290 



PROCEEDINGS OF THE AMERICAN ACADEMY 



by which amount F^ was longer than the meter of the Archives. The 

 amount by which the length was actually diminished will appear in 

 the discussion which follows. 



The upper surface of this bar was made nearly a plane surface when 

 supported at its neutral points. The amount of metal removed at the 

 middle point was about .012 inch. Tempered steel plugs were then 

 inserted with a close fit in the upper surface of the bar, according to a 

 scheme shown in the following sketch : — 



Diagram Illustrating Subdivisions of the Yard and Meter. 



f g h 



METRIC. 



a b = 1 Meter. 



a c ^ c b := i Meter. 



ae = ef=fg = gh^hc=:l Decimeter. 



a d ^ d e = 5 Centimeters. 



d e subdivided to Centimeters. 



ENGLISH. 



a' b' = 1 Yard. 



a' c = c b' = ^ Yard. 



c d' = d' e' = e' b' = 6 Inches. 



e' b' subdivided to Inches. 



By means of a device suggested by Mr. George F. Ballou, who 

 prepared the bar, the surfaces of these plugs were brought into the 

 same focal plane after they had received the final polish. At this 

 point it will be proper to explain the method by which the general 

 surface of graduated bars is made plane. 



As will be hereafter shown, a plate to which a microscope is attached 

 slides freely upon the ways of the comparator. The flexure of these 

 ways is controlled by means of flexure screws beneath. It is evident 

 that, if the objective of a microscope of high power remains in good 

 focus at every point along a polished metal surface lying parallel with 

 a plane passing through the points of contact at the two ends of the 

 ways, the movement of the microscope plate along the ways defines a 

 reference plane with which any other plane can be compared. The 

 following tests have been employed in the definition of this reference 

 plane. 



(1.) Mr. George B. Clark, of the firm of Alvan Clark & Sons, pre- 

 pared for me a steel bar, one meter in length, which has one optically 

 plane surface, when the bar is supported at its neutral points. Placing 

 the bar thus supported upon the comparator, and adjusting it so that 

 surfaces at the two ends are in a focal plane parallel with that defined 

 by the microscope plate in corresponding positions, it has been found 

 that the objective remains in focus at every point along the surface 

 of this bar. 



