414 Scientific Intelligence. 
this tap being carefully evacuated. Now the side branch was 
closed, and the tap leading to the supply was opened. Then we 
observed that the light of the electric discharge in the vacuum 
tube was at first greenish, and after a while grew yellow. By 
cutting off the current of gas after a sufficiently short time, we 
succeeded in making a vacuum tube which remained greenish. 
On examining it in a small spectroscope with which we could over- | 
look the whole spectrum, we found that the intensities of the lines 
had changed. ‘The yellow line was scarcely as bright as the green 
line 5016, and the red line 7065 had apparently decreased relatively 
to 7282 and 6678, although it was still stronger than 7282. The 
two lines that had decreased in intensity belong to the second set 
of series, while the others are members of the first set. The other 
visual lines of the second set could not very well be examined 
because they are more in the violet part. 
This observation confirms our spectroscopic result. The gas in 
cleveite may be taken to be a mixture of two gases of different 
density, of which the lighter one is more rapidly transmitted 
through the plug of asbestos. There is, however, the objection 
to be raised, that in the green tube the pressure is less, and that 
the difference of intensities is due to the pressure being different. 
This must be further inquired into. 
We were not satisfied with the visual observation of the change 
of intensities in our green tube, but thought it desirable to test 
the conclusion by the bolometric measurement of the two lines 
that we have discovered in the ultra-red part of the spectrum. 
If we were right, the ultra-red line of smaller wave-length, which 
belongs to the second set of series, ought to have decreased in 
intensity relatively to the other ultra-red line. This we found to 
be so indeed. In the yellow tubes the intensity of the smaller 
wave-length was to that of the other on an average as 3 to 1, while 
in the green tubes it was as 1°8 to 1. This confirmation we con- 
sider the more valuable as it does not depend on any estimation 
which may be biassed by the personal opinion of the observer, 
but is based on an objective numerical determination. 
Another contirmation may be gathered from the spectrum of 
the sun’s limb and that of several stars. Let us confine our 
attention to the six strongest lines in the visible part of the spec- 
trum : 
7066, 6678, 5876, 5016, 4922, 4472. 
The first, third, and sixth belong to the second set of series; 
the second, fourth and fifth to the first set. These six lines have 
all been observed in the spectrum of the sun’s limb, as Norman 
Lockyer and Deslandres have pointed out. Now, according to 
their appearance in the spectrum of the sun’s limb, they may be 
classed in two groups, one group being always present, the other 
group being sometimes present. OC. A. Young long ago called 
attention to the difference in the frequency of appearance of the 
chromospheric lines. He has given them frequency numbers, 
