40 Mr. W. Sutherland 



on a 



already given in this paper will be as before to measure 

 rigidity for each metal in terms of that at common tempera- 

 tures (about 15°) as unity, and then according to the data 

 given in the earlier part of this paper for lead, zinc, tin, and 

 magnesium, and in Table I., after Pisati, to calculate a value 

 of Sf/ri! from equation (1) for each value of n given for the 

 different values of 6. 



For example, in the case of tin we have the following 

 values, T=503° (absolute):— 



0... 289° 324° 339° 359° 376° 



n 2 / ni ... 1-0 -854 -804 -727 -666 



which give respectively the following values of N/^ : — 



1-492 1-461 1-472 1-480 1-510, 



with a mean value 1*483, the departures from which are 

 within the limits of experimental error. 



In the case of zinc, as already mentioned, there is dis- 

 crepancy between Tomlinson's results and my own ; Tom- 

 linson gives for the rigidity of zinc at any temperature the 

 formula 



n t =n (1- -00108* --0000049* 2 ), 



by means of which I have calculated the rigidities at the 

 same temperatures as my own determinations were made at 

 in terms of that at 18° as unit in the following, T=690° : — 



291° 321° 336° 355° 370° 



n 2 / ni (Toml.) 1-0 -955 -932 -896 -865 



These give the following values of N/n x : — 



1-217' 1-218 1-221 1-218 1-215 ; 



while at the same temperatures my own experiments give 



na/wi l'O '952 -906 -842 -784 



N/W] 1*217 1-215 1-188 1-145 1-101 



Thus, while Tomlinson's values confirm the formula with 

 wonderful accuracy, my own fail to do so, the reason being I 

 believe that my specimen of zinc was not a genuine wire, 

 whereas Tomlinson worked with a properly drawn wire ; the 

 mean value of N from his experiments is 1*218, with a 

 maximum departure in the individual experiments of -25 per 



