Wave-length of Light. 25 5 



derive the fullest benefit from the size of the telescopes ; over 

 and over again has the line in the spectrum appeared slightly- 

 displaced from the crosshairs, when no difference whatever 

 could be detected in the micrometer-readings. However, 

 there was gained the great advantage of using gratings of a 

 decimetre in length, giving spectra of great brilliancy and 

 superb definition, and which could be measured with vastly 

 greater exactness than is possible with the small gratings 

 generally employed. 



Gratings. 



Four gratings have been used in my experiments — two of 

 glass and two of speculum metal. The former are probably 

 the best of the very few glass gratings that have been ruled 

 on Prof. Rowland's engine. They are ruled on plane sextant 

 mirrors of rather hard glass. 



Grating I. contains 12,100 spaces in a length of very nearly 

 thirty millimetres, the lines being nineteen millimetres long. 

 It was ruled in Jan. 1884, at a temperature of 6°"7 C. ; gives 

 spectra of excellent definition, quite free from ghosts or false 

 lines, and having almost exactly the same focus on both sides 

 of the normal. 



Grating II. has 8600 spaces, with almost exactly the same 

 length and breadth as I., is free from ghosts and false lines, 

 and, like I., is very smoothly ruled, though it is somewhat 

 inferior to I. in the matter of regularity. The definition is 

 excellent, and the spectra alike in focus on both sides of the 

 normal. It was ruled in Nov. 1884 at ll°-6 C. 



Gratings III. and IY. are on speculum metal. The plates 

 are five inches square and five eighths of an inch thick, and 

 were worked plane with especial care. The ruled surface is 

 of the same size in each, four inches long by two inches 

 length of lines. 



Grating III. was ruled in April 1885, at a very nearly con- 

 stant temperature of 10° C. It contains 29,000 spaces, having 

 very nearly the same grating-space as II. It is a pheno- 

 menal grating both in its superb definition and extraordinary 

 regularity of ruling, and was selected from a large number 

 because of its very unusual perfection. The focus of the 

 spectra on each side of the normal is the same, and the ruling 

 is flawless. 



Grating IV. was ruled on the new dividing-engine just 

 completed by Prof. Rowland, and was one of the first large 

 ones completed. Although the new engine has even now not 

 received the finishing touches, it has turned out a few gratings 

 of remarkable excellence. One of these is IV., which was 



