46 Dr. E. Taylor Jones. 



and corresponding field strength on several nights in July and August, 

 1897, the temperature indicated by the coil thermometers being 

 about 20 C. As before, the temporary and residual contractions 

 were measured separately, the total contraction corresponding to any 

 field strength being obtained as the sum of the temporary effect 

 (observed as an elongation when the current was broken), and the 

 residual effect observed by first demagnetising the wire by reversals, 

 then quickly making and breaking the current. It was found that 

 the curves thus obtained were not the same on different nights, but 

 showed a continual diminution of contraction as time went on. This 

 is probably due to a slow hardening of the metal which seems to go 

 on for some time after the wire has been annealed. The effect of 

 this change can be seen by comparing Tables I (column 2) and III 

 (column 4), which contain values of the contraction of the wire in 

 million ths of its length for a series of fields, observed about the 

 middle of August, 1897, and at the end of December, 1897, respec- 

 tively.* The wire was annealed on July 3, 1897. 



In order to observe the effect of a change of temperature on the 

 magnetic contraction at any field strength it was necessary to take 

 readings at two temperatures immediately after one another, so that 

 the results would not be affected by the above-mentioned time effect. 

 This was done on several nights and with several field strengths, the 

 two temperatures used being about 20 C. and 55 C. The results 

 showed that at low fields (up to about 90 C.Gr.S.) this rise of tempera- 

 ture of 35 C. caused an increase, at higher fields a diminution, of 

 contraction ; and if the temperature was then lowered back to 20 C. 

 the contraction returned to its former value. 



Table I contains values of the change of length in millionths and 

 corresponding field strength at the two temperatures, determined on 

 two successive nights, the effect at both temperatures being observed 

 at the lower fields on one night, and at the higher fields on the next. 

 Each field was reversed several times before readings were taken, 

 and the wire was demagnetised on each night while the tempera- 

 ture was being changed. The curves representing these results 

 (contraction, field) have the same general form as was described in 

 the former paper, but the temporary contraction is here considerably 

 greater, though the residual contraction is much the same as before. 

 The effect of temperature is chiefly seen in the temporary contrac- 

 tion : the residual effect, however, appears to be slightly less at the 

 higher temperature at all fields. 



On referring to Professor E wing's 'Magnetic Induction in Iron 

 and other Metals ' ( 114, p. 169), I find that rise of temperature 

 causes increase of magnetisation in nickel at low fields and diminu- 



* The temperatures on the two occasions differed by about 10" C., but the effect 

 of this would be comparatively small. 



