the Recalescence of Iron. 107 



The temperatures given in the third column of the table 

 are deduced from the values of the modulus of torsional 

 elasticity, given in the second column, by means of the 

 formula 



* # =* (l--0002442*--0000002510* 3 )*, 



where z t and z are the values of the modulus at t° C. and 0° C. 

 respectively. This formula was itself deduced from some 

 very careful observations of the torsional elasticity of iron at 

 different temperatures between 0° C. and 100° C. ; but the 

 numbers given in the third column must only be regarded as 

 rough approximations to the true values ; inasmuch as in the 

 first place the above formula was calculated from observations 

 made with a different specimen of annealed iron, and in the 

 second it by no means follows that the formula would apply for 

 such high temperatures as are here recorded. It would seem, 

 however, from the last column of the table, that somewhere 

 near 550° C. there begins to be a rapid rise in the internal 

 friction of the iron, and another still more rapid rise at a 

 temperature of about 1000° C. At this last temperature the 

 internal friction becomes so great that only two or three 

 vibrations are made by the wire before it comes practically 

 to rest ; and it is necessary to multiply the observations very 

 considerably before any approach to accuracy in the results 

 can be attained. Some idea of the enormous friction en- 

 countered by the molecules may be arrived at by bearing in 

 mind that the logarithmic decrement of a torsionally vibrating 

 iron wire at 1000° C. is ten times the logarithmic decrement 

 of a tin wire at a temperature of 20° C, tin possessing the 

 largest internal friction of any metal yet examined at ordinary 

 temperatures. 



From the last line in the table it would seem that the 

 internal friction of iron begins to decrease as the temperature 

 rises from 1100° C. to 1200° C. ; but the numbers here given 

 must be received with some caution in so far as they are to be 

 used for determining tins pointf. 



We tread on surer ground when we compare the second 

 and fourth columns together ; and it is abundantly evident 

 that the decrease of torsional elasticity which ensues on rise 

 of temperature bears no comparison with the increase of 

 internal friction. It is a common lecture- experiment to 

 heat a bell by means of a burner placed underneath, and 



* Proc. Koy. Soc. No. 244 (1886), p. 343. 



t A second set of experiments made with another piece of the same 

 wire showed, however, a similar decrease of internal friction but at rather 

 a lower temperature. 



