufficient damping. 
We must regard this experiment as shewing that the steel 
tube though put into conditions that promote continuous heating 
or continuous cooling does not either rise or fall in temperature 
continuously. 
The galvanometer indications denote that there is a fall of 
temperature coincident with darkening during heating, and a 
rise of temperature coincident with reglow during cooling, and 
that the temperature of darkening is higher than that of reglow. 
Tn the latter point care has been taken, by reducing the rate of 
heating and cooling, to shew that the indications were not to 
be attributed to inertia in the thermometric arrangements but are 
really signs of something of the nature of what Prof. Ewing calls 
hysteresis. 
If a piece of steel wire is connected up with a galvanometer 
and these connections kept at constant temperature, then if at 
any intermediate part of the wire a part be heated there will 
be no resultant electromotive force unless there is want of homo- 
geneousness or of isotropy or of similarity of magnetic state. 
Let the heating be continued so that the part is heated above 
the temperature of darkening; still no E.M.F. results, if the heated 
part is kept the same. If however this part is shifted along the 
wire so that darkening takes place in front and reglow behind, 
an E.M.F. results, of the order of a milli-volt, and persists only so 
long as shifting continues. The fact that the temperatures of 
darkening and reglow are different shews that we must consider 
the steel between the points at which reglow and darkening are 
taking place as different from that outside those poits. It will 
be convenient to speak of this as altered steel or ‘hot’ steel as 
