Ejfed of Evro)\< in lialiiKj on a Dijfraction-Gradnrf. 551^ 



})lia?o. I liavc obtaiiKMl an interesting series of tliese, but as 

 the examples alreaily given are siitHcient for the ])urposes of 

 this paper, I will hold it over for publication at an early 

 date. 



15. It is interesting to note that in the device I have 

 described the quantity whose wave-form is to be determined 

 and analvsed is first differentiated with respect to the time, 

 then integrated by the galvanometer, the limits to the in- 

 tegrals being assigned by the commutator. It is possibly 

 that the same device miglit be applied to determine the 

 wave-forms of periodic quantities other than those that arise 

 in alternate-current working. All that would be necessary 

 for this would be to obtain an E.M.F. or a current pro- 

 portional to the time-rate of variation of the quantity ^^■hose 

 wave-form is required. 



In conclusion I wish to thank my friend Mr. W. Stone, 

 Electrical Engineer to the State Railway J)epartment^ for his 

 inA'aluable assistance. Xot only was the commutator which 

 was used in getting the preceding results built under his 

 supervision, but he, at great per>onal trouble, put at my 

 disposal one of his Parsons turbine-driven generators, and 

 assisted in the experiments whereby a number of the wave- 

 forms given in this paper were obtained. 



' smu 



LXX. Tlie Effect of Errors in Ruling on the A2:)2?eajxince of 

 a Diffraction- Grating. By H. S. Allex_, M.A., B.Sc."^ 



[Plate XXIV.] 



Part II. 



IX a former paper f imder this title 1 drew attention to the 

 bands which can generally be seen when a diffraction- 

 grating is viewed by the light going to form a single spectral 

 line, and explained the formation of these bands at points of 

 the grating where a change is taking place in the value of 

 the ruling. This method has proved of service in locating 

 the errors in Lord Blythswood's dividing-engine ; and has 

 been employed b}' him in determining the correction required 

 to eliminate the residual periodic error arising from the 

 ratchet-wheel and thrust-block. I have therefore thought fit 

 to give an account of a modification of the method formerly 

 described, which renders it possible to determine directly the 

 parts of the grating where the ruling is wider (or narrower) 

 than the average, instead of the places at which the value of 



* Communicated bv tlie Author. 



t I'liil. 31a--. p] vol. ill. p. 'd2 (iD02j. 



