ASTRONOMY: H. N. RUSSELL 
397 
and Scheiner,^^ and the photographic investigations of Rosenberg.^^ 
Fainter stars, down to the sixteenth magnitude, at least, can be 
reached by the determination of color indices. 
(b) In order that these color indices may be capable of full utiliza- 
tion, it is necessary, first, that trustworthy and homogeneous scales of 
visual, photographic and photovisual magnitudes be established over 
the whole range of about 47 magnitudes from the Sun to the faintest 
observable stars. This problem, which is fundamental in all stellar 
photometry, is already well advanced toward solution. But in the 
second p'ace, it is necessary that the physical meaning of the units of 
magnitude should be precisely known; that is, that the 'luminosity 
curve' which expresses the relative sensitiveness of the photometric 
receiver for equal energy of different wave-lengths should be exactly 
determined. And, above all, it is essential that this luminosity curve 
should be independent of the brightness of the stars under observation. 
These last two conditions are at present very imperfectly satisfied, if 
at all. Very little is known about the luminosity curves of the standard 
plates and apparatus used in the determination of photographic and 
photovisual magnitudes, and nothing at all about the luminosity curves 
of the eyes of the 'standard observers' at different observatories, — 
except that they must be very different under the conditions prevailing 
at Harvard and at Potsdam.^^ It is certain that the Purkinje effect 
alters the form of the visual luminosity curve as the brightness of the 
illumination varies, probable that this affects the visual comparison of 
the brightness of stars of widely different magnitudes, and uncertain 
whether, and to how great an extent, similar photographic influences 
exist.i9 
The direct determination of the luminosity curves for the principal 
instruments and methods employed in the determination of photo- 
graphic and photovisual magnitudes would be neither difficult nor la- 
borious. For visual observations it can be derived indirectly, if direct 
measures prove difficult. To make these investigations at once is 
urgently desirable, for the present bases of the scales of stellar magni- 
tude are not permanent. The photographic and photovisual scales 
depend on the properties of present commercial types of rapid plates, 
which may not be manufactured a few years hence if improvements are 
devised ; and the visual scales ai e based on the characteristics of the eyes 
of observers some of whom have already retired from active work. 
Such an investigation would also establish a connection between the 
scales of stellar magnitude and the physical units of measurement of 
light in the laboratory (which are now defined in terms of a definite 
