796 PROCEEDINGS OF THE SOCIETY. 



line, these readings had apparently in no single case been given, with the 

 result that the calculation of the mathematical probable error of the 

 result was impossible. Results claiming- such a high order of accuracy 

 should have been accompanied by the numerical result of control deter- 

 minations. Nothing, for instance, would have been easier than to 

 determine independently, and place side by side, the different comparative 

 measurements for five given spaces, say, obtained by using different 

 micrometers and different powers. 



.Mr. Ilopkinson said that it was well known to meteorologists that 

 thermometer tubes should not be graduated until thev had been made 

 for several years, otherwise they would alter, mercurial thermometers 

 requiring in course of time a minus correction owing to the contraction 

 of the glass, and he thought that if the errors in some of the 

 micrometers tested were generally in the same direction they might be 

 due to this cause. 



The Chairman thought the great advantage of this paper was not 

 so much the value of the results obtained, as that the writer had 

 brought the subject forward for consideration. There were many 

 matters which made it doubtful if they could accept the results as 

 being entirely correct ; also it seemed hardly fair in considering the 

 subject to take as micrometers for comparison standard rulings which 

 had no doubt been carefully selected from a large number, and to 

 compare these with unselected specimens of commercial articles. All 

 who were accustomed to this work knew how very difficult it was to 

 make these comparisons. The whole difficulty was much greater than 

 appeared at first sight, and it began with that of fixing a unit of 

 measurement, the standard hitherto adopted being based upon the 

 accepted standard English yard or French metre. He had, however, 

 been authorised to bring before their notice a machine which was 

 being made for Major Macmahon and Dr. Tutton on behalf of the 

 Standards Department of the Board of Trade in order to calibrate in 

 terms of the wave-length of light the standard yard deposited at "West- 

 minster. For this machine a large concrete bed had been sunk in the 

 ground, and a brick foundation raised upon it to carry a heavy metal 

 bed on which carriages would run containing a pair of very high power 

 Microscopes. For the purpose of making an index they had obtained 

 from Mr. Grayson, of Melbourne, a series of five lines ruled T 4oo or 

 an inch apart on speculum metal, and some plates were supplied which 

 had a number of such sets of five lines ruled at intervals upon them. A 

 double cobweb micrometer in the eyepiece of the Microscope could be 

 set so that one web was placed on either side of the central of the five 

 lines. This formed the settling device ; the Microscopes could then be 

 travelled from one to another of a series of such rulings, and the method 

 adopted to ascertain the amount of such travel did not depend on any 

 screw or other mechanical method of measuring distances which were 

 always open to sources of error. It depended on a direct measurement 

 by means of an interferometer of the distance expressed in wave-lengths 

 of light. One interferometer mirror was fixed to the base of the 

 machine and the second interferometer mirror on the travelling Micro- 

 scope itself. By an ingenious step-by-step method'with two sets of plates 

 and two travelling Microscopes, long lengths could be measured without 



