412 ME. J. J. MANLEY: OBSEEVATIONS ON THE 



Of Certain Experiments u>hich were made to Test the Efficacy of the Precautions 



advocated. 



Before concluding we would briefly allude to a few experiments of a comparative 

 nature. These were performed by means of the Oertling long beam and protected 

 Gallenkamp balances. Owing to the extremely low value of the temperature 

 coefficient of the former, and the almost complete freedom from strain effects of the 

 latter, these balances were particularly appropriate for the object in view. 



It was desired to ascertain how closely the results obtainable with two balances of 

 totally different types might agree when the precautions, which we now know to be 

 necessary, are taken. For this purpose a block of brass, the mass of which was 

 approximately 100 gr., was weighed at various times on successive days with the 

 screened and fully fatigued Oertling balance. The same block was then weighed a 

 number of times with the other balance. 



The results obtained are set forth in Table V. (see next page). 



As the R.P. coefficient of our Oertling balance was insignificant, the observed values 

 for the mass weighed have not been reduced to a common temperature. Such a 

 procedure would, in the present instance, have introduced a correction considerably 

 smaller than the possible experimental errors due to other causes. 



We note (1) that the greatest difference in the four daily mean values obtained 

 with the Oertling beam are O'lO, 0'07, 0'17, and 0'03 mgr. respectively; these are 

 irregular when referred to the corresponding thermometer readings, and are probably 

 chiefly due to some minute and corresponding variations in the temperature of the 

 two arms of the beam. Such irregularities cannot be completely avoided unless the 

 beam is protected in the manner already described. The difference in the values 

 yielded by the protected Gallenkamp beam was 0'02 mgr. (2) That the two final 

 mean values for the weight W of the block are 



By the Oertling W = 100'007065 gr. 



Gallenkamp = 100'007064 



The difference here observed is not greater than 1/1 7 of the whole. 



This remarkably close agreement affords still further proof of the effectiveness of 

 the precautions which are, in our opinion, necessary when a very high degree of 

 accuracy in weighing is desired. 



We may add that although the values given in Table V., and obtained by means 

 of the protected balance, show amongst themselves a maximum difference of 0'02 mgr., 

 yet further weighing proved that, when the beam was first fully fatigued, the 

 observed differences in the values were, in general, not greater than some few 

 thousands of a milligramme. The correctness of this view is likewise supported by 

 the results represented in the latter portion of the curve in Fig. XII. 



