504 EEPORT— 1889. 



extension due to the whole load. With long pauses at the middle load the hysteresis 

 appeared to be reduced to 27 scale dinsions, or -r^l^ of the whole extension. 



Next, a hard wire of high carbon steel was tested, 0-72 mm. in diameter, by- 

 having 14 lbs. permanently on, and loading and unloading 12 kilos. At 6 kilos, 

 the readings were : — ' up ' 9,220, creeping to 9,240 in two hours ; ' down ' 9,400, 

 creeping to 9,370 in two hours. Differences 180 and 130. 



The extension for 1 kilo, was 1,290, so that the above differences correspond to 

 -^^th and ifgth of the whole extension respectively. It is remarkable that the 

 hysteresis is decidedly greater here than in the wires previously tested. 



A brass wire 1*08 mm. in diameter tested in the same way with 12 Idlos. gave 

 readings 6,990 and 7,114 at 6 kilos., difference = 124, which was eqiuvalent to jl-th. 

 of the whole extension, since the extension for 1 kilo, was 1,100 scale divisions. 



With a copper wire 1-15 mm. in diameter, loaded with 10 kilos., the readings 

 at 5 kilos, showed a difference of 52 scale divisions, equivalent to jl^th of the whole 

 extension. 



These experiments make it clear that in conditions of loading such as those 

 they deal with, there is a decided departure from Hooke's law, one effect of which 

 is that some work is done upon the material when it is put through a cycle of 

 stress changes. If the relation of strain to stress were represented graphically, the 

 curves would form loops enclosing a certain area. This result has a sufficiently 

 obvious bearing on the conclusions of Wijhler with regard to the deteriorating 

 effect of repeated variations of stress. 



9. On the relation of the Ether to Space. 

 By Dr. G. Johnstoite Stonet, F.B.S. 



10. On the JE. M. F, produced hij an Abrupt Variation of Temperature at the 

 -point of contact of tvjo portions of the same Metal. By Professor Henry 

 Stkoud, M.A., b.Sc. 



-'■> 



The results of experiments on the E. M. F. produced when two portions of the 

 same metal are in contact, and one portion is Iwjyt at a high temperature and the 

 other at a low temperature, can be predicted from the thermo-electric diagram, 

 and conversely a simple and direct method may thus be obtained for the determi- 

 nation of the Thomson effect. The precautions necessary in the performance of 

 such experiments are discussed in the paper, and the results obtained are given in 

 the case of copper and iron, two typical metals. The E. M. F. established in the 

 case of copper is fi-om hot to cold across the junction, while in the case of iron it 

 is from cold to hot and of far greater magnitude. 



These experiments are being continued with a view to obtaining the relation 

 between the E. M. F. and the difference of temperature. 



MONDAY,' SEPTEMBER 16. 



The following Eeports and Papers were read ; — 



1. Fourth Mepiort of the Committee for inviting Designs for a good 



Differential Gravity Meter. 



2. Report of the Committee for the Collection and Identification of Meteoric 



Dust. — See Reports, p. 34. 



