604 Intelligence and Miscellaneous Articles. 



outbreak of war. Up to that time I had not obtained any evidence 

 of refraction. It may be of interest to indicate briefly the method 

 of procedure adopted in these experiments, and also the con- 

 clusions which were drawn regarding the order of magnitude of 

 the possible refractive index. 



The general arrangement of apparatus was very similar to that 

 adopted by Professor Barkla. By an arrangement of lead slits an 

 exceeding narrow beam of X-radiation gave rise to a fine hori- 

 zontal line about two or three centimetres long on a photographic 

 plate placed at about one metre beyond the levelling-table carrying 

 the prisms. In the path of the beam was placed a train of eight 

 prisms each- of 90° refracting angle with their refracting edges 

 horizontal and perpendicular to the beam. By the side of these 

 prisms was placed a similar set with their refracting angles 

 reversed. Any evidence of refraction would be recorded by a 

 lack of alignment of the two halves of the line image obtained 

 on the plate after several hours' exposure. No such evidence 

 was obtained. A simple calculation showed that the refraction of 

 X-radiation, if any, in these materials would correspond to a 

 refractive index lying between 1*000005 and -999995. 



H. B. Keeke, D.Sc, 

 University of Birmingham, Assistant Lecturer in Physics. 



Nov. 11, 1916. 



additions to peof h. s. carslaw j s paper on 

 napier's logarithms in the November number. 



The following footnotes should be added : — 



Page 479. To the sentence forming the third and second lines 

 from the bottom of the page: Whenr = 7, nl r 1 = 161180896-38 

 (cf . Construction Section 53) : add the following footnote : — 



The error in Napier's Second Table affects the accuracy of his Canon, 

 and this number should be 161180956-51. The correction can be made 

 from the corrected result given by Macdonald iu his English translation 

 of the Constructio, pp. 94-5, for it is not difficult to show that 

 nll=7nl 7 10 6 . 



Page 480. To the sixth line : and nl 7 l-nl 7 10 = 23025842-34 

 (cf . Constructio, Section 53) : add the following footnote : — 



This number should be 23025850*93, since it is easy to show that 

 nl 7 l— nl T 10=nl 7 10 6 , and Macdonald gives the corrected logarithm of 

 10 6 (loc. cit. pp. 94-5). 



