118 



On the Critical Point of Mixed Gases. [June 15, 



very regular curve, all the values being below those of Pawlewski's r 

 until a point is reached where the mixture contains about 1 7 per cent, 

 of carbonic acid, within which limit it evidently approaches to Paw- 

 lew ski's values. The reason for this apparent anomaly can onlv be 

 explained by the assumption that a small trace of air or other impurity 

 was present in the tube, for we know from Andrews' and my own 

 experiments, that even the -5-5-oth part of air makes a considerable 

 difference both in the critical point and tension of a gas. 



It is also conceivable that a trace of air may have more effect in a 

 mixture of two gases, than upon either individually, and this would 

 consequsntly complicate matters considerably, when a mixture of 

 several different gases is used. It was principally on this account 

 that tensions of the saturated vapour of the mixture at different tem- 

 peratures was taken, so as to judge of the amount of impurity in the 

 gases, and whether it materially affected the results. 



The curves on Plate I represent these tensions, the ordinates being 

 the pressure in atmospheres, and the abscissa? the temperature in 

 degrees C. The corresponding curves for pure hydrochloric acid and 

 carbonic acid are also shown, but although all the curves for the 

 different mixtures, except one, fall between the two, still the distances 

 are evidently not strictly proportional to the percentage composition, 

 which can only be explained by the presence of a small quantity of 

 air; now, as this impurity must have been infinitesimal, it is inte- 

 resting and curious to see how much it has modified the tensions 

 of the saturated vapour. No. Ill should, of course, have come 

 between No. II and No. IV, and must have had rather a larger 

 amount of impurity than the others, and this is also the case in 

 No. VII, which actually shows a tension higher than that of the 

 most volatile constituent of the mixture, which, of course, is unpre- 

 cedented. 



Having satisfied myself that these apparent anomalies were really 

 due to impurity, I filled another tube with extreme care, which con- 

 tained the same relative proportions of the gases, within "2 of a per 

 cent, as No. VII, and found that it now assumed its proper place 

 below the curve for carbonic acid, the critical point, however,' being 

 scarcely altered at all, showing that an amount of impurity, sufficient 

 to materially modify the tensions of the vapour, had very little effect 

 on the critical point. 



The present position of the question therefore appears to be, that 

 the critical points of mixtures of liquefied gases cannot be expressed 

 by Pawlewski's formula, the maximum difference between his cal- 

 culated value and mine, which occurred in a mixture of equal volumes 

 of the gases, being as much as 3°*6 C. This is '2 or 25 per cent, of 

 the whole difference between the critical points of the two gases r 

 hydrochloric acid being 51°*25, and carbonic acid 31° C. 



