OF ARTS AND SCIENCES. 239 



Rutherford UUrractiou pliito, and a plate ruled by uiyself for the pur- 

 pose of investigating the errors of my screw. The measures were all 

 made with a f\^ objective, and an eye-piece micrometer, 200 to the inch, 

 tlie lines of which were about x5ff(T«T o^ ''^^ '"^^ ^^ breadth. Using a B. 

 ocular, the value of one division was found to be xssiyff o^ ^^ ^"^b* 

 With this arrangement, it was found easy to measure any given space 

 to yuJ^y^y of an inch with considerable certainty. To eliminate errors 

 in the micrometer, the same divisions were used in all comparisons. 

 In the Nobert plate, the width of the lines is about xgjjj^ of an inch ; 

 in the Rutherford plate, s^^jj^ of an inch ; and iu my own, 55 J jj^ of 

 an inch. The space measured was ^^^ of an inch. 



Table I. contains the residuals obtained by subtracting the meas- 

 ure of each space from the mean of all the spaces. 



Table II. contains the residuals obtained by subtracting the error 

 of each space from the error of the next consecutive space. 



Table III. contains the periodic errors deduced from Table I. 



