Errors of a Straight Scale. 335 



much, as we shall see further on. First, however, I will briefly 

 describe Gill's method of observation and reduction. The 

 latter admits of demonstration without resorting to extensive 

 mathematical formulae. 



The scale under investigation, A, and the auxiliary scale, B, 

 are supposed to be graduated as nearly alike as possible, and to 

 be mounted so that one can be compared with the other. The 

 observations are now effected in a series of separate operations, 

 the diagrams given below showing how they would be arranged 

 for a scale of four spaces. 



Operation 1. 

 A 12 3 4 



e d c b a B 



Operation 2. 



12 3 4 



e d c b a B 



Operation 3. 

 12 3 4 



d c b a B 



Operation 4. 

 12 3 4 



e d g b a B 



Operation 5. 

 12 3 4 



d c b a B 



Operation 6. 

 12 3 4 



e d c b a B 



Operation 7. 

 12 3 4 



e d g b a B. 



