% § 2. Previous Determinations of the Isotherms of Hydrogen at 0° and 20° C. The isotherm of hydrogen at 0° C. has been determined at Leiden by Kamerlingh Onnes and Braak ;* and at the Physikalisch-Technische Reichanstalt, Berlin, by Holborn.f The Leiden measurements, which extend only to 50 atmospheres, are open to some objection. As is the general practice in the determination of isotherms at Leiden, the volumes occupied by one and the same quantity of gas at different pressures are measured,

more »

... so that the probable * * Phys. L ab.,' Leiden, Commn. N o. 100b. f * Ann. d. P hysik,' vol. 63, p. 674 (1920). on July 20, 2018 http://rspa.royalsocietypublishing.org/ Downloaded from on July 20, 2018 http://rspa.royalsocietypublishing.org/ Downloaded from 554 T. T. H. Verschoyle. (piezometer stem, small reservoir, and connecting steel capillary) were all directly calibrated by mercury, and it bas recently been pointed out by Penning'1 ' that small spaces of unknown volume may exist in the soldered joints between the couplings and the glass capillaries to which they are affixed. All the measurements at Leiden with divided piezometers are liable to objection on this ground (except where some suitable check has been afforded), more especially when the density of the gas in the small reservoir is not much larger than that in the stem, as is the case here. Moreover, only four points were measured, so that it is difficult to form any idea of the probable errors involved. The determinations of pressure and temperature were, as is always the case at Leiden, unexceptionable. The method in use at Berlin for the determination of isothermsf differs from the Leiden method in that the volume into which the gas is compressed is con stant, the quantities of gas used thus increasing with the pressure. The sensitivity of the measurements therefore does not decrease with increasing pressure, and, as comparatively large quantities of gas were used (the con fining volume was about 110 cm.3), the volume observations, cet. par., should be extremely accurate. Normal volumes are measured only after the observa tions at high pressures, in a series of flasks contained in a large water-bath. Pressures are measured on a pressure-balance fitted with a mechanism for imparting an oscillatory motion to the piston. The balance was calibrated with the help of an open manometer, so that the pressure determinations should also be very satisfactory. It appears from the results that the errors of observa tion seldom exceeded 1 in 4,000, and, up to 100 atmospheres (the highest pressure reached), the isotherm at 0° C. is undoubtedly very accurate. The isotherm at 20° C. has been repeatedly measured at Leiden. The determination by Schalkwijk, J using an undivided piezometer, stands in a class by itself, for it was effected with extraordinary care and precision; unfortunately it reaches only to 60 atmospheres. Kamerlingh Onnes, Crommelin, and Smid § have extended the isotherm at 20° C. to 100 atmospheres, and the pn-value found for the highest pressure differs by 1 in 700 from that derived from an extrapolation of Schalkwijk's equation. The lowest pressure measured was 65 atmospheres, and the pv-value found differed only by 1 in 3,000 from Schalkwijk's value. Since, however, the differences all lie in the same direction, * * * § * * Phys. Lab.,' Leiden, Commn. N o. 166. t Holborn and Schultze, * Ann. d. P h ysik ,' vol. 47, p. 1089 (1915). Downloaded from and Walstra extended this isotherm to 1,000 atmospheres, but, for several reasons, their results are not so accurate. § 3. The determinations of the isotherm of nitrogen at 09 C., both by Kamerlingh Onnes and van Urk, up to 60 atmospheres, and by Holborn and Otto, up to 100 atmospheres, are reliable; but the measurements by Smith and Taylor are open to serious objection. The isotherm at 20° C. has been directly on July 20, 2018 http://rspa.royalsocietypublishing.org/ Downloaded from 576 Hydrogen, N i t r o g e n , and Hydrogen-Nitrogen Mixture