7 353 



measure for the dispersion of the apparent sensibilitj' of the plates. The relative 

 small efficiency of the later plates is striking. 



The exposures have all been made by opening and closing the plateholder. 

 For this reason the shortest exposure time, 2^*, was rather uncertain and the different 

 parts of the plate were not exposed exactly alike. But this circumstance is of no 

 importance in the present case, because the intensity of each star image was estim- 

 ated independently when the effective wavelength was measured. 



3. The Measuring of the Plates. 



The plates have been measured by the writer at Potsdam in the time from 

 1915 March 17 to Aug. 25. For this purpose the same Toepfer machine was used 

 which has served for the measuring of my photographs of double stars (Potsdam 

 Publ. Nr. 75). By this machine the displacement of the whole plate is measured 

 micrometrically, which gives a guarantee for the constancj' of the screw value. 



The rectangular coordinates of the stars on each plate were noted to ^/lo of 

 a mm. The diameter of the central star image was estimated to Vioo of a mm by 

 the aid of a double wire in the focal plane of the objective of the microscope. 



On each plate the effective wavelengths were measured within a field, the 

 diameter of which was 1°. This is a somewhat greater field than indicated above 

 as fit for use and the star images near the border are rather unsharp. But in the 

 present case, where several overlapping plates are available, the possible systematic 

 differences between stars near to and far from the centre of the plate will to a 

 great extent be compensated by taking the mean of the different plates, as usually 

 the same star will on some plates be near to and on others far from the centre. 

 Only the stars more than about 1° distant from Alcyone make an exception to this 

 rule. — In other cases, where only plates with a common centre are available, it 

 is not to be recommended to use a field greater than 40' in diameter, when the 

 reflector is diaphragmed down to an aperture of 1 meter. 



The general appearance of the images is shown in Figure 2 for some stars 

 of different brightness (Alcyone and neighbouring stars). The distance between the 

 two spectra on each side of the central image was measured as between two star 

 images. When the spectra were unsymmetrical it was aimed to place the wire in 

 such a way that the area of the spectral image was cut into two equal parts. The 

 setting was read to ^/t corresponding to about 2A in the spectra of the first order 

 or about '"'01 in the colour index. 



When the thickness d of the wires of the grating is equal to the spare width 

 I between them, all spectra of even order disappear, while, as stated above, those 

 of uneven order are at their maximum intensitj'. For the grating here used d is 

 only approximatelj^ equal to /. Hence the spectra of the second order also make 

 their appearance in the images of bright stars. In fact we have d = 2"745 mm and 

 / = 3'255 mm. Corresponding to the proportion Vd = 1118 the central image is 



