ON THE MODULUS OF TORSION AL RIGIDITY OF METAL WIRKR. 



25 



went on. On allowing the wire to cool down and then re-heating to the same 

 temperature on the following day, the period was always very considerably less, the 

 difference being greater the higher the temperature at which the observations were 

 taken. In the table given below, the last column contains the number of hours 

 during which the wire had been maintained at approximately the temperature given. 

 The coincidence period at 33 C. was extremely awkward to work with, and the 

 observations taken at that temperature could not be relied on. They are therefore 

 omitted from the table. 



TABLE III. Soft Iron Wire. 



In Diagram I. the value of the rigidity modulus from the first determination at 

 each temperature is plotted against the corresponding temperature. It seems 

 VOL. cciv. A. E 



