MAY 19, 1923 PROCEEDINGS: PHILOSOPHICAL SOCIETY 217 
that the more nearly alike is this atomic environment in the two cases, the 
closer together are the ‘‘atomic radii” of an atom occurring in both. 
Author’s abstract: Laue photographic studies upon potassium and am- 
monium aluminum alums show that the previous X-ray spectrometer data 
(Vegard and Schjelderup) are incorrect and assign an incorrect atomic 
arrangement. A comparison of the spectrometer measurements with some 
spectrographic observations proves that the spectrometer inaccuracies arise 
from the entrance of reflections from other than the principal reflecting face 
into the ionization chamber. This difficulty is pointed out to be one in- 
herent in the spectrometer technique itself so. that the alums thus furnish 
an excellent and clear-cut illustration of the insufficiency of the original 
spectrometer procedure for the determination of the structures of crystals 
using X-rays. 
The unit cube of the correct structure contains four molecules of the 
composition RAI(SO,4)2-12HO:, where R is either K or NH;. The 12 water 
molecules fall into two groups of six each. The hydrogen atoms in the 
ammonium group present an interesting difficulty in the impossibility of 
arranging them into a chemically plausible radical which will possess a 
symmetry in keeping with that of the crystal as a whole. For these crystals 
the corresponding space group is T)°. 
879TH MEETING 
The 879th meeting was held in the Cosmos Club Auditorium on Saturday, 
February 24, 1923. It was called to order by President White with 24 
persons present. 
Owing to some delay at the projecting lantern, opportunity was afforded 
for informal communications. Mr. I. G. Priest presented an informal 
communication on a new variation in the use of the Nicol prism. Dr. W. J. 
Humpueeys presented an informal communication on a correct explanation 
of the diffraction phenomenon commonly known as the “glory” or “‘Brocken- 
bow.” 
Dr. C. G. Petmrs presented a paper on Changes in the index of refraction 
of glass at high temperature. The paper was illustrated with lantern slides 
and was discussed by Messrs. WILLIAMSON, GiIsH, HUMPHREYS, WASHBURN, 
HAWKESWORTH, WHITE, and SosMAN. 
- Author’s abstract: Using an interference method the changes in the re- 
fraction of nine different kinds of optical glass were measured over the 
temperature interval 20° to 650°C. 
The samples made in the form of a plate having two faces parallel were 
placed between two fused quartz mirrors and heated in an electrical furnace. 
The index at any temperature can be represented by the relation 
_ Ny + ANg 
Py No a 
_ Where N> is the number of light waves between the fused quarz plates at 
20°C., Nz the number of waves in glass between the two parallel surfaces of 
the plate at the same temperature, AN, the number of fringes that pass the 
reference mark on the quartz plates, and AN, the number that pass reference 
mark on the glass plate while the temperature is increased to t. 
The index of each glass increased with the temperature until the annealing 
range was reached which was near 500°C. for these glasses. As the tempera- 
ture was still further increased, the index decreased rapidly. The rapid 
