3(12 
PROFESSOR J. A. EWING ON THE EFFECTS OF STRESS AND 
the reading of a glass gauge-tube at the side of the tank, which was graduated to 
show each kilogramme of pull on the wire. In all the experiments described in 
Part I. the water-tank was used as the means of applying load ; but in later work it 
was found more convenient, and not less satisfactory, to load the wire by stringing 
discs of lead on a vertical rod which took the place of the tank. 
§ 3. The wire, which was a piece of moderately soft iron, was well annealed, to 
begin with, throughout its whole length by heating it to bright redness and allowing 
it to cool in the air. Its slack ends were soldered to copper leading wires, and these 
were connected to a mirror galvanometer whose resistance was 0‘25 ohm. The effects 
of stress in altering the thermoelectric quality of the iron were measured by the 
deflection of the galvanometer, that being proportional to the electromotive force due 
to the difference of temperature of the two ends of the stressed part. This electro¬ 
motive force was so small that the galvanometer required to be adjusted to a high 
degree of sensibility by using a controlling magnet to weaken very much the earth’s 
directing field on the needle. In consequence of this the zero position of the needle 
became somewhat variable, and to eliminate error from this cause, double readings of 
the deflection were always taken by manipulating a reversing-key in the circuit. At 
the two copper-iron junctions a very slight difference of temperature would have 
given rise to an electromotive force greater than that which it was the object of the 
experiments to measure, and much care had to be taken to keep their temperature the 
same. They were immersed close together (but, of course, without contact) in a large 
beaker of paraffin oil, which was stirred from time to time, and was screened from 
radiation. In later experiments the value of the thermoelectric electromotive force 
was determined in absolute measure. In those, however, which have first to be 
described its value is stated only in the arbitrary unit which corresponds to one scale 
division of the galvanometer. 
§ 4. The effect is called positive when the piece of iron affected has its thermo¬ 
electric position (with the assigned temperatures of the junctions) shifted towards 
bismuth in the thermoelectric series. In other words, the effect is called positive 
when it causes a current to pass from the part affected to the part unaffected through 
the hot junction. 
The following (§§ 5-16) is a summary of the results stated in the paper referred 
to above as “ Part I.,” illustrated by figures taken from that paper. 
§ 5. The first effects of loading are in many cases somewhat uncertain. This may 
be ascribed to the fact that continuous portions of any sample of wire differ more or 
less in thermoelectric quality, and the yielding which occurs during the first loading 
is apt to make some variation in the precise position along the wire of the hot 
iunction. On the whole, however, the experiments agree in showing that the first 
effect of loading an annealed iron wire is negative; that is to say, if there is no current 
before the load is applied, the effect of a moderate amount of load is to make a current 
flow from the unstressed to the stressed portion of the wire through the hot junction. 
