THE MASS-LUMINOSITY RELATION 
173 
At Hfi there is practically no interference, the scattered light of Sirius 
being weak. The displacements of Hfi measured on 4 different photographs 
(different methods of measurement giving 8 determinations in all) were 
found to be 
+ 31, 23, 24, 17, 31, 27, 28, 25 km. per sec. Mean + 26. 
At Hy it was found that the scattered light was just about equal in intensity 
to the true light of the companion; thus the line is presumably a blend in 
equal proportions of Hy for the companion and for Sirius. The measures 
should therefore be multiplied by a factor nearly equal to 2. The measured 
displacements of Hy were 
+ 13, 17, 2, 4, 8, 14, 12 km. per sec. Mean (corrected for blend) + 21 km. 
per sec. 
Fainter lines which could be measured gave, after multiplying by the 
factor for blend, the mean result -f 22 km. per sec. We have then 
General mean, Companion minus Sirius +23 
True Doppler Effect (orbital motion) + 4-3 
Einstein Shift +19 km. per sec. 
The device that was resorted to for Hy and the faint lines is, of course, 
unsatisfactory; but the evidence from Hf3 alone seems decisive. 
This observation is so important that I do not like to accept it too 
hastily until the spectroscopic experts have had full time to criticise or 
challenge it; but so far as I know it seems entirely dependable. If so, 
Prof. Adams has killed two birds with one stone; he has carried out a new 
test of Einstein’s general theory of relativity and he has confirmed our 
suspicion that matter 2000 times denser than platinum is not only possible, 
but is actually present in the universe. 
119. White dwarfs are probably very abundant. Only three are 
definitely known, but they are all within a small distance of the sun. It 
is only in rare conditions that we are likely to suspect, much less to establish, 
this condition of a star. In this book the phrase “ordinary stars” is to 
be understood to exclude white dwarfs; but if we say little about them it 
is because we know little and not because we regard them as a negligible 
minority. 
The bright component of o 2 Eridani suggests itself for an additional 
test. It is fainter than Sirius comes, but there is no bright star to interfere 
with its spectral measurement. The Einstein effect is, however, smaller. 
The companion of o Ceti has been suspected of being a white dwarf, but 
its spectrum seems to be peculiar and its nature is rather obscure. 
The conditions in the white dwarfs are outside the limits to which our 
theoretical investigations apply. In particular, our formula for k ceases 
to be a valid approximation and we have as yet no determination of the