AND SOME ANALOaOUS RAYS. 
477 
sizes and of diiterent thicknesses of glass. All showed the same general set of 
phenomena. At low exhaustions there was little or no electrostatic deflexion by 
a glass-protected electrode, whether cathodic or anodic. But, at the stage of 
exhaustion where splashing sets in, the electrostatic deflexion of the shadow made 
its appearance, wdth resulting enlargement if the object were made cathodic. In 
several cases the mercury tube pierced, the mercury slowly oozing in minute drops 
into the bulb. When this had occurred, the shadow of the drop was electrostatically 
sensitive at exhaustions lower than that which was necessary to render sensitive 
the shadow of the glass-protected thread of mercury; the shadow assuming in 
consequence a grotesque nodular form. 
It appears, then, that the electrostatic deflexion of cathode rays hy an electrified 
object is dependent upon the surfaee of that object, as to ivhether that surface is or is 
not conductive; and that for objects protected by a non-conducting layer there is a, 
certain minimum stage of exhaustion below ivhich they cause little or no electrostatic 
deflexion of the rays. 
3. The “ Splash ” Phenomenon. 
Many Crookes tubes show the phenomenon already twice alluded to, in which 
the glass surface opi 3 osite the cathode appears to be “ sjalashed ” by the cathodic 
discharge; creeping dendritic forms of an unstable kind ap^Dearing in the luminescence 
of the glass. This “ splash ” phenomenon is independent of the kind of glass used. 
It occurs with soda-glass, lead-glass, and uranium-glass tubes. It does not occur 
Eig. 10. 
on an anticathodal surface of metal, nor, apparently, on an anticathodal glass surface, 
coated internally with plaster of Paris, or with powdered scheelite. It occurs at a 
particular stage of exhaustion a little below that needed for the production of 
Bontgen rays, and at the stage, previously considered, at which there is a rapid 
increase in the electrostatic sensitiveness of the cathode rays. The dendritic forms 
assumed by the splash strikingly resemble the Lichtenberg’s figures (of the positive 
kind), but are in general less fine-grained. The spreading of the “ splashes ” is 
affected by electrostatic influences ; they spread on the inner surface of the tube 
