304 



TRANS. ST. LOUIS ACAD. SCIENCE. 



was directly over the bar. It was not quite at right angles to the 

 glass fibre, so that one end would touch the bar before the other 

 did. The vertical distance from the needle to the bar was there- 

 fore a matter of some uncertainty, and slightly different adjust- 

 ments, which all seemed equally good, gave reading's varying as 

 much as fifteen per cent. 



After taking a number of series of readings a smooth curve was 

 drawn through the points determined by the readings of the nee- 

 dle and the known differences of the potential, the curve being so 

 drawn as to best satisfy the observations. The scale values are 

 given in the annexed table, where d is the reading of the needle 

 and u is the corresponding difference of potential in volts. 



St. Louis, March 8, i88S. 



