322 PROCEEDINGS OF THE AMERICAN ACADEMY 



(1.) The stops U V are set at the same height above the table. 



(2.) The Hue bar is then adjusted so that the horizontal Hue is par- 

 allel with the cylindrical ways, and is at the same time in the focal 

 plane of the objective. The means of adjustment in three planes are 

 not shown in the figure, but they are the same as those already de- 

 scribed. The frame which carries the lever movements is fitted to the 

 bed-plate by the dowel pins 0. 



(3.) The jilate K is moved upon the cylindrical ways till a contact 

 between U and V takes place. While the plate remaias in this posi- 

 tion, micrometer M' is read for coincidence with the defining line at 

 this end of the bar. 



(4.) The carriage having been moved to the other end, the end- 

 measure standard is placed in position between the stops U and V. 



(5.) After firm contacts have been made between the stops and the 

 defining surfaces of the eud-measure standard, through the hand-wheel 

 R of the rack and pinion, microscope M^ is read for coincidence with 

 the defining line at that end of the bar. The difference in the two 

 readings will be the difference in the length of the two standards, pro- 

 vided the cylinders have no horizontal curvature. 



(6.) In order to eliminate the effect of horizontal cnrvatnre, the 

 comparison is again made with the line bar, placed at the same dis- 

 tance inside of the line U V. 



Good results have been obtained by this method, bnt it is of course 

 open to two objections ; first, that since the bar must be moved in 

 order to secure contact at V, the pressure at the two ends is unequal ; 

 and second, that the force required to make the contacts secure, may 

 pi'oduce indentations in the defining surfaces of the standard bar. It 

 will be seen from the series of observations with the end-measure 

 standard in melting ice, detailed on pp. 345-359, that the second ob- 

 jection will not hold if the steel is properly tempered. The diffi- 

 culty, if it exists, is easily remedied by the use of protecting pieces 

 having parallel surfaces. The first objection is a more serious one. 

 It is certainly open to criticism in a theoretical view, but the results 

 shown on page 350 seem to indicate that this objection is not a serious 

 one in actual practice. 



A second method of comj^aring line with end measures will need to 

 be described without the aid of a drawing. The apparatas is attached 

 to table S, Fig. 1. 



Three requirements most be met in the successful comparison of 

 line-measure with end-measure standards : — 



(1.) The position of the end-measure standard must not be disturbed 

 during the comparison. 



