ON MOLECULAR PHENOMENA IN MAGNETISED IRON. 147 
Secretary on the effect (a) of the diameter of the wire, and ()) of the 
tension on the wire, the effects observed being represented by a curve 
where expansion is plotted against the time in seconds taken in heating 
and cooling. A smooth curve is, of course, formed for the non-magnetic 
metals. A slight break is found in the up side of the curve for fine iron 
wire, and a sharp break on the down side, whereas for steel-wire a sharp 
break on the up side, and a much more marked break on the down side, 
was always observed. Up to Nos. 14 or 15 B.W.G. steel-wire, ¢.e. up to 
a diameter of about +%, of an inch, there is, at the critical temperature, 
decided retraction on heating, but at this diameter a halt, with just per- 
ceptible retraction, occurs, and as the diameter increases the halt becomes 
_ more and wore prolonged, until when No. 3 wire is reached (a rod of 
_ d-inch diameter) at the critical temperature a halt of six seconds takes 
place in the expansion of the wire during heating, and a halt of twelve 
seconds in the contraction of the wire during cooling: the wire in all 
eases being wnenclosed, and therefore freely cooling in the open air. Even 
: with the thickest wire of a quarter of an inch diameter a slight expansion 
~ accompanies the prolonged halt in the contraction during cooling when 
the critical temperature is reached. Several observations, giving con- 
cordant results, were made with each wire. 
Next, as regards the effect of tension on the wire. As might be 
expected, tension has the effect of diminishing the anomalous retraction 
on heating and increasing the anomalous expansion on cooling. In fact, 
with a soft iron wire 20 centims. long (No. 19 B.W.G.) under a tension of 
500 grams no halt is observed on heating, and at the fourth heating only 
a halt and no expansion on cooling; at the eighth heating the halt 
vanished, but a double tension, viz, 1,000 grams, caused an elongation 
amounting to yj55 of the whole Jength of the wire to reappear. In a 
No. 23 hard steel wire, up to a tension of 500 grams, an anomalous retrac- 
tion on heating and expansion on cooling is exhibited, but additional 
tension destroys the retraction on heating and increases the expansion on 
cooling. After, however, thirty re-heatings of this wire the jerk on 
heating had disappeared, even under the reduced tension of 300 grams, 
nor did sudden quenching in cold water restore it, but it reappeared in a 
feeble way under a tension of 59 grams; after the fiftieth re-heating all 
that could be observed was a momentary halt on heating, but the expan- 
‘sion on cooling was as marked as ever. So that even in steel the 
anomalous contraction on heating appears to wear out in thin wires, but 
not the anomalous expansion on cooling. In thicker wires of hard steel, 
Nos. 12, 10, 8, 7, 6, and 2, B.W.G., under tensions varying from 50 to 
_ 8,000 grams, continued heating and cvoling appeared to make but little 
difference.' Here, however, the tension in grams per sq. centimetre was 
“hot so great as in the thin wire. We shall discuss later the cause of this 
curious wiping out of the jerk by annealing or repeated heating and 
cooling, and are continuing the investigation with samples of steel of 
known composition. 
The exact amount of the retraction on heating and expansion on 
cooling was measured by means of a microscope with micrometer eye- 
_' It was noticed that when a flat strip of steel of the same volume as one of the 
thicker wires was tried, the jerk on heating vanished on repeated heating, only a 
momentary halt remaining, the jerk on cooling being as strong as ever; but further 
experiments are being made to determine how far the composition of the steel, or 
the effect of wire-drawing, was influential in these cases. 
Lu 2 
