392 
DIFFUSE MATTER IN SPACE 
We add for comparison the corresponding results for V Puppis and 
Krueger 60 adopting V = 8 and V = 50 km. per sec. respectively, which 
are about the average velocities for stars of their class. 
Star 
dM/dl 
L/c 2 
Ratio 
V Puppis 
1-8.10 11 
2-9.10 16 
165000 
Sun 
4-8.10 8 
4-2.10 12 
8800 
Krueger 60 
1-9.10 7 
4-7.10 10 
2500 
Probably R ought to have been taken greater than the photospheric 
radius of the star. I suppose that if an atom from space encountered the 
solar corona it would be tangled up with it sufficiently to be captured by 
the sun. But allowing for all uncertainties it seems clear that the loss of 
mass by radiation cannot be compensated by accretion in any ordinary 
type of star. 
267. Loss of mass by escape of atoms has been considered by a number 
of writers. According to E. A. Milne* the loss from a star by reason of 
thermal velocities is quite negligible for all elements. There seems to be 
some possibility of escape of the chromospheric atoms acted upon by 
intense radiation pressure; but calculations are not as yet very definite f. 
But in any case it is difficult to believe that the loss by escaping atoms 
can be at all comparable with the loss of mass by radiation. The radiation 
of the sun carries away a mass of 6.10 -11 gm. per sq. cm. per sec. We have 
found (§ 254) that the average density of the calcium chromosphere is of 
order 10~ 17 gm. per cu. cm.; so that the whole chromosphere would have 
to move steadily outwards at 60 km. per sec. in order to carry away as 
much mass as the radiation does ! 
The radiation of mass of the sun is equivalent to the escape of a billion 
calcium atoms per second from each square centimetre of its surface. 
It seems then that change of mass of a star by radiation far outweighs 
any material accretion or loss. If that is so, the calculation of the duration 
of stages of evolution in Table 41 must be accepted as definitive. As 
already explained grave difficulties then arise as to the coexistence of 
giants and dwarfs in the same cluster; it seems almost necessary to throw 
over the idea of any important advance in evolution in the life-time of the 
clusters, and it then becomes a question whether there is any point in 
retaining the idea for stars in general. Somewhere in the present tangle 
of evolution and sources of energy I have been misled; and my guidance 
of the reader must terminate with the admission that I have lost my way. 
* Trans. Camb. Phil. Soc. 26, p. 483 (1923). 
t The problem is discussed by M. C. Johnson, Monthly Notices, 85, p. 813 (1925)-