66 Mr. C. Binks on Eleclricity, 



equal to one measure of gas in 80 seconds, during the agita- 

 tion of the water occasioned by refilling the meter ; but after 

 it had become again tranquil the same measure was produced 

 in 95 seconds, or in a length of time greater by about one- 

 fifth. It is almost needless to remark that this increased 

 action is due solely to the mechanical effect of the water when 

 in motion displacing the air-bubbles as rapidly as formed, and 

 more rapidly than they would otherwise be displaced by rea- 

 son merely of their superior levity. 



41. III. Another mechanical effect of a similar kind, but 

 operating to a much greater extent than in either of the pre- 

 ceding instances, is induced by the mere position of the plate 

 from which the gas is evolved : if the surface of a copper 

 plate be placed in a horizontal plane, the gas which is gene- 

 rated on its under side will remain continually attached to the 

 plate ; and should that be the only surface operating, the vol- 

 taic action of such an arrangement would be almost com- 

 pletely impeded; but, on the contrary, if it be placed and 

 used in a vertical plane, then the only causes likely to obstruct 

 the ready dismissal of the gas from the surface are the com- 

 paratively minute ones just referred to in II. 



42. IV. A copper wire which had been stretched in order 

 to straighten it, previously to its being attached to a plate, 

 was found to give a voltaic effect about one-fifth less than 

 another, in all other respects the similar, but which had not 

 been so extended ; but the conducting power of the former 

 was again restored by exposing it to heat. And, again, two 

 wires, in all other respects alike, but which had accidentally 

 been heated in different degrees, gave a marked difference in 

 amount of action ; but had their uniformity of conducting 

 power restored upon exposure to an equal temperature. 



43. V. The same remarks apply equally to plates as to 

 wires ; but there is another cause of variation in the action of 

 plates not previously recognised. In an extensive course of 

 experiments connected with the subject of this paper, in which 

 I had completed about 15 tables of results, each table con- 

 taining, on an average, about 10 observations, I was sur- 

 prised, on looking these tables over, to observe that there was 

 presented invariably (under whatever conditions the experi- 

 ments might have been performed) a remarkably increased 

 amount of action opposite to a number representing one par- 

 ticular copper plate. Since no probable cause could be as- 

 signed for this recurrence, except that of some property pecu- 

 liar to the plate itself, another plate was substituted for it, 

 when it was found that throughout the whole inquiry the 

 plate in question had been giving nearly double its amount of 



