396 ME. J. J. MANLEY: OBSERVATIONS ON THE 



Experiment IV. One arm of the loaded beam was gently warmed. This produced 

 an increase in the deflection. As the beam cooled, the deflection assumed its original 

 value. When the other arm was warmed the deflection decreased ; but as the arm 

 cooled, the deflection assumed, as before, its normal value. 



The first three experiments lead us to infer (a) that even when the beam is kept 

 loaded for an exceptionally long period, on removing the stress, the recovery from the 

 strain is, for all practical purposes, instantaneous ; and (6) the moment the beam is 

 loaded, the flexure reaches a maximum. 



The fourth experiment is interesting because it indicates the necessity of maintaining 

 the whole of the beam at some uniform temperature. 



The beam was also tested with gradually increasing loads ranging from 50 to 

 1000 gr. In every case the flexure reached its maximum value at once, and the 

 recovery after unloading was equally rapid. 



The experiments just described were repeated with the Bunge and circular beams. 

 The results (Table II.) were very similar in the case of all three beams. 



TABLE II. 



The results obtained with the Becker beam are best represented by a straight-line 

 graph. The Bunge beam behaved in a very uniform manner, and all the points lie 

 upon a smooth curve ; but the circular beam showed some irregularity. As this beam 

 was subsequently found to be readily affected by changes in temperature, the observed 

 irregularities are probably to a large extent due to this cause. 



(c) Of the Possibility of Differential Flexure. 



Although the experiments cited in the preceding section support the view that any 

 flexure in a loaded beam is instantaneous, they do not enable us to decide whether 

 such flexure is equally distributed over the two arms or not. To determine this, the 

 following additional experiments were made : 



The beam was supported as before ; but, instead of using the optical lever, a small 



