lit ON- A TABLE OF STANDARD WAVE LENGTHS OF THE SPECTRAL LINES 
It is seen that some of tlie plates have only one standard upon them. With a 
plane grating 
Id be impossible to work them np, but with the normal spectrum 
produced by the concave grating only one is necessary, as the multi|)lier to red 
reading 
ly a constant. In working up a whole series of 
plates, there is no trouble in giving a proper value to the constant for any pi 
w 
the series which has only one standard. 
PJute 17 was measured twice by two dividing engines, and as it was a specially 
goorl plate, each measure was given a weight equal to one of the other plates. 
The principal error to be feared in these plates is a displacement of the instrument 
between i\\Q time of the exposure on the two spectra. This was guarded against by 
the method above described. In Plates 17 and 20 there was a portion of the plate 
on which both the spectra fell all the time, and thus gave a test of the displacement. 
This was found to be zero. The other plates overlap so much that there are gen- 
erally two or more determinations of each line. A comparison of these values shows 
little or any systematic variation in the different plates exceeding yj^ division of 
Angstrom. Plates IG, 17, 18, and 5, 6, 8, all give the region 3900 as derived 
from 5200 and 5850, and thus give a test of the relative accuracy of these latter 
regions. It is seen that the two results of the region 3900 differ by about .015 
division of Angstrom. Were the wave lencrths of the re^-ion 5170 to 5270 to be 
increased by .020 the discrepancy would cease. The amount of this quantity seems 
rather large to be accounted for by any displacement of the spectra on the plates, 
but still this may be the cause. Again, it is possible that different gratings may 
give this difference of wave lencrth from the cause I have mentioned above. This 
Ih 
Lve shown, exists in the same degree in plane gratings as in concave. 
I have not attempted to correct it in this case, but have simply taken the mean 
of the two values for the region 3900, and so distributed the error. This is the 
greatest discrepancy I have found in the results, except in the extreme red. 
Thus the region 3100 to 3200, a portion for which Plate 20 is to be relied upon, 
gives the wave length of the ultraviolet .01 division of Angstrom hi-her from the 
O 
f 
4200 than from 6300. As the discrepanc 
C'^"'""" '"O 
befo 
tue concave gratmg were often a whole division of Angstrom, I have regarded this 
result as satisfictory. Indeed, until we are able to make all sorts of corrections due 
to the change in the index of refraction of the air with the barometer and ther- 
mometer, it seems to me useless to attempt further accuracy 
\i 
of photographic plates into the table, especially the Ion 
required for metallic spectra, it becomes necessary to correct them for the departi 
■J'. 
