Nipher — On the Nature of the Electric Discharge. 17 



This figure also shows fainter images of the glass fibers 

 in the position they had been in on a still earlier exposure, 

 the hard rubber frame on which the fibers were mounted 

 having been turned end for end. In these exposures the 

 holder was placed between two plates acting as a con- 

 denser, and 100 spark discharges were sent across a 30 

 cm. spark gap in parallel with the condenser. A shadow 

 image of the hard rubber frame is also visible. 



These sources of disturbance having been discovered 

 the angle wires were surrounded by black curtains in 

 order to protect the photographic plates from the light 

 due to the sparks at the machine, and the plates were 

 then exposed directly to the wires at the angles. The 

 plates were supported on insulating supports at their 

 outer lateral edges. Below the plates was a layer of air, 

 separating the plate from the sheet of glass serving as a 

 table top. It was then found that the films exposed to 

 the angles in the positive line were acted upon as quickly 

 as those in the negative line. It was finally learned that 

 this result was due to negative discharges from the films 

 to the positive wire, and that it was not a fogging of the 

 film by a discharge from the wire to the film, as was ap- 

 parently the case when the film was protected by the hard 

 rubber holder. Fig. B of Plate X shows such a result. 

 Here the limiting effect of pencil marks on the film is 

 shown. Fig. C of Plate X shows a similar limitation of a 

 discharge from an angle in the negative line. Here the 

 negative leakage is outward from the discharge wire, 

 while in Fig. B the flow is inward towards the positive 

 discharge wire. In both cases the discharge lines or 

 fogged areas on the film begin to form immediately below 

 the wire, and elongate outwards, as has been explained. 

 It was this experience which led to the results given in 

 Plates II, III, IV and V, which have been previously ex- 

 plained. 



Some work has been done on the momentum effects 

 around the angles in the negative line, with plates uncov- 



