114 Mr. H. Tomlinson on Thermal Conductivity [Feb. 21, 



equal to the increase in the other, but it is the intention of the author 

 to make further researches into this part of the inquiry. 



The batteries employed in these and subsequent experiments were 

 slight modifications of Daniell's batteries, which, though having a very 

 small resistance, maintained a constant current for some hours. These 

 batteries the writer hopes to have the honour of describing to the 

 Society on a future occasion. 



Similar experiments were next made on a piece of hard steel, of 

 similar dimensions to those of the iron, but the length of the brass bar, 

 AE, was considerably shortened. 



The result of the experiments proved that there was a decrease of 

 flow, amounting to about 4 per cent., of the whole, when the bar was 

 magnetised longitudinally with a current producing a deflection of 18° 

 on the tangent galvanometer, and an increase of flow when magnetised 

 by a current of 10° C transversely, amounting to about 3 per cent, of 

 the whole (unfortunately, through accident at the time, the same 

 magnetising current as used for longitudinal magnetisation could not 

 be employed). 



This last result was rather unexpected, as, though Sir William 

 Thomson has shown (Phil. Trans., Feb. 28, 1856) that the electric 

 conductivity of hard steel is diminished when the steel is magnetised 

 longitudinally, some experiments of the writer in the same direction 

 (Proc. of Royal Society, June 17, 1875) seemed to show that, in the 

 case of hard steel, the contrary effect is produced, while Joule has 

 proved (Phil. Mag., 1847) that, while a bar of iron, or even steel 

 which is so hard that a file will just touch it, is lengthened by longi- 

 tudinal magnetisation, iron under great strain, or steel so hard that 

 a file will not touch it, is shortened. 



Of course, the mere act of soldering the steel to the brass would, to 

 a certain extent, soften the steel. A bar of steel, therefore, of circular 

 section, \ inch in diameter and about 12 inches in length, was made 

 so hard that it could not be touched with a file, and placed, with the 

 usual preeautions, on the poles of the electro-magnet, at a height of 

 about 3 millimetres above the cores of the electro-magnet, one of the 

 Gr. S. elements being secured to one end and the other end inserted 

 in a Leslie's cube. But here, again, there was undoubtedly a de- 

 crease of conductivity, when the bar was magnetised longitudinally. 

 Of course, however, the mere fact of heating the bar at one end to 

 the temperature of boiling water would tend to soften the bar, and 

 the writer is not quite satisfied but that it may be ascertained that, 

 for low temperatures, the conductivity of hard steel is increased. 

 It is intended, therefore, to make experiments on iron and steel, at 

 different temperatures, both high and low, with a view of definitely 

 settling this point. 



Some experiments were next made with the magnetising coil de- 



