PROCEEDINGS OF THE SOCIETY. 



79i 



that it was deteriorating after a few years' time. He indorsed Mr. 

 Conrady's suggestion that the Society might with advantage consider the 

 advisability of standardising Micrometers. 



Mr. F. J. Cheshire said that he had carefully read Dr. Ewell's inter- 

 esting paper, but scarcely thought that the results had been given in the 

 form most readily appreciable by an audience of microscopists who had 

 not already read the paper. He had therefore taken the liberty of ex- 

 tracting those results from the author's paper most likely to be of interest 

 to English microscopists, and had calculated from them certain com- 

 parative percentage figures, shown in tabular form on the blackboard. 



The speaker then went on to point out that the important question 

 to decide as a preliminary to the acceptance of Dr. Ewell's results was, 

 of course, the order of reliability of those results. The method adopted 

 and the figures obtained were not, unfortunately, set out in sufficient 

 detail to allow of a conclusive answer ; but as regards one important 

 matter, it appeared from the fact that (1) only low and medium powers 

 had been employed, and (2) that " the mean of from five to ten readings 

 of each end of each space," that no attempt had been made to utilise the 

 same part of the micrometer screw, as far as possible, for the different 

 measurements. The author would thus appear to have taken it for 

 granted that the various screws employed in the eye-piece micrometers 

 had been cut and mounted so as to realise an order of accuracy greater 

 than that claimed for the results. Let us see what this means. In the 

 first micrometer tested (No. 1 Bausch and Lomb) a tenth of a millimetre 

 is divided into about 4000 parts, so that if accuracy is claimed for the 

 last significant figure a screw with a maximum error of 40W P al 't °f the 

 space measured in the length of screw used, is postulated. Assuming 

 even that the correctness only of the third significant figure is claimed, 

 this works out as equivalent to the assumption that the screw employed 

 had for the length used an error not exceeding the Tin nnnr P a **t of an 

 inch, the accuracy obtained by Rutherford in a screw which took three 

 years to make, and the most exquisite workmanship ! It is to be re- 

 gretted that information on such a vital matter has not been given by 

 the author. Finally, the speaker pointed out that, although the author 

 states that from five to ten readings were taken of the position of each 



3 c 2 



