4;,0 Transactions of the Society. 



The general results of the examination of the half of each of 

 the scales show that they are remarkably free from accidental 

 errors in division, which probably do not reach 1 /* anywhere, 

 and that they are therefore eminently fitted to become standard 

 scales. There is, however, a very systematic and progressive error 

 in the divisions of both scales (i.e. in the halves examined) which 

 could not be overlooked. It consists in the graduations becoming 

 wider and wider apart from the beginning of the scale in each case ; 

 and this error seems quite unmistakable, being far in excess of 

 any conceivable error of observation, a millimetre in one instance 

 or ^jj in. in the other, in the centre of their respective scales, being 

 on an average about 1/* (or ^ p.c.) longer than a millimetre or ^ in. 

 at the beginning. 



The absolute value of at least one of the scales must also be 

 considerably in error, for from the mean of all the measures the 

 ratio of Grayson's millimetre to Grayson's inch was found to be : — 



25-369 : 1 

 instead of being 



25-400 : 1 



which is the correct ratio, Grayson's ratio being wrong to the extent 

 of one part in 800. 



It would appear, however, highly desirable that at least one of 

 these scales be sent to the National Physical Laboratory, when 

 they are ready to carry out such determinations, to be accurately 

 standardized — anyhow over its first half of graduations, which seems 

 sufficient for all useful purposes — after which it would be of great 

 value and an excellent addition to the resources of the Society. 



