ON A TABLE OF STANDARD WAVE LENGTHS OF THE SPECTRAL LINES. Ill 
4691.590 4690.326 7027.778 
.626 +.620 +2.785 
4690.904 4G90.952 7030.503 
4691.590 . 7040.092 
.626 -0.."'>17 
4690.904 7030.545 
Weight. h. 
t. 
1 4690.964 7030..-5G.? 
2 4090.958 7030.503 
2 4090.904 7030.515 
i^_ - — -" 
4090.962 7030.550 
The equation 3/^ — 2^=11.774 then rendily follows. 
Tables IV. and V. are from a 21| foot concave ^vith 7218 lines to the hirh, 
used on both sides, and thus equivalent to two gratings used on one side only. 
I have not yet determined theoretically whether the minor cnors arc perfectly 
neutralized in this manner: but it would evidently have a fcndency in thi^ 
d 
The photographic coincidences are given in the main table (X.), as not on 
the standards are compared by this process, but whole regions are photographed 
side by side. Both a 10,000 and a 20,000 concave were used for this work. 
Table VI. crives the collection of the equations relating to the visible .p-frum, 
the final results beinir given in Table VII 
The proper method of treating these twcntj-six equations would be by the 
jmed 
method of least squares. Bot it would be so long and tediou-s, and so b.ble (o 
mistake, that I have adopted the method of stnrting at one po.nt, and gn,ng to 
until all the equations are reached. Thus (T,l)le Xll), startnig with an n- 
value of e, we can calculate p, ■», I, J', h "> ^- r 7 f m a 
Using the eiglit values thus found once more, from f we have ^, .", ' '"' 
we have h,i,g; with similar results for the others. Collecting, we then hav /, 
.. /, y, ., I, 1 . „ „ i. Using these once ^^^ r^ '-']^:^ ^, 
standards. We could do this an, number ^^^^^^^^Z^^^ ,^^. ,-„%he 
but I thoudit this number was sufficient. Ihe second ( 
same manner, startino; from 
given ill Table VIII 
• TnhiP W Tnkim? the mean 
The results of the two calculations are give m Table . • ^^^^ ^^^^^^^ 
and adding the results of local micrometer measurement, 
marked "Relative Wave Lengths." 
