Note. 203 



that the designinj^ and computing of first-class optical systems must 

 of necessity be checked by workshop trials. By the methods of 

 design introduced by Abbe the concrete embodiments of a system 

 produced in a workshop do not test the conclusions arrived at by 

 the designer and computer, so much as they serve as means for 

 determining the accuracy with which the mathematical results can 

 be realized in the workshop — i.e. in first-class optical designing the 

 designer tests the work of the workman, the workman does not 

 test the work of the designer. 



p 2 



