gray: micrometer microscopes 51 



All that now remains is to determine the magnification and to 

 translate these curves into the figures that forms the final table. 



It is, of course, desu'able that the numbers in the correction 

 table, and also the tabular differences, be small. Tabular dif- 

 ferences arising from periodic irregularities cannot be reduced 

 except by improving the screw. Progressive errors and deviations 

 from the nominal screw-value, however, even if small, lead to the 

 accumulation of large corrections when many turns must be used. 

 But if the magnification is adjustable, these corrections can often 

 be materially reduced. The first of the series of curves plotted, 

 namely, that representing the progressive errors, was based on 

 deviations of the measured lengths of individual turns from the 

 length represented by the basis selected for comparison. If a 

 known length, such as a standard millimeter, be measured with 

 the microscope to which the screw S belongs, and the readings cor- 

 rected by using the curves described above, the distance measured 

 that would correspond to the basis of comparison is likely to differ 

 somewhat from the distance to be represented by 100 divisions 

 (one turn) in the correction table. A little reflection will make 

 it apparent that corrections for this divergence can be added to the 

 progressi\'e corrections by the simple device of drawing a line at 

 the proper angle thru the point selected for zero correction. The 

 distance parallel to the axis of ordinates between this line and the 

 progressive curve (Co) measures the total correction for any comb 

 reading. It is, therefore, evident that these corrections can be 

 reduced to a minimum for any given part of the screw by adjust- 

 ing the magnification of the microscope so as to give this line the 

 most favorable slope. A moment's inspection of the progressive 

 curve will decide what this should be. After the magnification 

 has been adjusted with sufficient closeness, a standard length 

 interval should be carefully measured, and the correcting line 

 accurately located. Several measurements using different parts 

 of the screw will afford valuable checks on the accuracy of the 

 correction curves. Table 2 represents measurements of this kind. 

 We are now ready for the final step of reading off the corrections 

 from the curves and entering them in the table. A translating 

 scale is used which reads the pbts in terms of the unit chosen for 



