404 Professors Liveing and Dewar on Sun-spots 
has one ultra-violet line (w.-l. 2852) which is far more easily 
expanded than any other of its lines. And for a large number 
of metals, when heated in the arc, we have observed that lines, 
or groups of lines, repeat themselves at intervals, and that of 
such series the alternate members are far more readily ex- 
panded than the others. This is the case with the doublets of 
the visible spectrum of sodium — which are alternately diffuse 
and sharp, the diffuse lines being more easily expansible — ^and 
with the triplets of the magnesium, calcium, and zinc spectra. 
In more complicated spectra it is very probable that the same 
thing occurs, but, from overlapping of the groups, the alterna- 
tions are not so easily detected. At any rate, they exhibit 
some lines which are more easily expanded than others. 
It is only necessary to watch with a spectroscope of suffi- 
cient dispersive power the spectrum of the arc while iron and 
other metals are feci into it, in order to see the great expansion 
of some lines and the comparative inexpansibility of others. 
Any theory which is to account satisfactorily for the expansion 
of lines must account for the difference in the amount of ex- 
pansion of the different lines of the same substance, as well as 
for the fact, noticed in several cases, that lines do not all expand 
symmetrically (some spread out more on one side than on the 
other), and also for absorption-lines expanding as well as the 
emission-lines, and for the fact that the tension of the vapour 
has more to do with the expansion of the lines than its tem- 
perature has. We do not know enough about the mechanical 
nature of the collisions, as they arc called, between the par- 
ticles of a gas, whereby the motion of translation due to tem- 
perature is supposed to be converted into vibratory mofions 
producing radiation, to say whether in a gas of high density 
the frecmency of the collisions between similar vibrating par- 
ticles may not affect the period of some of their vibrations 
much more than that of others. 
When we come to examine what lines are usually seen ex- 
panded in sun-spots, we see that a large number are those of 
iron ; and besides these the lines of magnesium, calcium, 
barium, sodium, titanium, and nickel are frequently enlarged. 
In fact the greater part of the lines of all these elements are more 
or less enlarged in most of the spots observed at Greenwich in 
1881. The hydrogen lines are sometimes enlarged, sometimes 
not. There may be, probably are, great variations of pressure 
in the region of a spot : and as it is not very likely that the 
material of the photospheric cloud contains hydrogen, the ten- 
sion of the hydrogen in the solar atmosphere will not be in- 
creased by the evaporation of the cloud. Hence the variations 
in the density of the hydrogen will be chiefly due to the cur- 
