GRATINGS FOR OPTICAL PURPOSES 491 



spectrum. Extremely brilliant gratings have been made with 43,314 

 lines to the inch, and there is little difficulty in ruling more if desired. 

 The following show some results obtained: 



Flat grating, 1 inch square, 43,000 lines to the inch. Divides the 

 1474 line in the first spectrum. 



Flat grating, 2X3 inches, 14,438 lines to the inch, total 43,314. 

 Divides 1474 in the first spectrum, the E line (Angstrom 5269-4) in 

 the second and is good in the fourth and even fifth spectrum. 



Flat grating, 2X3 inches, 1200 lines to one millimetre. Shows very 

 many more lines in the B and A groups than were ever before seen. 



Flat grating, 2 X 3 inches, 14,438 lines to the inch. This has most 

 wonderful brilliancy in one of the first spectra, so that I have seen 

 the Z line, wave-length 8240 (see Abney^s map of the ultra-red region), 

 and determined its wave-length roughly, and have seen much further 

 below the A line than the B line is above the A line. The same may 

 be said of the violet end of the spectrum. But such gratings are only 

 obtained by accident. 



Concave grating, 2X3 inches, 7 feet radius of curvature, 4818 lines 

 to the inch. The coincidences of the spectra can be observed to the 

 tenth or twelfth spectrum. 



Concave grating, 2X3 inches, 14,438 lines to the inch, radius of cur- 

 vature 8 feet. Divides the 1474 line in the first spectrum, the E line 

 in the second, and is good in the third or fourth. 



Concave grating, 3 X 5 inches, 17 feet radius of curvature, 28,876 

 lines to the inch, and thus nearly 160,000 lines in all. This shows 

 more in the first spectrum than was ever seen before. Divides 1474 

 and E very widely and shows the stronger component of Angstrom 5275 

 double. Second spectrum not tried. 



Concave grating, 4 X 5f inches, 3610 lines to the inch, radius of cur- 

 vature 5 feet 4 inches. This grating was made for Professor Langley's 

 experiments on the ultra-red portion of the spectrum, and was thus 

 made very bright in the first spectrum. The definition seems to be 

 very fine notwithstanding the short focus and divides the 1474 line with 

 ease. But it is difficult to rule so concave a grating as the diamond 

 marks differently on the different parts of the plate. 



These give illustrations of the results accomplished, but of course 

 many other experiments have been made. I have not yet been able to 

 decide whether the definition of the concave grating fully comes up to 

 that of a flat grating, but it evidently does so very nearly. 



