ARTICLES 575 



mines the direction of the rays, without any error due to 

 uncertainty as to where the rays 'started from. It does not 

 matter if the position of the prism or grating is altered pro- 

 vided that its orientation is unchanged. In the X-ray spectro- 

 meter there is no focussing lens, and the determination of the 

 reflecting angle must be made differently. 



In many cases observers have measured, not the angle 

 through which the X-ray is displaced right and left, but the 

 angle at which the crystal must be set to displace it. This 

 method avoids all uncertainties due to the volume nature of 

 the crystal reflection, but it requires the use of the ionisation 

 method and is not adapted to the photographic method. 

 Perhaps it cannot reach the accuracy of Uhler and Cooksey's 

 method, which certainly gives highly concordant results. 



The following extract from their results will give an idea 

 of the accuracy attained. The readings of the angular deflec- 

 tions by calcite for four independent sets of measurements of 

 the lines a 2 and a x were : 



An error of 5* in the reading corresponds to an error in the 

 wave-length of about one part in 9,000. 



From such examples it is clear that as an accurate mea- 

 surement the determination of X-ray wave-lengths has made 

 strides. Collected results have been given in tabular form by 

 various writers, e.g. by Siegbahn, Jahrbuch der Radioaktivitdt 

 und Elektronik, XIII. 3, September 191 6, by Maimer and by 

 Friman in separate publications issued by the University of 

 Lund in Sweden, by De Broglie in the Comptes Rendus, by 

 Duane and others in the Physical Review. 



As a result of this work of the last two or three years the 

 accuracy of our knowledge of X-ray spectra has been in- 

 creased considerably, certain conclusions which had been drawn 

 tentatively have been put beyond doubt, and fresh ground 

 has been broken. 



First of all with regard to the original law enunciated 



