84 



«' The cause was obvious, and the defect was attempted to 

 be remedied by procuring instruments in which the issue 

 holes of the burner were very large and close together, and 

 the gas-ways more than adequate to the supply. At length, 

 by repeated trials, an apparatus was obtained which gave an 

 excellent cylinder of dense flame, and the relative dimensions 

 of the parts were thus determined. 



" The advantages of this mode of lighting were obvious : 

 an explosion could never take place ; and the tubes for con- 

 ducting the air from the gasometer, instead of being metallic, 

 might be of thick paper. I constructed lights with paper 

 tubing which answered all purposes. 



" But a sufficiently large gas-holder is expensive and incon- 

 venient ; I therefore thought of adopting a double-bellows to 

 my paper tubing. I procured double, triple, and quadruple 

 bellows, most perfectly executed by an organ-builder, but 

 could not obtain a blast which supplied the burners equably ; 

 the lights rose high and sunk low with every stroke of the 

 feeders; and multiplying the feeders only multiplied the flick- 

 erino-s. At length the object was accomplished by causing 

 a very small double bellows to discharge air into a reservoir 

 bellows of comparatively large dimensions. When the two 

 feeders were worked at a certain rate, and then only, the 

 stream of air was equable, and the gas flames did not flicker. 

 In order to insure this certain rate of action to the feeders, 

 as well as to cause the apparatus to be self-acting, a piece of 

 wheel-work machinery became necessary ; and I soon suc- 

 ceeded in adapting a train of wheels, which, being actuated by 

 a weight and regulated by a fly, worked a crank that gave 

 the exact motion required for the bellows. This train sup- 

 plied the lights for several hours, and, when run down, could 

 be wound up without extinguishing the lights, as it was fur- 

 nished with a maintaining power. 



" Conceiving that the perfection of the invention would be 

 to embody it in the form of a table lamp, which should be as 



i 



