ASTRONOMY: H. SHAPLEY 
345 
plane beyond the confines of the local cluster: First, the distance of the 
center of the system and the total extent of the Galaxy may be con- 
siderably greater than inferred from the visible globular clusters; sec- 
ond, the transition from globular to open clusters, if there be such an 
evolution,^ is not necessarily rapid and inevitable when the globular 
clusters enter thickly populated galactic regions.^ It should be noted 
that the diameter-parallax correlations^ are some insurance that a 
hypothetical partial obstruction has exaggerated neither the distances 
of globular clusters nor the galactic dimensions. If dark clouds obscure 
distant globular clusters they are remarkably thorough (except in one 
or two possible cases^), the obscuration at any point appearing to be 
absolute or non-existent. According to the magnitude results now 
available for faint galactic stars, the obstructing clouds also are of such 
character that they do not affect star colors appreciably. 
Several years ago the correlation of the apparent diameter to the 
brightness of stars in open clusters was pointed out and the possibility 
noted of obtaining relative parallaxes from measures of magnitude or 
diameter.^ My values of the relative parallaxes of these open groups, 
however, as well as the provisional absolute values, have remained un- 
pubUshed in the hope that the accumulation of data on magnitudes and 
spectra would permit the determination of more definite absolute 
parallaxes. The open clusters contain few, if any, Cepheid variables, 
and appear almost without exception to be quite beyond the reach of 
direct measures of distance. Parallaxes of an accuracy comparable 
with that for globular clusters seem unattainable; but, notwithstanding 
some uncertainty in diameters due to looseness of structure and to inter- 
mixture with the Milky Way stars among which the open clusters usu- 
ally lie, it appears quite possible to use their angular dimensions to 
determine a system of relative distances and to use the apparent mag- 
nitudes of the red giants or the B-type stars in as many clusters as 
possible to establish the scale and zero point for absolute distances. 
My observational work on open clusters comprises: (1) Thirty spec- 
trograms of some 200 faint stars in various northern groups, made with 
a slitless spectrograph of small dispersion at the 80-foot focus of the 
60-inch reflector; (2) more than a hundred direct photographs at the 
primary focus of the 60-inch; (3) the determination of magnitudes and 
colors of about two thousand stars; (4) the measurement of the diame- 
ters and form of all known open clusters on Franklin- Adams charts, 
Harvard photographs, or Mount Wilson plates. This work has been 
supplemented by similar data from other sources, chiefly from the 
