Vol. 8, 1922 PHYSICS: R. C. WILLIAMSON 255 
the commonly accepted value, 0.862 gm./cm.^ partly no doubt on account 
of the low temperature of the crystals, and partly perhaps because the 
values usually quoted have been obtained for metal not perfectly crystal- 
line, to judge by the X-ray evidence. As regards the purity of the potas- 
sium here used, it can only be stated that the original material was pur- 
chased as "c. p." from Eimer and Amend, was dried and freed from the 
usual oily surface layer, and was then melted and double-distilled in vacuo 
using only pyrex glass. The samples were prepared for me under the 
supervision of Dr. H. E. Ives from material used by him in the preparation 
of photoelectric cells. His suggestion that a change in the crystalline 
condition of the metal at low temperatures might be responsible for certain 
anomalies in photoelectric emission was the starting point of this research. 
The observed crystalline structure does not persist when the tempera- 
ture is allowed to rise again to about 20° C. No effort has been made to 
determine the highest temperature at which crystallinity can still be 
detected, since it seems probable that the loss of structural symmetry is 
not abrupt but gradual. 
1 A. W. Hull, Physic. Rev., (2) 10, 661-696 (Dec, 1917). 
2 L. W. McKeehan and P. P. Cioffi, Ibid., (2) 19, 444-446 (April, 1922). 
PRELIMINARY REPORT ON 
THE IONIZATION OF POTASSIUM VAPOR BY LIGHT 
By R. C. Williamson 
Department of Physics, University of Wisconsin 
Communicated June 13, 1922 
As Hughes states in a Bulletin of the National Rksearch Coun- 
cil,^ experimental results relative to the ionization of metallic vapors by 
light, have been very meagre, and it is difficult to point to any phenomena 
which are due, without doubt, to ionization by radiation of optical frequen- 
cies. Steubing^ describes ionization of mercury vapor by light transmitted 
through fused quartz, but his results seem probably to be explained as 
ionization in the vapor by collision of photoelectrons from the electrodes. 
Anderson^ and Gilbreath^ have published results obtained in potassium 
vapor which seem to be due to wall emission. Kunz and Williams,^ in a 
brief abstract, state that while caesium was not ionized by radiation of 
wave-lengths greater than 3130 A, a marked effect was secured at 2530 A. 
No details as to the method have as yet come to the writer's notice. 
The equation Ve = hv, where V is the voltage corresponding to the 
