of Copper to the Measurement of Electric Currents. 181 



and made to revolve rapidly, while the plate is held lightly 

 against it by means of a pad. This avoids touching either 

 the paper or the part of the plate to be put in the liquid with 

 the fingers. A clean circular brush or a soft silk or velvet 

 pad at the end of the cylinder, and revolving with the paper, 

 is Convenient for cleaning off any dust or loose glass from the 

 surface of the plate after it is polished. A set of five or ten 

 plates may be cleaned in this way in a few minutes, and never 

 fail to give a firm deposit and accordant results. The method 

 of washing and drying the plates after they are removed from 

 the cells, which was recommended in my former paper, has 

 been adhered to throughout. As the plates were removed 

 from the cells they were rinsed two or three times in clean 

 water very slightly acidulated, and then placed in a shallow 

 glass tray containing pure water. From this they were 

 almost immediately taken, dried first in a clean pad of blotting- 

 paper, and finally in front of a bright fire, care being taken 

 not to heat the plates much. 



The perfect regularity of the results of experiments such 

 as those here referred to is greatly interfered with by the 

 accidental inaccuracies of small amount which are almost 

 unavoidable. The total difference of weight between one 

 plate and the next of a series did not as a rule amount, at 

 ordinary temperatures, to more than a few tenths of a milli- 

 gramme ; and this was subject to error, which might be com- 

 parable with its total amount, from inaccuracy in weighing, 

 from want of perfect insulation between the cells, from slight 

 oxidation of the plate in the drying, and so on. The results 

 of the individual experiments were, however, in very good 

 agreement; and I believe the mean results as given in the 

 tables and curves below will be found to be very close to the 

 truth. 



An inspection of the tables and curves will show that the 

 effect of temperature, especially when that reaches about 30° 

 Centigrade, is very important. The results become so sensi- 

 tive to difference of temperature at about 35° C, that a 

 difference of a degree or two between the different cells pro- 

 duces marked irregularity in the curve. At the ordinary 

 temperature of a laboratory in this country, say from 10° C. 

 to 15° C, the effect of difference of temperature is not im- 

 portant, and the numbers given for 12° C. in the table will be 

 found near enough for most purposes. In hot climates, how- 

 ever, where the temperature of the laboratory may be 35° C, 

 or even more, the variations of temperature would require to 

 be very carefully attended to. 



Some experiments made for the purpose of testing whether 

 PHI. Mag. S. 5. Vol. 25. No. 154. March 1888. ' 



