DI.MK.NSIONS ()K KIMiS AND KoDS OK II{o.N AND OTIIK.K MKTAI.S 227 



The iron was a piece of soft charcoal wire, 7'5 cm. in length, and 0'32 cm. in 

 diameter. The nickel was also a drawn wire, its length being 7 '5 cm., and diameter 

 0*3 cm. The cobalt was a short rod turned from a cylindrical casting. When finished 

 its length was 4'1 cm., and diameter 0'5 cm. 



The nickel and cobalt were supplied by Messrs. JOHNSON and MATTHEY, who 

 prepared them expressly for these experiments. Both were probably purer than the 

 specimens previously used. The new cobalt rod in particular was much softer than 

 the old one, and no great difficulty was experienced in turning it in the lathe an 

 operation which I performed myself. 



The results of a series of experiments with these rods are given in Table VI., and 

 plotted as curves in fig. 6, the points at which observations were made being distin- 

 guished by crosses. 



It was not attempted, and, indeed, under the circumstances, it would not have been 

 possible, to make the observations with the same degree of precision as in the first 

 described experiments. The values of the elongations and retractions are therefore 

 given in two-million ths of the lengths (expressed for convenience as multiples of five 

 ten-millionths) instead of in ten-millionths. To this degree of approximation identical 

 values of the elongations could be obtained for the same magnetising forces with 

 tolerable certainty, and the accuracy of the tabulated results may be accepted as 

 amply sufficient for the purpose in view. 



TABLE VI. 



NOTE. Where no observations are recorded none were made. 



A second series of observations with magnetising forces up to about 400 units was 

 subsequently made in my own laboratory (a battery of seven GROVE'S cells being 



2o2 



