292 PROCEEDINGS OF THE AMERICAN ACADEMY 



the other microscope had been adjusted for focus will remain in focus 

 under tlie left-hand microscope. 



This operation will test the movement of the carriage over only one 

 half of the distance moved, viz. from the middle to the right end. But 

 by adjusting the left microscope upon the right end of the bar, and 

 moving the carriage from right to left, the remaining half can be tested 

 in a similar manner. It is found that only two or three trials need to 

 be made in the adjustment of the bed-plate by means of the flexure 

 screws in order that the defining lines at each end of the reference bar 

 shall remain in sharp focus under both microscopes. Of course, the 

 lines at the middle of the bar will now remain in focus only when they 

 are in the same plane as the terminal lines. In the case of the Clark 

 reference bar this condition is fulfilled with great exactness. It will 

 be noticed that this method does not require that the surface of the 

 reference bar shall be a true plane. Any two points may be chosen, 

 but these points must remain in the same plane during the entire op- 

 eration of testing. 



Since this method has been adopted, much better results have been 

 obtained than it was possible to obtain before. 



(5.) A fifth method avoids dependence upon the variation of focus 

 as the test of a plane surface. In this method the microscope is at- 

 tached to the carriage in a horizontal position, the line of sight being 

 a little above the surface of a bar which is supposed to be a plane sur- 

 face, and with which the plane described by the microscope carriage is 

 to be compared. 



If a small piece of metal having one plane surface is placed at differ- 

 ent positions upon the surface of the reference bar, a given point upon 

 the front face, formed by the intersection of a perpendicular and a 

 horizontal defining line, will remain in a plane parallel with the sur- 

 face of the bar. This test-piece is first placed at one end, and the 

 defining horizontal line is brought into coincidence with the defining 

 line of the microscope micrometer. The carriage is then moved to the 

 other end, and the test-piece is moved upon the surface of the bar till 

 the cross lines are brought into the field of the microscope. The co- 

 incidence with tiie micrometer line is then made by the vertical adjust- 

 ment of the bar. The test-piece is now placed at the centre of the 

 reference bar, and, if the coincidence of the defining lines is still main- 

 tained, the surface described by the carriage is parallel with the surface 

 of the bar. The deflection of the ways at any point can thus be meas- 

 ured by means of the micrometer screw of the microscope. 



This method of preparing the surfaces of bars for graduation obvi- 



