TRANSACTIONS OF SECTION G, 573 
8. On the Importance of previous Magnetic History to Hngineers.* 
By Professor E. Wiuson, B. C. Cuayton, and A. E. Power. 
If iron be subjected to a considerable force of, say, 20 C.G.S. units, and 
be then tested for hysteresis loss, it is known that for a given value of the 
magnetic induction B, the loss may be considerably larger than would have been 
the case had the effects of previous magnetic history been wiped out by careful 
demagnetisation.2_ Experiments made with a 3-kilowatt 60-frequency transformer 
show that the watts due to hysteresis and eddy currents are increased 30 per 
cent. or more when B has a value of about 300, and the magnetising force for this 
value of B is increased by nearly 50 per cent. These results are confirmed by 
ballistic galvanometer tests with specimens of Stalloy and Lohys. In Stalloy 
the ergs per cycle per cubic centimetre are increased by as much as 45 per cent. 
for a value of B of 500. Experiment shows that these effects can persist, and 
are not so easily wiped out as might be supposed, and it is suggested that 
engineers should bear these facts in mind, especially when expecting accurate 
results from current transformers and meters employing iron cores. 
9. The Behaviour of Ductile Material during Torsional Straining. 
By C. KE. Lararp. 
The author presented the following considerations with a view to discus- 
sion :— 
(1) The exact recorded limit of ‘proportionality between strain and stress’ 
may depend on the instrument used and the degree of fineness to which it is 
possible to indicate the strain. 
(2) The yield period in torsion and the yield load are dependent on the speed 
rate of loading. If the test be made very slowly there is no sudden plastic yield 
experienced similar to that obtained in the case where a test is carried out 
moderately quickly. In fact the stress-strain curve is smooth and continuous. 
(3) The approximate laws of flow of ductile during twisting at a uniform 
angular velocity may be stated as follows :— 
(a) The rate of increase of the torque with respect to time varies inversely 
as the time and therefore inversely as the strain or angle of twist. 
(6) The acceleration, which decreases, or as it may be called the decelera- 
tion of the torque with respect to time varies inversely as the square 
of the time or inversely as the square of the angle of torsion. 
(c) The relationship between torque and time, and therefore between 
torque and twist, follows the compound interest law. 
(4) Two theories have been advanced to explain the method of fracturing of 
a specimen. The author’s theory,3 which may be stated: When the maximum 
torque is reached shearing takes place over an annulus near the periphery, the 
shearing extending from annulus to annulus of decreasing mean radius and under 
a diminishing value of the torque until, owing to the irregular form of the 
sheared area setting up a wedging action, a more or less central core of the 
material is fractured by tension. 
A theory advanced by Dr. Wm. Garnett,* in which he suggests that the out- 
side layers being in helical tension and the inside layers in compression, 
resulting in one cylindrical layer being in a state of pure shear so that the first 
part of the failure is due to tension of the outside annulus, while the second 
1 See Zhe Electrician, September 13, 1912. 
* See Journ. Inst. Elec. Hng., vol. 34, pt. 170, p. 55; Roy. Soc. Proc., A, 
vol. 83, 1909, p. 1; Phys. Soc. Proc., vol. 23, pt. 4, June 15, 1911, and vol. 24, 
pt. 5, June 28, 1912. 
* Winnipeg meeting of the British Association (see Hngineering, September 3 
and 10, 1909). 
“ In a discussion following the autior’s Paper before the Institute of Auto- 
mobile Engineers, January 1911. 
