RIGIDITY OF ROCKS A.ND HYSTERESIS FUNCTION. 33 



THE EFFECT OF TEMPERATURE. 



The present arrangemenl being unsuitable for heating and cool- 

 ing the specimen, my intention was rallier to ascertain only the 

 order of the experimental error due to the temperature-variation in 

 the laboratory. To get rid of the effect of the yielding, the specimen 



to be tested was subjected to a constant couple during three days and 

 nights. Then, the creeping due to the yielding having become com- 

 paratively small, the temperature-variation of the twist was observed. 

 The time- variations of both the temperature and the twist were, as 

 they are shown in Figs. 50-52 in PI. XXII., similar to each other. 

 In Fig. 53 in the same plate, we see the increase of twist plotted 

 against the corresponding rise of temperature. The curve, as a 

 whole, expresses the simple proportionality between (lie two clement*. 

 \\ e find, however, one remarkable fact that the temperature-variation 

 of the twist has a minimum value in the neighbourhood of 9°C. It 

 may he questionable whether this is a general property of all rocks or 

 merely special to that specimen alone. By the way, it may be cited 

 here that, from the experiment of G. Wertheim (I) . the stretch modulus 

 of iron and steel seems to have a maximum value at or a little below 

 100°C, since at 200° it is sensibly the same as at — 15°. 



Lastlv, we must not neglect to remark here that though the 

 temperature-variation of the twist is unexpectedly great, yet it is 

 ahnest negligible as compared with the total amount of the twist and 

 also with that of the yielding. Indeed, the increase of twist per 

 degree of temperature-rise is less than ",\ x 10 " 3 of its total amount in 

 sandstone. 



(1) G. Wertheim. Annales de chéinie. T. KV. 1S15. 



