128 Mr. F. L. 0. Wadsworth on a 



a quantity quite negligible in comparison with the errors of 

 setting. This indicates another method of using this new 

 form of instrument to good advantage ; i. e., the method of 

 superposition of object and comparison scale. To accomplish 

 this, the mirror A is half silvered and the scale is viewed 

 directly through the unsilvered half in the direction o' (see 

 fig. 4), the object being at the same time seen by reflexion 

 from the silvered half in the direction of 1 *. In this case we 

 may make the measurement either by determining the distance 

 between the image of the point and the image of the nearest 

 millim. division on the scale with an ordinary form of 

 micrometer ; or better, by bringing these two images into 

 coincidence by means of a Rochon double micrometer, an 

 opthalmometer or a parallel-plate micrometer. The second 

 method of coincidence has the decided advantages both of 

 greater rapidity, only one setting and reading being necessary 

 instead of two, and of greater accuracy for the same reason, 

 since any error, due to a change in position of any part of 

 the apparatus in the interval between two settings, is thus 

 avoided. This last advantage fully balances the disadvantage 

 of the greater effect of a given error of levelling on account 

 of the greater angular distance between scale and object. 

 One additional cause of error is introduced, i. <?., that due to 

 a want of parallelism between the mirror A and the axis of 

 rotation C ; but since this should not exceed a fraction of 

 a second, if the first adjustment has been properly made, its 

 effect is negligible. 



Of the various instrumental means for obtaining coincidence 

 perhaps the simplest and most convenient, as well as one of 

 the most accurate, is the parallel- plate micrometer first 

 invented by Clausen, and quite recently reinvented and much 

 improved by Poyntingf, who was the first to adopt it for 

 cathetometric measurements. Figs. 4 and 5 show in plan 

 and elevation a form of this micrometer modified slightly from 

 that described by Poynting, to better adapt it to this particular 

 instrument. It consists simply of a plate of plane-parallel 

 glass, P, rotating on an axis C / at right angles to the axis C, 



* The mirror should be half silvered horizontally, i.e. the line of 

 separation of the silvered from the unsilvered portion should be parallel 

 to the axis of rotation, both because the maximum resolution is required 

 in a vertical direction and because, as will be seen later, this management 

 is the better adapted to the use of ceitain forms of micrometer. 



t Phil. Trans, vol. clxxxii. 1891 A, p. 588. See also " On a Parallel 

 Plate Double Image Micrometer," Monthly Not. of the Royal Astron. 

 Soc, vol. liii. No. 8, 1892, p. 556. 



