166 Trans. Acad. Sci. of St. Louis. 



water-pipe system. The other terminal was grounded 

 on a gas pipe. The lead wires were clamped between 

 the leaves of two small brass door-hinges, one leaf of 

 each hinge being soldered to one of the ends of the rods 

 at the gap. Between the gap in which the leaves were 

 mounted and the ground was placed a large battery jar 

 filled with four litres of a solution of common salt. The 

 wire was parted at this point, and the ends were im- 

 mersed in this solution which thus formed part of the 

 ground circuit. One of the wires penetrated the liquid 

 to a depth of about half a cm. Between the gap in which 

 the wires were mounted and the machine was a spark 

 gap, between knobs of equal radius. The condenser con- 

 sisted of forty large sheets of glass (36x36 inches) hav- 

 ing upon them 2x40 square feet of tinfoil. The machine 

 was driven by a single-phase electric motor. Below the 

 lead wires a sheet of white paper was laid upon a plate 

 of glass. 



It was found that there was a marked difference be- 

 tween the effect of the positive and the negative dis- 

 charges upon the lead wire. The resistance of the gap 

 in the salt solution could be so adjusted, that the nega- 

 tive discharge, or the compression wave, would cause 

 the lead wire to fuse and drop in hot globules upon the 

 paper below. The effect upon the paper is shown in Plate 

 XLW, Fig. 1. 



With the same adjustment, the positive discharge 

 causes the lead wires to rise in a cloud of dust. If the 

 paper were placed 3 cm. below the wires, it would usually 

 be practically unaffected when the positive discharge 

 was used. In a few cases, as in Fig. 2, it was slightly dis- 

 colored by the lead fumes. 



Of course the discharge could be made greater, so that 

 either discharge would cause the lead to be dissipated in 

 a cloud. It could be made less, so that fused metal would 

 fall upon the paper when the positive terminal was con- 

 nected with the ground line containing the wire. In all 

 cases the cloud effect was the more marked in case of the 



