435 
a peculiar apparatus which I described. ‘This I have done, 
though not as far as I wish. The plan of closing the bell which 
is mentioned there, failed on trial; and I used asa valve a disc 
of iron, haying in its centre a convex projection fitting the 
hole, round which is a thin washer of vulcanized caoutchouc. 
This is kept tight by a screw, the head of which can be caught 
by a hook on the sliding rod of the receiver, and the bell may 
thus be raised (for I found its flotation uncertain). The bell 
holds 19 inches, and the little jar which I use to transfer gas 
into it holds ~4, so that the density when it is in its normal 
position =;5'-7- At first some air adheres obstinately to the 
bell, and some escapes from the pores of the iron core, in spite 
of varnish; but after repeating the process three or four times, 
no trace of it appears. The distance between the iron disc, 
and the surface of mercury which is the lower electrode = 5''5. 
‘©16. When no gas is introduced, I presume the vacuum 
contains nothing but mercurial vapour; with Smee the dise 
was covered with a film of pink ; then there were broad, bright, 
green, spherical shells extending across the bell to a radius of 
15. From them a faint haze spread down in a cone, so as 
actualiy, or almost, to meet a similar cone rising from a stratum 
of bright, blue light (in which no stratification was visible), 
floating on the mercury. The whole is so like the appear- 
ance in hydrogen, that I feel almost sure the latter is a gaseous 
metal. ‘There was no fluorescence; but as I had learned 
from (2) that it might be concealed by the absorption of the 
glass, I lately repeated this experiment, placing on the inside 
of the bell’s dome a spot of sulphate of quinine, and one of 
platino-cyanide of potassium. When the disc was positive, so 
that these tests were enveloped in the green menisci, they 
showed no fluorescence whatever; when it was negative, and 
they were exposed to the blue light, the platino-cyanide 
showed the faintest possible trace; the other none. This 
vacuum is a worse conductor than the external air one, 0-15. 
“17. Introducing the jar with air, which gives the vacuum 
