422 
Proceedings of the Royal Society 
The precipitate purpureo-cobaltic chloride was collected on a filter, 
well washed with very dilute hydrochloric acid, and dried carefully, 
great care being taken to avoid contamination by iron. The dried 
purpureo-cobaltic chloride was then reduced in a current of hydro- 
gen in small reducing crucibles, and fine skeletons of spongy 
cobalt were obtained. 26 gms. of cobalt were made in this way ; 
the remainder of the purpureo-cobaltic chloride was heated to 
expel chloride of ammonium, dissolved in hydrochloric acid and 
water, and when evaporated to dryness, yielded 32 gms. of chloride 
of cobalt. This was made up to 900 c.c. with distilled 'water.” 
Having obtained these, some experiments were then made as to 
the substitution of aluminium for the carbon-electrode used by 
Mr Beardslee, and as to the strength of current required. The 
aluminium was found to answer admirably. These preliminary ex- 
periments over, we set about preparing such a bar as would serve us 
instead of a wire in our thermo-electric work. For this end 
we took a strip of aluminium about 100 mm. long, and 8 mm, 
broad, carefully covered it with wax on one side, and used it as 
the negative electrode. To form the positive electrode a piece of 
glass tube, 10 mm. wide, and 20 mm. long, was taken and closed 
at one end, through which a platinum wire was passed ; the one 
end of this was connected with the positive pole of the battery, 
while the other end was coiled into a spiral inside the tube, 
into which some of the spongy cobalt was tightly packed so as to 
make good contact with the wire. These electrodes were placed 
in a porcelain bath containing some of the chloride of cobalt solu- 
tion, rendered very slightly alkaline by the addition of ammonia. 
The battery found most suitable consisted of two large Bunsen 
cells ; when less was used the deposition went on too slowly, 
while pieces prepared with a stronger battery-power were inco- 
herent. The action was usually allowed to go on for two or three 
days, at the end of which time the metal had a considerable thick- 
ness. The cobalt was separated from the aluminium by boiling it 
in a strong solution of caustic soda. The pieces of cobalt thus 
prepared were of very great hardness and compactness, being with 
difficulty filed. 
It was with these pieces of cobalt that the thermo-electric obser- 
vations were made. The firsi observations showed that the cobalt 
