424 Mr. A. Miiller on an X-Ray Bulb 



sometimes (owing to the fact that the mercury vapour gives 

 electrons off very easily) had to be suppressed by a rectifier. 

 A photograph has been taken in order to get an estimate 

 of the potential from the end radiation. The wave-length of 

 the end radiation was found to be about - 22 A.U. This 

 corresponds to a potential of approximately 56,000 volts. 



5. Accuracy and- Corrections of the Wave-Length 

 Measurements. 



The wave-lengths of the mercury lines have been measured 

 in taking the copper Ka line as a standard (after exposing 

 the plate to Hg radiation, the spectrometer was shifted to a 

 second X-ray bulb which had a copper anticathode, there 

 the plate was exposed for a short time to the copper radi- 

 ation). Let « be the angle of reflexion corresponding to 

 the copper line, / the distance between the two symmetrical 

 positions of this standard line on the plate ; a. and / the two 

 corresponding quantities for the line for which the wave- 

 length lias to be found. Then we have the equation 



tan2a = /// . tan2a (1) 



A correction had to be made to this formula for the fact 

 that the surface of the crystal is not exactly in the axis of 

 rotation. a is supposed to be known, the relative accuracy 

 of a depends therefore upon the accuracy of the ratio /// . 

 Readings of the centre of the stronger lines could be made 

 on the Hilger Comparator with which the distances I have been 

 measured, to a few thousands of a millimetre*. The distance 

 between the copper standard lines was over 50 mm v between 

 those of the mercury y } line about 30 mm. If the lines are 

 well-defined, the error in the determination of l/l should not 

 be more than '01 to *02 per cent. This limit of accuracy has 

 been nearly obtained in a series of measurements of the ratio 

 ^Cno/?Cuj8 ]? where the following values resulted : — 



Plate IJl. 



33 143334 



34 1-13347 



35 1-13316 



36 1-13334 



37 1-13355 



In the case of /cua/^Hga the agreement is not quite so good : 



Plate IJL 



33 1-2988 



34 1-2978 



35 1-2976 



36 1-2985 



* Reproductions of some typical photos are given in Plate XIV. 



