Precision-measurements in the X-Ray Spectra. 651 



As to the estimated accuracy of this value, it has been 

 found after continued working with the vacuum-spectrograph, 

 that the method formerly used for measuring the distance 

 rotation axis to plate was not as reliable as desired. This 

 may answer for some units in the last decimal. A second 

 reason for a little inaccuracy in the value lies in the fact 

 that the lattice constant of rock-salt has not been corrected 

 for the temperature. It may be mentioned that the above 

 value is a mean from measurements both with rock-salt and 

 calcite. 



For determining the Cu wave-length with the new spectro- 

 graph two specimens of calcite have been used. One of these 

 was the same that was employed for the vacuum-spectro- 

 graph. No real difference between the values of these two 

 crystals can be found, as seen from the following table. 



Reflexion -angles of Cu-Ka^ by Calcite. 



Plate. Angle. Crystal. 



14 ... 14°42'7"'21 



15 3"-7 I ni] 



16a 3"'3i Old specimen. 



166 13"-3i 



17 6"-8j 



18a 0"*5 \ New specimen. 



186 3"-8J 



Mean ... 14° 42' 5"'5 



\=1537-44X.U. 



The two values for the wave-length found with vacuum- 

 spectrograph and the new spectrograph thus differ by about 

 0*005 per cent. This shows firstly that the penetration effect 

 with spectrographs of the rotating crystal type has no 

 influence on wave-lengths for which it has hitherto been used 

 for exact measurements. 



It may be mentioned that at these wave-lengths the 

 vacuum-spectrograph is able to give the greater accuracy 

 as the angle to be measured is then the fWr-fold reflexion- 

 angle. On the other hand, very sharp lines may be obtained 

 with the last described spectrograph up to wave-lengths as 

 short as 02 A.U. (a gold-slit being used). 



Lund, Physical Laboratory, 

 July 1919. 



