ANOMALOUS BEHAVIOUR OF DELICATE BALANCES, ETC. 395 



(2) Of Flexure, (a) General Effects. 



It is well known that the sensibility of a balance almost invariably changes with 

 the applied load. The changes may be attributed to the flexure of the beam : and, 

 according to the view most generally held, such flexure would, for all practical 

 purposes, take place uniformly and almost, if not quite, instantaneously. If this view 

 is correct, it follows, provided all other conditions which can affect the position of the 

 R.P. remain constant, (l) that for any given load both the sensibility and the R.P. 

 will remain invariable from the moment the beam is. released; and (2) that any 

 observed deviations will fall within the limits of certain experimental errors, the 

 magnitude of which can be estimated with a fair degree of accuracy. Such errors 

 would naturally assume positive values in some cases, and approximately equal 

 negative values in others ; the curve representing the mean result of a series of 

 determinations of the R.P. would be, if drawn after the manner adopted in Figs. I. 

 and II., a horizontal straight line. 



If, however, the flexure is neither uniform throughout the beam, nor completed 

 within an inappreciable time after releasing the beam, we should expect to find some 

 such effects as those already noticed. Want of strict uniformity of flexure in the 

 two arms would lead to a shifting of the R.P. to the right or left of a truly normal 

 position. Uniform and equal flexure of both arms, whether instantaneous and 

 complete or not, would affect the sensibility only. 



(b) Of the Nature and Degree of Flexure. 



With the object of discovering whether the flexure of the beam, consequent upon 

 the application of any given load, is instantaneous and complete or not, three beams 

 were tried, namely, a Becker solid beam, a Bunge solid beam, and a circular beam. 



In each case the terminal knife-edges of the beam were suitably supported upon 

 glass blocks resting upon a rigid bench. The pan for carrying the required weights 

 was suspended from the central knife-edge by means of a stout wire terminating 

 above in two hooks, between which the beam was inserted. The depressions of the 

 centre of the beam, resulting from the application of various loads, were measured 

 with the aid of an optical lever. With the apparatus used, the smallest depression 

 that could be observed with certainty was equal to 0'0004 mm. 



The following experiments were made : (1) Becker beam. 



Experiment I. A load of 400 gr. was kept in the pan for 13 hours. The load was 

 then removed and the position of the image upon the scale at once noted. Ten minutes 

 later the deflection was still the same. 



Experiment II. The pan was loaded with 900 gr. and the deflection immediately 

 read. During a lapse of 10 minutes the deflection remained unchanged. 



Experiment III. The load (Experiment II.) was now removed and the deflection 

 reading taken. No change in its value occurred during the ensuing 5 minutes. 



3 B 2 



