Report of the Gravity Survey Committee. 223 



the amplitude of vibration usual in pendulum experiments, 

 at least in the early portion of the swing the effect of eddies is 

 not wholly insensible, and therefore it may well be that the 

 formula in which the motion is assumed to be small enough 

 to be regular may not be quite applicable to the actual 

 experiments. However, beyond the discrepancy between the 

 calculated and observed decrement of the arc of vibration, 

 which I have mentioned in my paper, and also the decrement 

 being not quite strictly in geometric progression, there was 

 nothing to indicate that the formulas were in any way in fault, 

 so very good seemed the agreement between theory and observa- 

 tion, until it was shown that the correction of the assumed law 

 as to the relation between the viscosity and the density still left 

 the numerical value of the index of friction as determined from 

 the pendulum experiments slightly too small. But in merely 

 differential observations, such as those carried on with invariable 

 pendulums, T think any uncertainty of this kind would be quite 

 insensible provided that care were taken that the observations 

 should be strictly differential or very nearly so. Hence, if you 

 wish to connect a group of Australian stations with Indian 

 stations it is a perfectly open question whether you shall choose 

 a pressure of say twenty-eight inches for the Australian set or a 

 pressure of say four inches (or whatever the usual Indian 

 pressure for India was). That is on the supposition, which I 

 gather from your letter is intended, that you mean to construct 

 new pendulums. The pendulums being different, the two series 

 cannot be connected till the new pendulums are swung at one of 

 the old stations, unless you are ready to trust to a reference of 

 each series to an absolute detei-mination belonging to it. But in 

 either case if the higher pressure were thought the more 

 convenient for the Australian stations, and it were not wished to 

 trust to a correction for so great a difference of pressui'e as 

 twenty-eight and four inches, it would merely be requisite to 

 swing the invariable pendulums twice in succession at the 

 reference station, once at the higher pressure, to connect with 

 the Australian series, and once at a low pressure to connect with 

 the Indian series or with the absolute determination as the case 

 may be. If the vacuum apparatus be not quite staunch, as I 

 fear may prove to be the case, it might be better, as a matter of 



