New Evg'nie J or Ruling. 73 



indicate a movement of rather less than one micron per division of 

 the micrometer scale and values of less than half that amount, or 

 1/50000 inch were both appreciable and reliable. In operation 

 the lower bearing point of the indicator — a polished steel sphere — 

 was brought inte contact with the face of the polished quartz plate 

 Just mentioned, the indicator needles being inclined at about 45°; 

 an angle both correct and convenient for the position of the micro- 

 scope reading scale, and other predetermined constants of the 

 apparatus. These various adjustments called for the expenditure 

 •of some little time and care at the outset, but any difficulties were 

 soon mastered and the instrument proved reliable and effective 

 throughout its subsequent use. 



As soon as the foregoing method afforded the necessary assurance 

 ■of reliability, a series of tests were made and the readings care- 

 fully tabulated, the results being recorded in graphic form. One 

 of the graphs, the earliest obtained, has been reproduced as Fig. 1, 

 Plate XIII., and will serve to illustrate the actual condition of the 

 «crew at this time, l^efore any attempt had been made to correct 

 the errors revealed by this method of testing. Seeing some explana- 

 tion is necessary to make clear the interpretation of the results 

 -obtained from the measurements and graphically recorded as Figs. 

 1,2, Plates XIII., and Fig. 1, Plates XIV., we may first indicate the 

 procedure followed in obtaining the data for constructing the 

 ■graphs here presented. 



The curve in the upper part of the figure was based upon a 

 record of some 16 readings (see below) taken at half-inch or 10-thread 

 intervals over the full length of the screw in the following manner. 

 The three-inch nut was first carefully fitted to the screw and run 

 into position, back towards the thrust end of the screw up to the 

 limit of the threads available, all the threads of the nut being 

 engaged. With the spherical end of the indicator in contact with the 

 quartz plate, the screw was slowly revolved through a complete 

 "turn, the observer meanwhile noting any change of index point, in 

 relation to the scale in the field of the microscope. It is here worth 

 drawing attention to the fact that the maximum and minimum 

 readings always corresponded with the same points on the divided 

 circle used for locating these positions, indicating that the bend 

 upon the screw was of the nature of a plane curve. 



A record of the mean deviation of the index per revolution for 

 three revolutions, was usually made for each position. This was 

 -done for every half-inch interval over the range of threads available 



