46 



C. V. L. Charlier 



where such designations of the spectra as A+, G — K±, Ko — M etc. are by no 

 means rare. For investigating whether systematic errors in the determination of 

 the color-indices have crept in through this cause I have found it desirable to 

 make a new determination of the color-indices using the Harvard values of the 

 spectral types. 



Of the 631 stars in Parkhurst's table there are 296 stars for which the 

 Harvard determination of the spectral type is known either from the » Revised 

 Harvard Photometry » in H. 50 or from Circular 180 from the Harvard Observatory, 

 where an extract from Miss Cannon's »Revised Draper Catalogue» is given, em- 

 bracing all stars in the Durchmusterung brighter than 8.3 situated within ten de- 

 grees of the north pole. 



The results of Parkhurst have to be challenged even in another respect. I 

 have already stated the systematic difference between the results of Pakkhurst and 

 those obtained at Harvard, and suggested that this difference may be explained 

 through a colour equation in the Harvard photometric measurements. That this 

 colour equation, found by Messrs Müller and Kempf, is applied to Harvard 

 and not to Potsdam is supported by the fact that the comparison between Park- 

 hurst's photovisual magnitudes and the visual magnitude of P. G. (— Potsdam 

 General Katalog) show no noteworthy systematic differences, a result even confirmed 

 below. It is therefore of interest to determine the color-indices combining the 

 photographic magnitudes obtained at Harvard with the visual magnitudes of P. G, 



24. Consequently I have performed these two investigations. The photographic 

 Harvard magnitudes are to be found in H. 59, where Mr Edward S. King first 

 determines the photographic magnitudes of 33 bright stars (H. 59, N:o 4), than in 

 N:o 5 extends this investigation to 77 stars mainly brighter than 3.5 and finally in 

 N:o 6 adds further 76 stars. The photographic magnitudes of these 153 stars were 

 obtained from Table XII (H. 59, p. 177). For 109 of them the visual magnitude 

 could be obtained from P. G. The spectral type is taken from H. 50. 



First a provisory computation was made for determining the zero-point. It 

 was found that the différences between, the Harvard photographic magnitudes {= '^-h) 

 and the P. G. visual magnitudes (= inj>) are to be augmented by -|- 0.223 for giving 

 the visual magnitudes in the »international v scale*. In this manner the following 

 values of the color-index is obtained for the different spectral types. 



* This constant was obtained not from tlie stars of the spectral-type AO alone, but from all 

 stars considered, comparing the results according to the method of least squares to a straight 

 line representation. 



