Decomposition of Water by the Galvanic Battery. 95 



time. I have found that when the battery is not very powerful, 

 the briUiant hght comes only from the positive coke point. For, 

 if a coke point and a copper point be used instead of two coke 

 points, and if the coke point be connected with the positive end 

 of the battery, the light is as brilliant as if two coke points were 

 employed. But if the coke point be connected with the negative, 

 and the copper point with the positive end, the light will be 

 worthless. I forgot to make this experiment when I had occa- 

 sion to use a very powerful battery. I may mention here that 

 with 60 cast-iron cells in which the zinc plates were 2 inches 

 by 4, I was able to fuse and deflagrate a round piece of steel 

 \ inch diameter. Although I made a far greater number of 

 experiments on the lime light than on the coke light, I am not 

 able to state positively the size of the battery, which, by decom- 

 posing water, would produce a lime light equal in illuminating 

 poorer to that of a pair of coke points ignited by a battery of 

 40 cast-iron cells in which the zinc plates are 2 inches by 4. 

 I can only say that I think it probable that a battery twice, or 

 very nearly twice the size, would be required. With regard to 

 the length of time the two batteries would work with a given 

 charge, I am equally uncertain. In decomposing water, all the 

 cells must be arranged in rows of four each, so that the intensity 

 of the current may not exceed that of 4 cells. Now, when a 

 current is sent from 4 cells through a good conducting fluid, 

 the quantity of electricity which passes through the fluid and 

 through each cell of the battery is only about one-half of the 

 quantity which would pass if the ends of the battery were con- 

 nected by a short, thick wire. This I have found by the galva- 

 nometer. Hence the battery would work about twice as long in 

 the former as in the latter case. ^\'lien the coke points were 

 ignited by 40 cells, I measured the angle of deviation of the 

 sine galvanometer, and found it to be 31° when the needle was 

 in the axis of the coil, and nearly 10 inches from its centre. 

 When they were ignited by GO cells, the deviation was 40°. 

 Unfortunately, I did not measure the angle of deviation which 

 would be produced by the current passing directly through the 

 coil of the galvanometer. The experiments were made before 

 my class, and I intended to measure on the next day the angle 

 of deviation wliicli would be produced by the current from 4 cells 

 of the same size, sent through a fluid. I was prevented from 

 making the experiment until I had occasion to remagnetize the 

 needle, and consequently I cannot now depend on its indications 

 being the same tliey would have been on the day the battery of 

 40 and 00 colls was used. I have since tried the ettect of 

 a current from four of the cells on the sine galvanometer after 

 passing it through a solution of carbonate of soda, and found 



