﻿an 



d the Buhner Series Spectrum. 545 



highly transparent to the ultra-violet region in the vicinity 

 of the end of the Balmer series. 



The grating was selected by illuminating it with the light 

 from a quartz-mercury arc, mounted in a closed box provided 

 with a window of dense nickel-oxide glass. This arrange- 

 ment gives a powerful beam of radiation of wave-length 3660, 

 which was rendered convergent by a lens and reflected from 

 the grating to a screen of barium platinocyanide. 



In this way a grating was found which was enormously 

 bright in the 5th-order spectrum for the region of the end of 

 the Balmer series (X = 3676). 



With this grating and the new tubes of pyrex glass, the 

 20th line was photographed with certainty, and probably the 

 22nd line. The 21st is so nearly in coincidence with a strong- 

 line of the secondary spectrum that it cannot be identified 

 under present conditions. 



H a and H^ were not photographed, as we have Curtis's 

 values, which are of the same order of accuracy as those 

 about to be given. 



The plates were measured and the wave-lengths computed 

 by Mr. Arthur E. Ruark, one of my students, to whom I am 

 indebted for a large amount of very faithful and accurate 

 work. The reference iron-lines were the tertiary standards 

 determined by St. John and Babcock *, lines showing no 

 polo effect being used in most cases. 



Reduction to vacuum was done by the table of Meggers 

 and Foote t- 



. , . N \ f \ Difference between these values 



Line - X («*>■ X < vaC -)- and those of Curtis. 



1 *6562-793 



2 *4861-326 



3 4340465 4341681 -0001 ] „ m 



4 4101-731 4102-884 -'007 .„ , ,° , 



5 3970-073 3971192 - -002 \ 0. s observed 



6 . 3889064 3890-161 + "012 J wave-iengtns. 



7 3835-397 3836-481 + -010 ") 



8 . 3797-910 3798-984 + -010 



9 3770-634 3771-701 + -001 



10 3750-152 3751-214 + '002 | 



11 3734-371 3735-429 -000 



12 ... 3721-948 3723003 4- -007 



13 3711-980 3713032 4- "007 



14 3703-861 3704-911 + -006 , 



15 . ... 3697-159 3698-197 4- '005 I wa ™- len g tlis 



16 . . 3691-553 3692-600 + -004 I 



17 3686-833 3687-878 - -001 ! 



18 3682-825 3683-869 4- "015 I 



19 3679-372 + -015 | 



20 3676378 + -013 J 



* "Values given by Curtis. 



From 

 )■ C.'s calculated 



* Astropbj8. J. vol. liii. (1921), or Mt. Wilson Contrib. No. 202. 

 t Bur. of Standards Sci. Papers, No. 327 (1918). 



Phil. Mag. S. 6. Vol. 44. No. 2G1. Sept. 1922. 2 N 



