PHYSICS: A. H. PFUND 
521 
present theory, and no diagram of H3 can be made without in one 
manner or another spKtting the single unitary valence bonds into 
fractions, as for instance is possible with the conception of J. Stark. 
H3 has been shown to be consistent with Bohr's model of the hydrogen 
atom. But the chemist will require a more definite picture to show 
the structure of this compound and to reconcile its very ordinary 
existence and properties with his theoretical conceptions. Hydrogen 
seems no different from oxygen in this regard. Indeed, it is probably 
only the injunction due to the valence theory which has prevented the 
observation of this variety of hydrogen at a much earlier time than the 
present. 
1 Rays of Positive Electricity, Longmans, 1913. 
2 Duane and Wendt, Ithaca, Physic. Rev. 10, 1917, (116-128). 
3 S. C. Lind, Easton, Pa., J. Amer. Chem. Soc, 41, 1919, (545). 
* F. L. Usher, London, J. Chem. Soc, 97, 1909, (400). 
6 Irving Langmuir, /. Amer. Chem. Soc, 34, 1912, (860, 1310); 36, 1914, (1708); 37, 1915, 
(417). 
« A. J. Dempster, London, Phil. Mag., 31, 1916, (438-43). 
A VAPOR-FREE VACUUM SEAL 
By a. H. Pfund 
Department of Physics, Johns Hopkins University 
Communicated by J. S. Ames 
The behavior of stop-cocks in vacuum manipulations is, as a rule, 
quite satisfactory so long as the degree of exhaustion is not very high. 
However, when the utmost vacuum must be reached and maintained, 
the ordinary stop-cock is useless. Quite apart from its tendency either 
to leak or to fill up with grease, its main defects are: 1. Evaporation 
of volatile constituents of the grease into the evacuated chamber. 2. 
Comparatively small diameter of opening. Another element has been 
brought to light by Shrader who has recently shown (Physical Review, 
Ithaca, June, 1919, p. 434) that the gases liberated from the seal, 
when a glass bulb is sealed off the pump, increase the pressure several 
hundred per cent over that which had existed previously. An attempt 
has been made to overcome these difficulties by designing a vacuum 
seal of the type shown in sketch. 
The apparatus takes the form of the well-known trap, slightly 
modified. A hollow cylinder of platinum (A), either seamless or made 
