EXPERIMENTS WITH THE DISPLACEMENT INTERFEROMETER. 63 



AN. Plates in contact 247 



AN. Suspended plate free ooo 



AN. Distance apart of movable and fixed plate . . . .247 



If the mean results be taken from the figure, the following data appear: 



Position of fixed plate 2.05 cm. 2.10 cm. 2.15 cm. 



.165 .067 



.010 .022 



.165 .045 



These results are shown in fig. 3pA. They are such as to indicate an appa- 

 rent attraction of the fixed and movable disks, increasing as their distance 

 apart diminishes. Observations were now made on the other side to see 

 whether the results would be corroborated, or whether in the two hours of 

 observation the shift of the position of equilibrium of the pendulum was so 



A B 



70 



73 



FIG. 39. 



great as to obscure the true conditions completely. The new results for the 

 " small " side are given in fig. 38. They exhibit the familiar inflected curves 

 whenever the pendulum separates from the position of contact, which are 

 observed until the pendulum begins to swing to and fro. In series 14 and 15 

 the movable disk was probably not quite free. The mean results may be 

 estimated as 



0.65 cm. 

 -.050 

 .064 



0.70 cm. 

 -.142 

 .070 



.212 



0.75 cm. 

 .240 

 .071 



Position of fixed plates 0.60 cm. 



AN. Plates in contact 055 



AN. Suspended plate free 086 



AN. Distance apart of fixed and mov- 

 able plate 031 .114 .212 .311 



When the plates are 0.150 cm. apart (0.75), the air-damping is insufficient. 

 The ellipses move to and fro in the spectrum. These results are given in fig. 

 396, and, apart from the observation at 0.60, they contain no evidence of any 

 gravitational effect beyond the limits of error. The time needed was 2 

 hours, within which the shift of the position of equilibrium of the horizontal 

 pendulum is not guaranteed. 



It does not seem possible, therefore, to obtain any definite results from 

 methods which necessarily consume as much time as the present. 



