HIGH FREQUENCY AND HIGH POTENTIAL CURRENTS. 135

frequency discharges are directed through the wire, it is heatedmore on the ends than in the middle portion, and if the fre-quency, or rate of charge, is high enough, the wire might aswell be cut iu the middle as not, for most of the heating on theends is due to the rarefied gas. Here the gas might only act asa conductor of no impedance, diverting the current from thewire as the impedance of the latter is enormously increased, andmerely heating the ends of the wire by reason of their resistanceto the passage of the discharge. But it is not at all necessary thatthe gas in the tube should be conducting; it might be at an ex-tremely low pressure, still the ends of the wire would be heated ;however, as is ascertained by experience, only the two endswould in such case not be electrically connected through thegaseous medium. Now, what with these frequencies and poten-tials occurs in an exhausted tube, occurs in the lightning dischargeat ordinary pressure.

From the facility with which any amount of energy may becarried off through a gas, Mr. Tesla infers that the best way torender harmless a lightning discharge is to afford it in some waya passage through a volume of gas.

The recognition of some of the above facts has a bearing uponfar-reaching scientific investigations in which extremely liiglifrequencies and potentials are used. In such cases the air is animportant factor to be considered. So, for instance, if two wiresare attached to the terminals of the coil, and the streamers issuefrom them, there is dissipation of energy in the form of heatand light, and the wires behave like a condenser of larger capac-ity. If the wires be immersed in oil, the dissipation of energyis prevented, or at least reduced, and the apparent capacity isdiminished. The action of the air would seem to make it verydifficult to tell, from the measured or computed capacity of acondenser in which the air is acted upon, its actual capacity orvibration period, especially if the condenser is of very small sur-face and is charged to a very high potential. As many import-ant results are dependant upon the correctness of the estimationof the vibration period, this subject demands the most carefulscrutiny of investigators.

In Leyden jars the loss due to the presence of air is compara-tively small, principally on account of the great surface of thecoatings and the small external action, but if there are streamerson the top, the loss may be considerable, and the period of vibra-