390 



ScienHfic Frocpedings. Jini/al Diihlin Societ//. 



volume V^ of tlie gas J, witli volume Vs of gas B.) A similar equation 

 is approximately tl'ue for tlie density. An indirect confirmation of this 

 law follows from some observations recorded later. 



Tlie experiments with carbon dioxide are represented in tlie curves 

 Nos. 14 to 16 (Plate XXXIII.). Subsequently a mixture of tliis gas with 

 air was made by admitting some air. Curve No. 17 was obtained while 

 using the mixture. 



Perliaps it will be simplest to compare each curve witli the corresponding- 

 one for air. Tliis is done in the following table. 



Tahi.k IV. 



The first seven columns give the particulars of the CO2 experiments. In 



column 8, the values of - — = K are given ; in column 9, the values 



of K ior air. In the case of the tliird tube (3'8 cms.) the numbers for air 

 were taken immediately after those for the CO. — air mixture ; and they are 

 given completely on the third line of the table. For the copper tube, the 

 two values of IC differ by 3 per cent. ; for the narrower glass tube, by 4 

 per cent. ; and for the wider tube the difference is less tlian i per cent. 

 The gas mixture was analysed by drawing ofi about 100 ccs. through the 

 tap into a Hempel gas-burette; after measuring tlie volume of the 

 mixture, the COj was absorbed in a potash bulb; and tlie volume of air 

 remaining was measured. The mixture contained almost exactly 60 per cent, 

 of air by volume. The viscosity was calculated by the formula (5) above. 

 The density by the corresponding formula, 



P" ~ V ^ r ' 



'' A ' ' B 



which gave results sufficiently accurate for our purpose, 



(6) 



