HIGH FREQ TJENCY AND HIGH POTENTIA L CURRENTS. 383

about tlie capacity of two spheres, say, ten centimetres in diam-eter and two to four centimetres apart. The discharge then as-sumes the form of a serrated band exactly like a succession ofsparks viewed in a rapidly revolving mirror; the serrations, ofcourse, corresponding to the condenser discharges. In this caseone may observe a queer phenomenon. The discharge starts atthe nearest points, works gradually up, breaks somewhere nearthe top of the spheres, begins again at the bottom, and so on.This goes on so fast that several serrated bands are seen at once.One may be puzzled for a few minutes, but the explanation issimple enough. The discharge begins at the nearest points, the airis heated and carries the arc upward until it breaks, when it is re-established at the nearest points, etc. Since the current passeseasily through a condenser of even small capacity, it will be foundquite natural that connecting only one terminal to a body of thesame size, no matter how well insulated, impairs considerably thestriking distance of the arc.

Experiments with Geissler tubes are of special interest. Anexhausted tube, devoid of electrodes of any kind, will light up atsome distance from the coil. If a tube from a vacuum pump isnear the coil the whole of the pump is brilliantly lighted. Anincandescent lamp approached to the coil lights up and gets per-ceptibly hot. If a lamp have the terminals connected to one ofthe binding posts of the coil and the hand is approached to thebulb, a very curious and rather unpleasant discharge from theglass to the hand takes place, and the filament may become in-candescent. The discharge resembles to some extent the streamissuing from the plates of a powerful Toepler machine, but is ofincomparably greater quantity. The lamp in this case acts as acondenser, the rarefied gas being one coating, the operator’s handthe other. By taking the globe of a lamp in the hand, and bybringing the metallic terminals near to or in contact with a con-ductor connected to the coil, the carbon is brought to bright in-candescence and the glass is rapidly heated. With a 100-volt 10 c.i>. lamp one may without great discomfort stand as much currentas will bring the lamp to a considerable brilliancy; but it can beheld in the hand only for a few minutes, as the glass is heated inan incredibly short time. When a tube is lighted by bringing itnear to the coil it may be made to go out by interposing a metalplate on the hand between the coil and tube; but if the metalplate be fastened to a glass rod or otherwise insulated, the tube