186 
ON" A TABLE OF 
WAVE LENGTHS OF THE SPECTRAL LINES 
Remarks, Critical and Otherwise. 
In looking over this paper, the following thought occurred to me of a critical 
nature. First, it is to be noted 
observations have extended over ten years 
anv of 
been made for other purposes than for the use 
made 
of 
tliem. Ilencc the paper seems to lack unity of purpose, as many of the standards 
were added after the observations had begun, and do not appear in the earlier series. 
But it must be remembered that no good map of the spectrum existed at the time 
they were commenced, and I had no means of avoiding poor standards and selecting 
I 
do at present. To 
the old observations I was thus 
obliged to retain many poor standards. Again, for the method of coincidences only 
pour standards may exist in the limited region where the measurements were made. 
The effect of 
many gaps 
in the table. But many of the gaps 
also come from other causes which could not be avoided. 
The series of metallic plates w\as primarily made for measuring the spectra of 
metals. It is unfortunate for the present use that they nearly all begun and ended 
at the same ^place, overlapping very little. However, I have nearly always measured 
one or more plates to fill in the gap, and the coincidence of the values on these 
with the others shows that no error has been committed. 
Indeed, the accuracy 
with which the parts of this patchwork table fit into one another gives more con- 
fidence to me than if it had been made 
of rearular 
up oi regular parts aesig 
desigrned for their 
purpose 
