The Microscope. 259 



anade to correspond in })itcli with equal siibdivisions of a standard 

 yard traced upon a metal bar. The screw for the engine constructed 

 for Cornell University was made in this manner. Professor Anthony 

 has shown that the maximum accumulated error of the screw does 

 not reach 2 mikrons for a limit of 20 inches, while the actual error 

 of any selected point will not reach 1 mikron. This screw was cut 

 in the manner indicated, in the third story of a building occupied by 

 machinery, which produces a decided tremor in every room. It was 

 only found necessary to make the attachment of the microscope to 

 the compound rest of the lathe very firm and to brace the head of 

 the lathe very securelv from the floor. 



The writer was recently called upon to "level up" the head of 

 a veiy heayj' planer, having ways 18 feet in length. Several days 

 had already been spent in securing as good an adjustment as could 

 be obtained with the aid of a spirit level of special construction. A 

 plank 22 feet in length, S inches in width and 2 inches in thickness 

 ■was set up edgewise beside the plating of the planer, but insulated 

 irom it. A groove half an inch wide and half an inch deep was 

 ploughed in the upper face of the plank, and after having stopped 

 both ends, the groove was filled with mercury. The surface of the 

 mercury then formed an invariable plane of reference. The micro- 

 scope was securely attached to the platen and adjusted for sharp 

 focus upon the surface of mercury at one end. The platen was 

 then moved along until the microscope occupied a position near the 

 other end of the groove. This end was then adjusted by elevation 

 or depression, as required, until the surface of the mercury was 

 sharply in focus. After two trials, it was found that the surface of 

 the mercury was at the same constant focal distance from the microscope 

 as indicated by sharpness of definition. Notwithstanding the fact 

 that extreme care had been taken in the original adjustment by the 

 aid of the spirit level, it was found that as the platen moved toward 

 the central part of the bed, the focus became more and more indis- 

 tinct, indicating that the central part was too low. The proper 

 elevation was then made at these points by means of heavy set-screws, 

 when it was found that the mercury was sharply in focus under the 

 objective throughout the entire range of motion. As a check upon 

 the accuracy of the adjustment, a surface plate, S-i^ feet in length, 

 was now planed, when it was found that the deviation from a true 

 surface did not, at any point, exceed the third part of the thickness 

 of tissue paper. Two facts of considerable importance are to be 

 noticed in connection with this experiment. First, that the time 



