384 
INTERNATIONAL ASTRONOMICAL UNION Proc. N. A. S. 
reflectors, for different objectives seldom behave exactly the same and 
not infrequently very large differences are to be met with. This is best 
illustrated by the following comparisons of different series of photographic 
results. Here the complicating influence of a filter does not enter, because 
no filter has been used, and the effect of the kind of plate is largely elim- 
inated, because with some exceptions the same brand of plate has been 
used throughout. The comparisons show only the term depending on the 
color-index C. 
MW— Greenwich =+0.270 1^^^ ^ 
MW-Dziewulski ^ = -0.06C / ^^^^^^ 
MW— 16-in. Metcalf = +0.33C ] 
MW—8-in. Draper = +0.06C ^ . r^o 
o ■ A • r» -7- ^ Mt. Wilson Cont. No. 98, p. 5 
8-m. Draper — 0.5in. Ross-Zeiss = — 0.13C 
8-in. Draper— 4-in. Cooke = 4-0.21C J 
Other examples might have been added, but those given illustrate the 
magnitude of the effect and show the impracticability of restricting the 
photographic measurement of stellar brightness to any closely specified 
region of the spectrum, for the observer must use the telescope which is 
available. 
Emphasis should, therefore, be directed, not toward standardization, 
but toward the publication of results in such a form that they may be re- 
duced to any system for purposes of comparison and combination with 
other results. Apparently this is to be accomplished only by giving the 
color of a star in addition to its brightness. Spectral type cannot be re- 
garded as ^a satisfactory substitute for color, for it is now known that the 
color-indices of stars of the same type-color, especially the G^'s and K's — 
may differ by half a magnitude or more. - 
Admitting the impracticability of basing the observations upon any 
prescribed system of color (in the sense that the measurements are to be 
made within certain specified limits of wave-length), one might still ask 
if it is not desirable that the results be reduced to a standard system in 
advance of publication, in order that all catalogue values may be directly 
comparable. The answer here is that in the present state of photometry 
the reduction factors for color will necessarily be uncertain, and hence 
subject to correction. It thus appears important that the results referred 
to the color system of the equipment used should always be included in 
the published results, even when reduced values are also given. 
These details are presented as an indication that reservation and caution 
must characterize the decisions of a central committee in case its activities 
are to be helpful ; at the same time they reveal the disconcerting fact that 
measures of stellar brightness will attain their full value only when results 
for the colors of the individual stars become available. This is strikingly 
shown by an attempt to compare the results of the Durchmustenmg of 
the Selected Areas recently issued as Harvard Annals 101 with photographic 
