Chemistry and Physics. 603 



Since certain necessary data were lacking in the literature of 

 the subject, H. Augustin has recently undertaken the experi- 

 mental determination of the quantities needed. The first problem 

 was to find the value of the density of liquid hydrogen at a known 

 temperature. The experimental method consisted in ascertaining 

 the loss of weight of a quartz sinker immersed in the liquefied 

 gas. The volume of the quartz cylinder and the apparent loss of 

 weight in liquid hydrogen were about 23*3 cm3 and 1*6342 g. 

 respectively. These numbers may give some idea of the dimen- 

 sions of the apparatus employed and of the degree of accuracy 

 attainable. Because of the low temperature of the liquid a num- 

 ber of experimental difficulties had to be overcome. The final 

 form of apparatus used was too complicated to admit of detailed 

 explanation in this place. Suffice it to sa}*- that the apparatus was 

 ingeniously designed and that the data obtained by its aid seem 

 to be very reliable. The final value for the density of liquid 

 hydrogen is given as 0*07105 gram per cm 3 , with a mean error of 

 six units in the last decimal place. (The corresponding pressure 

 and temperature were 745*52 mm and -252*83° C.) 



The next problem was the determination of the indices of 

 refraction of hydrogen and of nitrogen in the liquid state. These 

 data were found by E. Wiedemann's method of total reflection. 

 By using monochromatic light and a Gauss eye-piece, Augustin 

 was able to make accurate settings of the cross-hairs on the well- 

 defined boundary between the light and dark regions of the field 

 of view. Earlier investigators who employed the same general 

 method did not isolate the colors and hence they obtained only 

 mean values of the indices of refraction. The radiations used by 

 Augustin had the following wave-lengths : 404*7, 435*9, 546*1, 

 579-0, and 656*3/x/x. The first four lines are emitted by mercury 

 and the last by hydrogen. The corresponding indices of refrac- 

 tion of liquid hydrogen relative to hydrogen vapor were found to 

 be 1*11262, 1*11179, 1*11003, 1*10974, and 1*10924 in the order 

 named. The relative indices for liquid nitrogen are given as 

 1*20125, 1*20010, 1*19788, 1*19746, and 1*19715 respectively. 

 When the refractive indices are laid off as ordinal es with the 

 wave-lengths as abscissae the points fall almost exactly on smooth 

 curves, thus indicating a high degree of relative accuracy of the 

 data. 



In order to calculate the form-number, u, mentioned in the first 

 paragraph, it was necessary to use the questionable process of 

 extrapolation to obtain the indices of refraction of the vapors of 

 hydrogen and nitrogen at the respective boiling points, since the 

 formula involves the absolute indices of the liquids. For hydro- 

 gen the values of u corresponding to all the colors except the 

 violet were found to be 2*02, 2*02, 2*04, and 1*90, giving a mean 

 value 2*00 i 0*02. It thus appears that the molecules of hydrogen 

 behave dielectrically as if they were true spheres. The author 

 remarks that this result (u = 2) is especially noteworthy because 



