430 Messrs. Mills and Mackey on 
series; having found these sufficient to enable us to draw, 
with fair approximate accuracy, the line of which we were in 
quest*. 
The zinc we employed was prepared electrolytically from a 
solution of highly purified acetate, pressed together, and then 
rapidly melted in a porcelain crucible by the heat of an alcohol 
flame. It was then flattened out between recently polished 
steel surfaces until almost exactly 0'5 millim. thick ; ribbon- 
shaped pieces cut from it, 1 millim. wide and weighing 05 
grm., were coiled in the form of a short helix. Each zinc, 
before immersion in the experimental liquid, was dipped in 
alcoholic potash for three minutes, rapidly washed with water, 
and drained on clear filter-paper; the metal was thus perfectly 
cleaned without material loss. A fresh zinc was used for each 
experiment. 
The hydric sulphate was prepared by fractional distillation 
from an extremely pure purchased sample. We failed to 
detect in it any foreign body excepting water. As the basis 
of our apparatus, we employed a constant-temperature bath, 
designed by one of us some time since (Edinb. Phil. Trans. 
1881, p. 567, to which place we refer for a detailed drawing 
and description of the instrument). Into the central or " con- 
stant " compartment of this bath we introduced, as will be 
seen in the Plate, a glass tube (15 x 155 millim.) closed at one 
end, which always held 5 cub. cent, of the sulphate under 
examination. Thermometer 40803, on which 0*01 C. could 
be read by estimation, was next inserted, and when this 
showed a sensibly constant temperature the zinc helix was 
dropped in. The delivery-cap (see Plate VIII. fig. 1) was now 
slid down over the reaction-tube, its point inserted under the 
collecting-cylinder, and time noted when the first bubble of 
gas came off. The bubbles were in all cases very small. 
In some cases, especially in the first three or four sets of 
experiments, the cessation of the reaction was difficult to 
observe. In these the gas came off sluggishly during the 
first minute, rose to a maximum of delivery-rate during the 
third or fourth minute, and then gradually became slower. 
It is among these groups that the longest times were recorded ; 
and chemical change was considered to have ceased when the 
rate fell below one bubble a minute. In the other sets, on 
the contrary, there was no difficulty in determining the point 
of termination ; action commenced briskly at once, and con- 
cluded with considerable abruptness. It will be observed that, 
as the temperature rises, the time (as might have been ex- 
pected) diminishes. All our experiments were performed as 
* Compare Set TV., exp. 1 in Table I., with the subsequent values of a. 
