224 
ASTRONOMY: H. SHAPLEY 
for bacterial affinities in some ascomycetous and sphaeropsidaceous forms, 
the hyphae of which are similarly very minute. It is doubtful whether far- 
reaching taxonomic generalizations can be based upon the 'acid-fast' staining 
reaction, especially as this reaction has not played a very important role in 
mycological research. There seems to be no adequate reason why the genus 
should not be classed, in an unqualified manner, with the Hyphomycetes, as a 
Mucedineous group with tendencies toward an erect Isaroid habit. 
A more complete illustrated account will appear shortly in the Botanical 
Gazette. 
STUDIES OF MAGNITUDES IN STAR CLUSTERS, VIII. A SUM- 
MARY OF RESULTS BEARING ON THE STRUCTURE 
OF THE SIDEREAL UNIVERSE 
By Harlow Shapley 
Mount Wilson Solar Observatory, Carnegie Institution of Washington 
Communicated by W. S. Adams, May 21, 1918 
In the preceding communication of this series 1 methods were discussed for 
the determination of the relative distances of a considerable number of globu- 
lar clusters. The methods have now been developed so as to give not only 
relative values but also fairly reliable absolute distances for all globular clus- 
ters, and for all variable stars of the Cepheid class for which periods and ap- 
parent magnitudes are known. A rather detailed summary of the procedure, 
its accuracy, and the results of a thorough application of the methods, has 
been given in the February issue of the Publications of the Astronomical Society 
of the Pacific. The present"note will be confined to a synopsis of the more im- 
portant results pertaining to the probable extent and arrangement of the 
sidereal system. The detailed discussion is appearing in a series of papers in 
the Astro physical Journal, and will be separately published as Contributions 
from the Mount Wilson Solar Observatory, Nos. 151-157. 
Extending to the globular clusters the work of Miss Leavitt, Hertzsprung, 
and Russell on the Cepheid variables of the Small Magellanic Cloud and of 
the galactic system, we have been able to establish beyond question the inter- 
dependence for these variables of absolute luminosity and period of light vari- 
ation. By combining the apparent and absolute magnitudes, the distances and 
positions in space have been determined for about 140 Cepheid variables, 
most of which are much more distant than any objects for which parallaxes 
have been directly measured. Figure 1 shows their distribution. 
The distances of globular clusters are of a different order of magnitude from 
those heretofore entering stellar investigations. Although the average naked- 
eye star is near as compared with many Cepheid variables, the most remote 
Cepheid now known is not so far away as the nearest globular cluster. The 
