Brown — Mechanical Stress and Magnetisation of Iron. 485 



From the curves we see that tlie maximum twist or deflection on the 

 scale occurs for all loads when the wire is in a longitudinal magnetic field of 

 2*5 units. If for this field we plot the values of the load as abscissae and the 

 values of the corresponding maximum deflections as ordinates, the points will 

 be found to lie very approximately in a straight line, and show that when 

 the load on the wire is increased six times the twist or deflection on the scale 

 is decreased 27 per cent. By comparing the twist obtained with this wire 

 when the load was lO'^ grammes per sq. cm. and when in a longitudinal 

 magnetic field of 2.5 o.g.s. units with the twist obtained with another No. 

 16 iron wire with the same load and in the same magnetic field/ we get 

 from Table I., above, a deflection of 38 mm., and in the former case the 

 deflection when reduced to the present conditions becomes 36 mm. This 

 difference of 2 mm. is no doubt due to (1), the former wire being annealed 

 by means of a Buusen flame when it was suspended in a horizontal direction, 

 and the present wire being annealed by the flame when it was hanging 

 vertically and heated from the top downwards, (2) the cross-sectional area of 

 the former wire being 20'6 x 10"^ sq. cm., and that of the latter being 

 20"8 X 10"^ sq. cm., therefore the maximum current through the two wires 

 was 2*06 and 2*08 amperes respectively. Tliis straight-line relation 

 between the load and twist holds very approximately for other longitudinal 

 magnetic fields higher than 2'5 c.g.s. units, and would no doubt also hold 

 for lower fields, if sufficient points had been determined on the rising pai-ts of 

 the curves. The magnetic field of 2'5 units was chosen because the highest 

 deflections occur with this field. 



In order to test how the circular magnetism changed with the increased 

 longitudinal load on the wire, the longitudinal magnetic field round the wire 

 was kept constant at 2"5 c.g.s. units, and a complete cyclic curve taken when 

 the wire was stretched by each of the five different loads. 



An electric current was sent through the wire which was increased by 

 steps up to a maximum of 2'08 amperes, then decreased to zero, reversed and 

 increased to a negative maximum, then back once more to the positive 

 maximum, the twist on the end of tlie wire or deflection on the scale being 

 read off at each step of the cycle. 



A complete cycle was taken in this way when each load was on the wire, 

 and the results plotted on millimetre paper where on the axis of abscissae two 

 centimetres represented one ampere and on the axis of ordinates one centi- 

 metre represented ten divisions on the scale ; the total area of each cyclic curve 

 was then measured in sq. cm., and the results are shown in the last line of 



' Scient. Proc. Eoy. Dub. Soc, 1909, vol xii., p. 186. 



