264 Dr. Collie and Prof. Ramsay. Behaviour of [Feb. 13, 



of tinfoil, through the air. The stopcock H was then shut, and by- 

 pressing down the little reservoirs on the capillary tabes, the latter 

 broke at the scratches, and each (J -tube became filled with the gas 

 contained in the reservoir placed over it. The stopcock above H 

 was then shut, and by opening H a charge of gas entered the 

 space between the two stopcocks. H was again shut, and the 

 upper stopcock opened (D being shut), and the charge of gas 

 filled the vacuum tube, being dried by passage through the pent- 

 oxide contained in F. It should here be remarked that the gases 

 underwent a preliminary drying by contact with fragments of solid 

 potash in the little reservoirs at K. The charge of gas in the 

 vacuum tube was, after a short time, removed by pumping until 

 its spectrum disappeared and the vacuum tube ceased to conduct. 

 This process was repeated several times as a precautionary measure. 



The space between the stopcock H and the one above it held 

 0*066 c.c. ; hence it was possible to introduce that amount or any 

 multiple of it into the vacuum tube, and when required, by open- 

 ing the stopcock communicating with the pump, to remove five-sixths 

 of that amount, leaving 0*011 c.c. in the vacuum tube; and as this 

 could be done with either gas, it was possible to alter both quantity 

 and pressure at will. 



The first set of experiments was made with helium and hydrogen. 

 The pump and vacuum tube were empty after washing with helium. 



1. 0*066 helium ; 0'066 hydrogen in 57*5 c.c. 

 P = 1-74 mm. 



The hydrogen red was brilliant, the blue strong, the violet 

 visible. 



The helium red faint, the yellow visible, the green faint, the 

 violet doubtful. 



2. 0-066 helium; 0*132 hydrogen in 57*5 c.c. 

 P = 2-61 mm. 



Helium spectrum invisible. 

 Volume increased to 347*5 c.c. 



P = 0*43 mm. Helium spectrum strong; reds visible. 



P = 0*072 mm. Helium spectrum very strong ; reds comparable 



with hydrogen red in intensity. 

 P = 0*012 mm. Helium relatively stronger. 

 P = 0*00033 mm. Both very faint, but still visible. 



It is thus seen that while at moderately high pressure one part of 

 helium in two parts of hydrogen is not detectable, on reducing pres- 

 sure the brilliancy of the helium increases relatively to that of hydro- 

 gen. These spectra were examined with direct-vision prisms: the 

 narrow part of the vacuum tube served for slit. The great advantage 



