410 



MR. J. J. MANLEF: OBSERVATIONS ON THE 



produced, that the discs r, s, t (Fig. X.) were added. These intercept and deflect the 

 convection currents away from the openings and cause them to expend their energy 

 upon the solid plate. 



Of the Minor Undulations to be found in a Resting-Point Curve. 



We are now in a position to determine experimentally whether the minor and 

 persistent sinuosities found in nearly all our K.P. curves are due to temperature 

 fluctuations more or less periodic in their nature. 



We know that a periodic variation in the temperature at one end of a bar will give 

 rise to heat-waves which will journey towards the other end. If the oscillations in 

 the temperature take the form of a simple harmonic motion, then for a homogeneous 

 and uniform bar the temperature variations will be correctly represented by a curve 

 of sines. As successive heat-waves follow each other, the bar will be alternately 

 lengthened and shortened ; and if the variations in length be represented graphically, 

 we again obtain a curve of sines. 



Now, any appreciable changes in the relative lengths of the two arms of a balance, 

 whether harmonic or not, should manifest themselves by corresponding variations in 

 the K.P. Hence, from the form of any given RP. curve, we ought to be able to 

 draw a fairly correct inference as to the nature of the heat-waves that may be 

 travelling along the beam at the time of weighing. 



Conversely, if it can be shown that the temperature of the beam remains strictly 

 uniform during the operations of weighing, then the R.P. curve for a perfectly 

 fatigued balance will be a right line, horizontal in direction. 



This theory was tested by means of the protected Gallenkamp balance, which for 

 the purpose was loaded with 20 gr. (Any other load would have been equally 

 convenient.) The beam was released and observations commenced and continued, 

 without any preliminary fatiguing of the beam, until sufficient data had been acquired 

 for 19 independent determinations of the R.P. The results obtained in this experiment 

 are shown in the curve, Fig. XII. 



tos-o. 



104-0. 



103-0. 



IOZ-0. 



101-0 



Loa.d.^20 c/rms. Temp Uniform 



-e e- 



'/! Complete Vibrations. 



02468/0 14 18 22 26 30 34 38 



Fig ZH 



