130

INVENTIONS OF NIKOLA TESLA.

any dielectric which is subjected to great, rapidly changing elec-trostatic stresses. Since it is difficult to exclude the gas perfectlywhen solid insulators are used, it is necessary to resort to liquiddielectrics. When a solid dielectric is used, it matters little howthick and how good it is; if air he present, streamers form,which gradually heat the dielectric and impair its insulatingpower, and the discharge finally breaks through. Under ordi-nary conditions the best insulators are those which possess thehighest specific inductive capacity, but such insulators are notthe best to employ when working with these high frequencycurrents, for in most cases the higher specific inductive capacityis rather a disadvantage. The prime quality of the insulatingmedium for these currents is continuity. For this reason prin-cipally it is necessary to employ liquid insulators, such as oils.If two metal plates, connected to the terminals of the coil, areimmersed in oil and set a distance apart, the coil may be keptworking for any length of time without a break occurring, orwithout the oil being warmed, but if air bubbles are introduced,they become luminous; the air molecules, by their impactagainst.the oil, heat it, and after some time cause the insulationto give way. If, instead of the oil, a solid plate of the bestdielectric, even several times thicker than the oil interveningbetween the metal plates, is inserted between the latter, the airhaving free access to the charged surfaces, the dielectric ivari-ably is warmed and breaks down.

The employment of oil is advisable or necessary even with lowfrequencies, if the potentials are such that streamers form, butonly in such cases, as is evident from the theory of the action.If the potentials are so low that streamers do not form, then itis even disadvantageous to employ oil, for it may, principally byconfining the heat, be the cause of the breaking down of the in-sulation.

The exclusion of gaseous matter is not only desirable on ac-count of the safety of the apparatus, but also on account ofeconomy, especially in a condenser, in which considerable wasteof power may occur merely owing to the presence of air, if theelectric density on the charged surfaces is great.

In the course of these investigations a phenomenon of specialscientific interest was observed. It may be ranked among thebrush phenomena, in fact it is a kind of brush which forms at, ornear, a single terminal in high vacuum. In a bulb with a con-