EXPERIMENTS WITH THE DISPLACEMENT INTERFEROMETER. 65 



35. Observations. Long periods and inversion. The unsatisfactory results 

 obtained in the last paragraph induced me to give a final trial to the original 

 method of alternating the 

 sign of the repulsion, by 

 moving each of the two 

 fixed plates in turn near 

 the suspended plate. The 

 results are given in table 

 6 and fig. 40. In series i, 

 2, 3 the equilibrium posi- 

 tion is approached from 

 two opposed directions and 

 for two positions on the 

 large side. The same is 

 the case in series 5 and 6, 

 while in series 4, 7, 8, 9, 

 similar observations are 

 made on the small side. 

 The mean results are 



ZO 40 



60 SO 

 FIG. 40. 



100 120 140 



Fixed disk at 2.10 2.15 0.70 2.15 0.75 



AN. Plates in contact 075 .150 .075 



AN. Movable plate free 0.015 .012 .045 .016 .053 



d = 031 .097 .030 



0.70 

 .150 

 .052 



.101 



0.65 cm. 

 .048 cm. 



Hence there can be no further doubt that the repulsions are real, although 

 their nature has not been made out. When the distance between the disks 

 is larger than a millimeter, the air-damping is insufficient and the free disk 

 unavoidably oscillates, as, for instance, in case of series i and 4. The evidence, 

 however, is none the less definite. In series i and 3, 5 and 6, 7 and 9, the equi- 

 librium position is approached from opposite directions (the displacements of 

 the horizontal pendulum are in half centimeters). 



In order to obtain some reason for this result, one may dismiss the effect of 

 electrical repulsion at once. Experiments, moreover, are to be made in the 

 next section, but rather for the purpose of corroborating the force equation 

 used. Furthermore, friction at the pivots may be excluded, since the pen- 

 dulum is usually in motion, swinging about its position of equilibrium, so that 

 friction would have operated both ways. There remains the possibility of an 

 excess of pressure in the film of air within a metallic fissure as compared with 

 the surrounding air. To obtain some quantitative data, since F' R = 6 

 one may note that the average values of 2AAf were roughly as follows: 



Table 6 2A7V =0.040 cm. 



Tables 1,2 2AAT=o.oi3, 0.022, 0.038 cm. 



Table 4 2AW = o.O4O cm. 



Hence 



Maximum, F' = 65. 2X0.020 = 1.3 dynes. 



Minimum, F'n = 0.4 dyne. 



