( 618 ) 
oxygen isotherms with the aid of the law of corresponding states. 
Subsequently, when we could avail ourselves of a portion of our 
experimental data, the expansion coefficients or rather the densities 
which are in this case easier to manipulate were graphically deter- 
mined. The degree of accuracy afforded by this method was found 
to be quite sufficient for our purpose. The second approximations 
were found to be sufficient for all the isotherms. 
In the calculations we have made no use of the Ramsay and 
Travers ') isotherm at + 11°.2 C., for their observations shewed such 
marked deviations from ours that we could place no reliance upon them. 
$ 5. The results are contained in the following tables. In the first 
column is given the date, in the second the series, and in the third 
the number of the observation; in the fourth is given the temperature 
in Kelvin deerees 7) — 273°.09; the fifth and sixth contain. the 
pressure in atmospheres (p) and the density in terms of the normal 
density respectively, the results gives the produet of p into the 
volume expressed in terms of the normal volume pv, and finally 
the eighth contains the values of vy. 
Such observations as were for any reason deemed less accurate 
than the others are placed between square brackets. 
Observations marked by an asterisk were made by way of control 
after the measurements proper were completed because uncertain 
indications of one of the metal manometers made us for a moment 
suspect that something had happened to the pressure measurement. 
The table of deviations given in a subsequent section shows that 
this fear was baseless. 
§ 6 A general survey of the region which the present isotherms 
in connection with earlier papers’) enable us to regard as known 
is given by the accompanying pva diagram for all observations *). 
Observed points are represented by small circles. For data for the 
boundary curve we may refer to the other papers just mentioned. 
Points calculated from the liquid and vapour densities are given by 
squares, and points extrapolated from the isotherms are given by 
trianeles °). 
1) W. Ramsay and M. W. Travers. Phil. Trans. (A) 197, 47, 1901. 
2) Proc. Dec. 1907. Comm. N° 101b. 
3) Proc. May 1910. Comm. 115, Proc. Dec. 1910, Comm. NO. 118a. 
Higa, 
5) In order to make the diagram not uselessly long, the volumes above 0.040 are 
omitted. The isotherms of + 20.39 C., 0°.00 CG. and -— 570.72 C. therefore ought 
to have been drawn until the line v4 = 0.040. 
