( 271 V 
(a, that by plunging it into the glycerine, the latter might bestirred 
as little as possible), while the other end was closed with a rubbei 
can The photograph « on plate II shows three drops, successively 
forced out of the same little tube. They appear to change their shape 
rather slowly. The one at the right hand is the oldest; there diffusion 
has proceeded farther than with the other. In the drops, solutions of 
different specific gravity seem to form, part of the contents moving 
downward, and another part upward. We observe this more clearly 
still in ft where three drops at a later stage are shown. Inside all 
these drops the optical density is a minimum, also in the tiny fila¬ 
ments where diffusion has proceeded tar. Everywhere the rule relating 
to case A is fully confirmed. 
Some instances of case B are represented in y, d, 8. The vessel 
was filled with the salt-solution, and a few drops of the optically 
denser solution of glycerine were forced into it. Owing to the smaller 
viscosity of the medium, the drops change and move more quickly 
in this case, than in the former one; and because the photographmg 
had to be delayed until a suitable stage of the diffusion process 
had been reached, these drops appear very irregularly lengthened 
out, as a rule. But the general character of the optical phenomenon 
is perfectly evident in this case too : a region of greater optical 
density appears dark toward the centre, and bright toward the edge 
of the source of light. All the thin filaments even, separating from 
the big drops in e, show the same peculiarity. 
And we see, that near the middle of the disk a condensation as 
well as a rarefaction can only reduce the brightness. If, therefore, 
we introduced into our vessel a great number of drops of greater 
or smaller optical density, so that the medium were honeycombed 
with density gradients, the central parts of the luminous field would, 
show a much lesser average brightness, than if the medium were 
homogeneous. Incidentally, circumstances might co-operate to repro¬ 
duce by refraction nearly the maximum brightness of the source, at 
certain places; those places would then highly surpass the average 
back-ground in brilliancy. 
Places considerably brighter than the average photosphere are 
really observed now and then, even on the central parts of the disk. 
This fact has been cited by FIsnyi ') as an argument against the 
Idea, that the light of prominences might be refracted photospheric 
light, because no optical system, no reflexion nor refraction, can ever 
make the intrinsic brightness of an object increase. Our above expla- 
*) Memorie 'della Soc. d. Spettroscopisti italiani XXXVil p. 182 (1908). 
