232 
PHYSICS: J. A. ANDERSON 
Proc. N. a. S. 
outside of the wooden block. By very slightly amalgamating the surface 
of the wires they explode normally, giving fine continuous spectra crossed 
by their own absorption lines, the lines due to Hg not having been observed. 
That a layer of these metallic vapors only a few centimeters thick should 
give a continuous spectrum is somewhat surprising, especially so, since the 
average pressure is certainly much lower than it was earlier^ thought to be. 
Professor Henry Norris Russell suggested that it would be of some interest 
to find out whether the vapors are transparent to radiation or not. The 
following experiment was accordingly performed: A short spark gap con- 
nected in series with the wire to be exploded was (a) placed immediately 
in front of the wire, so that the light from the explosion had to pass through 
the spark on its way to the spectrograph ; (6) placed immediately back of 
the wire, so the light from the spark had to pass through the explosion 
vapors before reaching the spectrograph. The spark and explosion would 
be exactly simultaneous, since they were connected in series. Iron wires 
were used, and the spark terminals were made of brass in order to make 
use of the bright emission lines of zinc throughout the spectrum and those 
of copper in the extreme ultra violet. 
Using the arrangement (a) the spectrogram showed the bright zinc and 
copper lines very distinctly superposed on the regular iron absorption 
spectrum. It was also clear that the iron fines were less dark than usual, 
that is they were partially filled up by the continuous background in the 
light from the spark. With the arrangement (b) no trace of the zinc or 
copper lines could be seen, the iron spectrum being exactly the same whether 
the spark gap was used or not, thus showing that the light from the spark 
can not pass through the explosion vapors. 
This experiment proves that 4 centimeters of iron vapor as here used is 
perfectly opaque, but it does not show how far the light from the spark was 
able to penetrate the vapor before it was absorbed, in other words, the 
absorption coefiicient has been shown to be fairly large but has not been 
measured. Other observations make it probable that a layer about 2 mm. 
thick is very nearly opaque — but further experimental work is required, 
and will be undertaken as soon as possible. 
1 Proc. Nat. Acad. Sci., 6 1920, pp. 42-43; Astroph. J., 51 1920, pp. 37-48. 
2 The thermo-couple was kindly placed at my disposal by Mr. Edison Pettit. 
3 Astroph. J., 51 1920, pp. 44-46. 
Acknowledgment. Since my Note on the Definition of a Linear Functional was 
published in the February number, I have learned that the principal theorem stated on 
page 27 has been proved by Frechet by means of Fourier series, and that it is also a 
special case of a theorem by F. Riesz. See Trans. Amer. Math. Soc, 8, p. 439, and 
Math. Ann., 69, p. 475. Chari^ks A. Fischer 
Erratum. Page 145, line 20. For "the majority of the stars of" read about 20% of 
the stars of the. 
