Ramsay and Travers, appears to cut that of oxygen, as the following data show: We do not feel entitled, therefore, to draw any conclusions in the case of argon and carbon monoxide until the critical temperatures have been redetermined. In the curves shown in Fig. 4, the values of γ have been plotted 90° 100° 110° 120° 130° 140° 150° 160° 170° Absolute temperature. 70° 80° Variation of surface tension with temperature. against temperature. The dotted lines indicate a possible extrapolation up to the critical points, which was effected by making use of the formula given by van der Waals, where A and B are constants and m is the reduced temperature. The values given below for A and B were calculated by taking as the critical temperature in each case the value found by continuation of the molecular surface-energy line to cut the temperature axis: 3 Q* VOL. LXXXI. The formula y = A(1 —m)3 gives an independent method of smoothing experimentally determined values of y, as it indicates a linear relation between logy and logT. Values obtained by this method, from our experimental results for oxygen, agreed very closely with those obtained by the method previously employed. During the course of this investigation, L. Grunmach has published measurements of the surface-tension of liquid air of varying composition, using the method of capillary ripples. The temperature was measured by a petroleum-ether thermometer, and the experiments were all made at a mean temperature of 190.3°. We have compared his results with our own on the assumption that the surface-tensions and densities of mixtures of the two liquids can be calculated with fair approximation from those of the constituents by the simple law of mixtures. This assumption is justified, in general, for liquids which do not influence each other's molecular state, by the results obtained by Ramsay and Aston. The following table exhibits the comparison : The agreement is not particularly good. No details are given by Grunmach with respect to the variations of temperature from the "mean" value, - 190.3°. Verschaffelt (communications from Leiden, No. 18) has measured the surface-tensions of liquid carbon dioxide and nitrous oxide at various temperatures, using the method of rise in capillary tubes. His results may be summarised as follows: In both cases, the molecular surface-energy line cuts the temperature axis at points lying about 6° below the corresponding critical points. Verschaffelt points out that the values found by him for the A and B of van der Waals's formula agree approximately with those found for a number of other liquids, such as ethyl ether, ethyl acetate, benzene, chlorobenzene, and carbon disulphide. This is not the case with the values of A and B, calculated from our results, but as van der Waals's formula is practically only an empirical formula, and as our measurements apply to temperatures which are very much lower than those at which the measurements of Verschaffelt and other previous observers were made, the divergence in the values of A and B is not surprising. In conclusion, we wish to express our best thanks to Prof. Ramsay for the interest which he has taken in our work. CHEMICAL LABORATORY, UNIVERSITY COLLEGE, W.C. |