IX THE PRESENCE OF DUST-FREE AIR AND OTHER GASES. 
291 
possible that an exceedingly small trace of air would be sufficient to cause the slight 
condensation which is observed. It must be remembered that even if only one drop 
separates out it will be seen. Colours like those observed in the other gases made 
their appearance when the expansion exceeded that required to produce the dense 
condensation. Measurements, however, were not made of the expansion required to 
produce a given colour. 
Expansion required to produce Dense Condensation in Hydrogen. 
Date. 
B. 
d 
c 
TT. 
Pv 
Pr 
Pi- 
Pi- 
Pl/Pi — 
Aug. 4. 
763 
22-0 
20 
670 
1412 
283 
1025 
1-378 
„ 6. 
767 
20-5 
18 
661 
1409 
267 
1015 
1-.388 
„ 6. 
767 
21-0 
18 
652 
1400 
269 
1017 
1-377 
» 6. 
767 
210 
18 
653 
1401 
269 
1017 
1-378 
Mean 
1-380 
Chlorine. 
This was prepared by heating hydrochloric acid with potassium bichromate. 
The apparatus is shown in fig. 4. The U-tubes contained water. A supply of the 
gas was collected in the tube SP, by allowing a rajiid stream to pass for about three- 
quarters of an hour, and then closing the stopcock, and sealing ofi‘ at P with the 
blow-pipe. 
Fig. 4. 
The tube was then cut at S. The part from S to P was then removed, and the 
open end joined to L (fig. 2) in place of the stopcock Tg. The chlorine was then 
pumped into the expansion apparatus when required, the reservoir B (fig. 2) with the 
tubes connected to it being of course first completely filled with well-boiled water. 
The tube SP was throughout protected from the light by a wrapping of black paper. 
2 p 2 
