110 PHILOSOPHICAL TRANSACTIONS. [ANNO 1797. 



an electrical discharge through it, no combustion was excited ; but on passing a 

 discharge while the compasses were in contact with the tube, as just mentioned, an 

 illumination and violent commotion were produced, with the rushing up of water, 

 to leave only 4- of the gaz as a residue. On repeating this experiment with 1 mea- 

 sure of atmospherical air and 2 of hydrogen gaz, combustion could not be excited; 

 nor with 2 measures of atmospherical air and 1 of hydrogen ; nor with 2 measures 

 of hydrogen gaz and 1 of atmospherical air ; but oil adding to this last mixture 1 

 measure of oxygen gaz, the electrical discharge produced the phenomena of com- 

 bustion just mentioned, with the rushing up of water, to occupy about -§- of the space 

 which was occupied by the gazes. 



Exper. 3. Having passed 12000 discharges through water, with the apparatus of 

 the preceding experiment, and thus obtained only 4- of a cubical inch of gaz ; and 

 having observed that the quantity of gaz was not greater than it was when only 

 8000 discharges had been passed, and yet bubbles had been seen to be produced on 

 each discharge as copiously, or more so, by the last 3 or 4000 discharges as before; 

 I began to suspect that part of the gaz had been destroyed during the process, or 

 had been absorbed. While I was considering how to account for this disappear- 

 ance of gaz, and was at the same time looking at the tube through which the dis- 

 charges were passing, I observed one of them to be attended with a diminution in- 

 stantly, of about -i- of the gaz produced, and with a slight commotion. I was 

 now sure, from this phenomenon, and from the unequal augmentation of the bulk 

 of the gaz at given times during the process, that combustion had been excited 

 several times before ; not only in the present experiment, but perhaps in the for- 

 mer ones, without observing it. I conceived that a gradual combustion also, very 

 probably, took place in this process, by the kindling of bubbles of gaz in their 

 ascent through the water. I now perceived that the discharges ought to be pro- 

 duced more slowly, or the tubes to be wider, to allow the bubbles to pass quite 

 through the water, in order to avoid the accension of gaz during the process. My 

 calculation also, that 35 to 40000 discharges were requisite to produce 1 cubical 

 inch of gaz from water, containing its usual quantity of common air, was rendered 

 much more vague by this accension, so often liable to be occasioned. To the gaz 

 which remained in the tube in this experiment was added an equal bulk of nitrous 

 gaz; the mixture diminished to 1.5; and on adding to the residue 4- its bulk of 

 oxygen gaz, and passing through it the electrical spark, no accension or diminu- 

 tion of bulk was produced. Hence all the hydrogen gaz and oxygen gaz, pro- 

 duced by the decomposition of the water, had been burnt during the process ; the 

 oxygen gaz thus detected being considered to be only that expelled from the water. 



Exper. 4. By means of electrical discharges, with the apparatus used in the pre- 

 ceding experiment, I obtained gaz from New River water ; letting it up into a re- 

 servoir as soon as about T '^ of a cubic inch was produced, till I had collected -L of 

 a cubic inch. To this was added an equal bulk of nitrous gaz ; on which the mix- 

 ture diminished to 1 .2 ; and on the addition of a little more nitrous gaz, no further 



