OF ARTS AND SCIENCES. 291 



(2.) A similar test has been applied with a surfece plate 42x1^x6 

 inches, prepared by Professor Morris of Cornell University. I have 

 two of these surface plates. One was prepared in the usual way, but 

 the other was not touched with a tool after it left the planer. I think 

 it would be difficult to find a better specimen of planing. Although 

 each plate weighs about tweuty-five pounds, one will lift the other by 

 the contact of their surfaces. 



(3.) If a shallow dish of mercury is placed upon the comparator, 

 the surface of the mercury forms a reference plane with which the 

 plane described by the microscope plate can be compared. Under a 

 half-inch objective supplied with a Tolles opaque illuminator, one can 

 focus upon the minute globules of mercury with considerable sharpness. 



Having proved that the microscope plate describes a true plane dur- 

 ing its movement along the ways, the bar whose surface is to be pre- 

 pared is placed upon the comparator, supported at its neutral points, 

 and the surface is worked down with soft stone till every part remains 

 sharply in focus under the objective. 



Experience has shown that the surface of every bar prepared in this 

 way falls into a plane surface within a short time after it is placed upon 

 its supports. 



(4.) Recently a fourth method of testing the deviation of the plane 

 described by the microscope carriage from a true horizontal plane has 

 been applied with good success. It may be described as " the method 

 by two microscopes." A flat bar, about sixty centimeters in length, 

 was attached firmly to the carriage in a direction parallel with the 

 ways. At each end, the arm which supports the microscope was fas- 

 tened with an adjustable motion iu every direction. The two micro- 

 scopes when placed in position are one half-meter apart, and equidistant 

 from the centre of the carriage. 



The Clark reference bar having been placed in position upon the 

 bed of the comparator, the defining lines of each end were adjusted for 

 focus under the right-hand microscope. The flexure screws beneath 

 the comparator were then moved till the defining line at the middle 

 point was in sharp focus. It is now found that every part of the sur- 

 face of the reference plate remains in good focus. But it does not 

 follow that either the plane described by the microscope carriage or 

 that of the reference bar are real planes. It can only be said that they 

 are parallel. We now adjust the left-hand microscope upon the refer- 

 ence bar at the left end. Now, if the microscope carriage describes a 

 true plane in passing from the middle point to the right end of the 

 comparator, the defining line near the middle point of the bar at which 



