— 
TRANSACTIONS OF SECTION A. 635 
It should be observed that Dr. Johnstone Stoney’s investigations on the present 
subject are based on the assumption that helium cannot remain in our atmosphere, 
and he has assumed a temperature of — 66° in his calculations. The present results 
throw doubt on one or both of these two assumptions, 
4. On some Novel Thermo-Electric Phenomena. 
By W. F. Barrert, F.R.S. 
For some time past the author, in conjunction with Mr. W. Erown, B.Sc., has 
been investigating the physical properties of various new alloys of iron, which had 
been prepared by Mr. Rh. A. Hadfield of Sheffield. In the course of this investiga- 
tion a particular nickel steel, to which five per cent. of manganese had been added, 
was found to possess some remarkable physical properties. The analysis of this 
alloy, kindly made by Mr. Hadfield, was as foliows:— 
Iron . . 68:8 per cent. Manganese . - 5:0 per cent, 
Nickel . . 25:0 i Carbon " Fons lg? 2s 
The specific electrical resistance of this alloy was found to be 97°5 microhms 
per c.c. at 15° C., and its temperature coefficient comparatively small, viz. 0-08 per 
cent. per degree C. The thermo-electric behaviour of this nickel-manganese steel, 
when coupled with iron, the author discovered to be most anomalous. Upon heat- 
ing the couple a rapid rise of E.M.F. took place till a certain temperature was 
reached, and then the E.M.F. appeared to remain practically constant in spite of 
increasing temperature up to a white heat, the cooler junctions of the couple being 
kept at 0° C. Careful pyrometric measurements were made as the couple was 
gradually heated, with the result that between 300° and 1100°C. the E.M.F. only 
varied 4 per cent. above or below that at 300°C. The mean E.M.F. in this wide 
range of temperature from a low black to a white heat was 4,000 microvolts, and 
the thermo-electric curve, representing the relation between the temperature and 
E.M.F. of the couple, was therefore nearly a straight line after 320°. Not quite a 
straight line, as the mean E.M.F. was intersected at four points, viz. at 310°, 540°, 
810°, and 1030°C. Had the cooler junction been kept at 310°, instead of at 0°, 
there would have been three successive small inversions of E.M.F. at the points 
named above, that is to say three neutral points. 
The effect of low temperatures (— 80° C.) was tried, and also coupling this alloy 
with some other metals, but no anomalous behaviour was observed. 
When the neutral points of a copper-iron couple were carefully determined, the 
author noticed that the temperature of the neutral point was not the same during 
heating as in cooling. This was specially noticeable in a copper-steel couple, and, 
moreover, in each successive heating the temperature of the neutral point fell until 
it became nearly constant. Here are the neutral points of a particular copper- 
steel couple examined :— 
First heating , 328° Second heating . 283° Third heating. 268°C. 
» cooling , 258° » cooling . 241° » cooling . 241°C. 
Hence the curve representing the relation between the temperature and E.M.F. 
of a copper-steel couple is not the same for a rising as a falling temperature, a con- 
siderable area being enclosed by the two curves. This thermo-electric hysteresis, 
as it may be called, the author also found to exist in many other couples, one 
element of which was iron or an alloy of iron. The explanation is probably to be 
found in the phenomenon of recalescence, and is intimately connected with the dis- 
covery made by Dr. Trouton, F.R.S.,! of a thermo-current produced in a closed 
circuit of iron wire by moving a flame steadily along the wire. 
' See British Association Report, 1889, p. 517. 
