418 



SCIENCE. 



[X. S. Vol. XXI. No. .333. 



differences arise in the visual estimates of 

 colored stars by different observers or by 

 the use of telescopes of different apertures. 

 The 'color correction' amounting to one or 

 two magnitudes for a red star in the photo- 

 graph, is no greater than the difference be- 

 tween the simultaneous Harvard and 

 Rousdon visual estimates of the brightness 

 of such stars. In fact, we find discrep- 

 ancies of the same kind and similar in 

 amount between different observers, dif- 

 ferent telescopes, visual and photographic 

 results, and different brands of plates in 

 photography. If the statement is made 

 that no known relation exists between 

 visual and photographic magnitudes, the 

 retort can be made that a normal visual 

 scale does not exist. 



The advantages arising from the use of 

 orthochromatie plates have long been rec- 

 ognized, hxit Scheiner dismisses them with 

 the statement that they can never yield 

 visual magnitudes. The suggestion was first 

 (as far as I am aware) made by Schwarzs- 

 child that the difference between the magni- 

 tudes of a colored star on ordinary and 

 orthochromatie plates can be taken as a 

 measure of the star's color. If this dif- 

 ference is a function of the color it only 

 remains to find the form of the function, 

 and then complete allowance can be made 

 for the effect of color and that troublesome 

 factor can be eliminated, making pof-sible 

 the reduction of photographic magnitudes 

 to visual, or vice versa. Two methods are 

 available for finding the form of the func- 

 tion. First, by trial on known stars of 

 dift'erent color (spectral type). To fix our 

 ideas, suppose, for example, that a star of 

 color 5 on Chandler's decimal scale was 

 7.0 magnitude visually, but photographed 

 8.0 magnitude on an orthochromatie plate 

 and 9.0 magnitude on an ordinary plate. 

 For such a star the orthochromatie plate 

 gives half the color correction. It is evident 

 that by such experiments with standard 



stars of known magnitude and color, the 

 form of the function can be found. This 

 work is being done by the writer, under a 

 grant from the Carnegie Institution, using 

 Cramer isochromatic plates in connection 

 with ordinary plates, on the 24-inch re- 

 flecting telescope of the Yerkes Observa- 

 tory. Provisional results thus far obtained 

 are very promising. An independent 

 method which will also be used for finding 

 the form of the function, consists in com- 

 paring the intensity curves of the spectra 

 of stars of different types with the in- 

 tensity curves of the solar spectrum on the 

 two kinds of plates used. It is evident 

 that the photographic effect is the integral 

 of the product of these two curves. 



If the objection is urged that the diffi- 

 culty of coordinating the results obtained 

 with different brands of ordinary and or- 

 thochromatie plates, will be equal to the 

 difficulty of harmonizing the visual and 

 photographic systems, it may be met by 

 the suggestion that any brands of plates 

 used should be calibrated by observations 

 of a carefully selected list of standard 

 stars, including each spectral type. 



No less important than the choice of 

 plates is the kind of telescope to be used. 

 It seems to the writer that the reflector is 

 the only telescope suited for this work, since 

 by it the rays of all wave-lengths are 

 brought to the same focus. 



Emphasis is needed on two further 

 points in regard to the adaptability of the 

 reflector to this work. When extreme 

 ratios of aperture to focal length are 

 avoided, first, the field is very nearly flat; 

 second, the action is very rapid, so that 

 the work can be extended to faint stars. 

 This flatness of field has been denied, both 

 from theoretical reasons and from so-called 

 measures of reflector plates; but it should 

 be stated that the theory is incomplete, not 

 taking proper account of the distribution 

 of light in the 'blurred' image; also that 



