74 Mr. C. Binks on Electricity ^ S^c, 



show the precautions which have been observed to ensure ac- 

 curacy, as to avoid the necessity for a continual reference to 

 such details afterwards. 



68. The effects now to be sought for, are estimated gene- 

 rally by the length of time needed to produce a certain mea- 

 sure of gas, or for the expenditure of a certain weight of zinc. 

 But it has been shown in some former experiments that the 

 quantities of zinc expended and of hydrogen produced by it, 

 (under certain conditions of voltaic action,) are not always 

 equivalents of one another. In my former paper* a case of 

 this kind is noticed, in which a small battery when used to de- 

 compose water, lost zinc, in quantity one third greater than 

 the equivalent of the hydrogen evolved. In the investigation 

 that now follows no uncertainty will be permitted to remain 

 upon this point; and the labour occasioned by such exami- 

 nations will be found to have been not altogether without use. 

 (See section 9th.) The measure of gas most commonly used 

 is one tenth of a cubic inch, and the time is taken in seconds. 

 With small plates, this measure will be yielded in a length of 

 time ranging between 30" and 900" according to the strength 

 of the acid or the distance of the plates. The moment at 

 which this measure is completed by the meter, may be deter- 

 mined with certainty within two or three seconds of the real 

 time. A closer approximation than this is seldom attempted 

 in the following experiments, nor is it in any respect neces- 

 sary. The difference between 30 seconds and 900 seconds is 

 an extreme one, and but seldom occurs : most commonly the 

 difference is much less marked, and in some instances is so 

 minute, that no other method, that I am aware of, than that 

 now proposed is adequate to its detection. 



69. The capability of this method to detect minute differ- 

 ences in quantity will be apparent upon a little consideration 

 from what has been already said ; but the following compai'i- 

 son between it and that afforded by an ordinary magnetic 

 galvanometer, will serve to mark that capability more clearly. 



70. The galvanometer here used, though of the common 

 construction, was an exceedingly good one, by Newman ; 

 and such an instrument in all respects as would have been 

 employed in the kind of experiments now made had the indi- 

 cations afibrded by the needle been desired. 



71. The instance selected is an average one of the effects 

 to be estimated. A small arrangement was used at first with 

 its plates one fourth of an inch apart, and afterwards at the 

 distance of thirty-eight inches. At the first position the one 



• Phil. Mag., p. 88, July 1837. 



