412 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Figure XI was devised. The protuberances of the bulb I, instead of 

 being provided with platinum terminals, were left open, and to them 

 rubber tubes were connected. The other ends of the rubber tubes were 

 connected to reservoirs containing mercury. To the top of the bulb I 

 was fused a capillary tube of uniform bore closed above by a stop-cock. 

 The vessels B B were stationary, and the bulb I could be lowered with 

 stop-cock open so as to fill with mercury. The bulb I was then raised 

 with stop-cock closed to a height greater than the barometric column, 

 and was thus exhausted. Traps T T in the protuberances of I pre- 

 vented air bubbles, which escaped from the rubber tube, from entering 

 the bulb. The bulb, which could thus be exhausted to any desired 

 degree, was connected through the mercury columns in series with the 

 primary of a Tesla coil, of which the discharge at the secondary served 

 as a test of the correctness of the vacuum. The vacuum could be 

 measured by lowering I with stop-cock closed, and bringing the residual 

 gas under atmospheric pressure into the capillary C. It was found that 

 with the particular voltage at my disposal (15,000 volts), the Tesla coil 

 gave its best action when the pressure in the cold bulb, before the dis- 

 charge, was about .02 mm. When the pressure was two or three times 

 this amount the bulb gave a brilliant arc, while the spark at the terminals 

 of the secondary was feeble. For pressures lower than .02 mm. (cold) 

 the bulb showed pale green luminescence resembling somewhat the glow 

 in a Roentgen tube. Under these circumstances the condensers seemed 

 not to discharge. 



Jefferson Physical Laboratory, Harvard University, 

 Cambridge, Mass., Dec. 20, 1903. 



