Vol. 6, 1920 
ASTRONOMY: H. SHAPLEY 
293 
to us for the first time the striking dissimilarity in chemical behavior be- 
tween these two types of isomeric compounds. 
It seems safe to conclude, therefore, from the results of this new work 
that any isothiocyanate formed by interaction of potassium thiocyanate 
with a secondary chloracetanilide will be found to be unstable and easily 
transformed into its corresponding thiohydantoin. Whether Beckurts 
and Frerichs were actually dealing in their work with such normal thio- 
hydantoins instead of the pseudo modifications must be decided by fur- 
ther investigation. In the case of tertiary anilides of isothiocyanacetic 
SCN.CH2CO.NR2, such cyclic rearrangements cannot take place, and in 
such cases we expect to obtain acyclic combinations containing the group- 
ing NCS functioning as a true isothiocyanate. Investigations dealing with 
various phases of this interesting problem are now in progress and the re- 
sults will be published in the Journal of the American Chemical Society. 
1 Wheeler and Merriam, /. Amer. Chem. Soc, 23, 1901 (283). 
2 Amer. Chem. J., 28, 1902 (121) ; Johnson, /. Amer. Chem. Soc, 25, 1903 (483). 
^ Arch. Pharm., 253 (233); Chem. Zentrhl., 2, 1915 (614); Chem. Abstracts, 10, 1916, 
(888). 
4 Aschan, Ber. chem. Ges., 17, 1884 (420); Marckwald, Neumark and Stelzner, Ibid., 
24, 1891 (3278). 
^ Loc. cit. 
STUDIES OF MAGNITUDES IN STAR CLUSTERS. XI. FRE- 
QUENCY OF CURVES THE ABSOLUTE MAGNITUDE AND 
COLOR INDEX FOR 11 52 GIANT STARS. 
By Hari^ow Shapley 
Mount Wilson Observatory, Carnegie Institution of Washington 
Communicated by G. E. Hale, April 8, 1920 
Kapteyn, Schwarzschild, Parkhurst, Russell, and other investigators 
of statistical problems in stellar astronomy have paid special attention 
to the relative frequency of successive values of the absolute brightness 
of stars (the luminosity law) and the relative frequency of different 
spectral or color types. The laws of the frequencies of absolute magnitude 
and spectrum are indeed fundamental in studies of stellar evolution and 
the arrangement of stars in space; but to determine these laws, at the 
same time keeping the errors due to unavoidable and vitiating selection 
of data at a minimum, is by no means a simple process. Insufficient 
knowledge of stellar distances, and frequently of apparent magnitudes 
and spectra as well, presents a serious obstacle. In particular, it is 
difficult to obtain satisfactory luminosity curves for each spectral type, 
or representative spectral curves for small and clearly defined intervals 
of absolute magnitude. The luminosity curve that does not differentiate 
