94 OPTICAL PROJECTION 



disarranged by the shock, so as not to resist a second one in 

 the same degree. This has been proved experimentally ; and 

 therefore these chambers should always be examined, and if 

 necessary refilled with fresh-sifted material, after any notice- 

 able ' snap ' of a jet, or at reasonably short intervals even 

 when none may have been noticed. Also it is to be observed 

 that, as the violence of an explosion will depend upon the 

 amount of gas in front of the chambers, these should in all 

 cases be used as close to the nipples as possible. They should 

 never be used at the supply nozzles of the dissolver, but on 

 the nozzles. of the jet; and where this light is habitually 

 used, the safety chamber should be arranged under the nipple 

 itself, with only a very small empty chamber above. Here 

 also it should be constantly examined. This makes the only 

 real * safety ' jet I know ; but the obstruction to the flow is 

 great. This, however, can now be overcome by the higher 

 pressure at command from cylinders. 



The other precaution was to add gas-chambers above the 

 tank one from which the oxygen passed to the surface of 

 the ether, another into which the vapour passed before 

 leaving by the exit-tube. Thus, if the tank were jarred or a 

 little tilted, ether would not run back, but be stopped in these 

 chambers. At the end of 1887, however, another explosion 

 occurred with a tank thus constructed, under circumstances 

 which for a time baffled speculation. The tank was purchased 

 in Feb. 1887, and worked well the rest of that season ; but 

 when got out in Nov. 1877, it behaved differently. Before 

 any pure oxygen was turned on at the jet, the latter heated 

 the lime red hot ; and this got worse and worse, until finally 

 a severe explosion ruptured the tank, in spite of two pumice 

 chambers attached to the dissolver. The probable reasons for 

 the explosion passing these have already been pointed out ; 

 but the explosive condition of the tank itself puzzled everyone. 

 It was carefully examined by several, including myself ; but 

 none could discover any apparent leakage which might have 



