200 



PROCEEDINGS OP THE AMERICAN ACADEMY 



TABLE X. 



It is very difficult in such an investigation to prove the absence of 

 error. It would also seem to be nearly as difficult to imagine a source 

 of error which should reverse its sign, as in Tables III. and IV., when 

 two condensers change places, and yet give consistent results shov^ang 

 an increase of capacity in both cases, — which should also show a 

 change of sign when we pass from the use of condensers whose plates 

 are perpendicular to the lines of force to those whose plates are paral- 

 lel to those lines, — and, finally, which should change in amount with 

 the intensity of the field. 



