ASTRONOMY: A. S. KING 
373 
little iron vapor in the tube-arc, the displacement becomes very small, 
although stronger and wider lines on the same plate retain their one- 
sided character. 
It has long been known that a strongly condensed spark tends to 
produce unsymmetrical Hnes in its spectrum. The writer has made a 
study of this effect using a powerful and very disruptive transformer 
spark. The diss3anme tries produced are strikingly similar to those 
of the tube-arc. Examination of different regions of the spark showed 
the dissymmetry to be such as would result from the interior vapor of 
the spark giving a Hne displaced toward the red, while the outer vapor 
gives a symmetrical line coinciding with the usual arc line. Tube-arc 
lines displaced toward the violet show strong shading on that side in 
the condensed spark. 
For the cause of these dissymmetries, which amount in each case 
to a displacement of the maximum of the spectrum line, it appears 
that we must look to an agency other than pressure. Dissymmetries 
in the spark have sometimes been ascribed to this, but they correspond 
closely to those of the tube-arc, which is operated in a partial vacuum. 
Also, in both sources there is no tendency for the displacement to in- 
crease in magnitude with increasing wave-length, which is a feature of 
the pressure effect. 
A high vapor density increases the dissymmetry of tube-arc lines, 
but apparently cannot by itself cause such dissymmetry. Evidence 
on this point is supplied by the electric furnace, the spectrum lines of 
which are notably symmetrical with high vapor density. In the arc 
also, lines may be very wide when much vapor is present, and yet 
remain symmetrical. When the poles of the arc are approached, how- 
ever, the occurrence of enhanced lines indicates a condition resembHng 
the spark, and in this region the lines more sensitive to displacement 
become one-sided, an effect which is magnified in the tube-arc and in 
the disruptive spark. 
At present these effects seem to be harmonized by considering as a 
necessary condition the presence of electrified particles moving at high 
velocities, these being given in the arc and spark by the strong potential 
gradients and in the tube-arc by the enormous consumption of energy 
by the ruptured carbon tube. If such swiftly moving electrons are 
present, the effect of their velocities in causing a disturbance of vibration 
period should be increased by the crowding together due to high vapor 
density. On this account the relation of the effects here described to 
the effect of the crowding due to pressure may be a close one, as is borne 
out by the similarity of the observed displacements. 
