396 PROCEEDINGS OF SECTION B. 
primary battery ; but on including the secondary condenser, they 
were at once apparent.* From the blurred nature, want of 
delicacy, and general resemblance to air-lines, it seemed probable 
the two lines under notice might be due to the oxide of the 
metal. This conclusion was strengthened by careful considera- 
tion of the spectra and physical properties of the natural group 
of elements, Be, W7g, Zn, Cad, He. 
For example, if the spectra of this group be examined, con- 
siderable homology is observable between them, the refrangibility 
of the principal lines of each element increasing inversely as its 
atomic weight. It is worthy to note, in this connection, the 
absence in the spectra of Be and Mg of lines corresponding to w.7. 
4923-8, 4911°2 in Zn, and 5378°8, 5337-5 in Cd. A possible 
reason for this is to be found when the melting and _ boiling 
points of the elements of this group are considered. Similarly 
with their oxides ; but with this great difference, that, whilst the 
metals melt and vaporise at comparatively low temperatures, the 
oxides, for the most part at least, are non-volatile. 
Now, whether the lines in Zz and Cd are due to the oxide or to 
the metal, it seems probable, from what has already been noted 
in the case of Zz, that if the spectra of Be and AZg were examined at 
sufficiently high temperatures, by the use of a more powerful 
induction apparatus and extra condensers, lines corresponding to 
those of Zz and Cd would be found. 
The fact that Ca is the most volatile element of this group, 
except “7g, led me to examine its spectrum, in order to find 
whether the lines w./, 5378°8, 5337-5 would be influenced by the 
temperature of the spark, as in the case of Zz ; it seemed probable 
they might be visible at a lower temperature. On subjecting 
that element to the action of the spark the lines w./. 6438, 5085, 
4799, 4676°8 were the only ones visible ; 5378-8, 5337-5 required 
the condenser. Another trial was made without the condenser, 
increasing the strength of the primary battery. At this tempera- 
ture they were seen as a nebulous band. 
It has been stated elsewheret that zzsoluble and non-volatile 
substances do not yield spark spectra, and as this statement rather 
upsets the present considerations, zinc oxide, specially purified, 
was submitted to the action of the spark. It gave lines identical 
with those of metallic zinc, both at low and high temperatures. 
In order to settle the question as to the lines w./, 4923-8, 4911-2 
in the zinc spectrum being due to the oxide, there appeared only 
one course open. The spectrum must be observed in an atmosphere 
free from oxygen and its compounds, pure dry hydrogen being 
thought most suitable for the purpose. 
* Jn a photograph of the are spectrum of zinc, by J. R. Cafron, F.R.A.S., where the arc 
‘was produced by 40 Groves’ cells, no trace of the lines, w.l, 4923°8, 4911°2, is shown. This 
will give some idea of the extraordinary influence of a condenser in raising the tempera- 
ture of the spark. 
+ Professor W. N. Hartley in a paper on “Spectrum Photography,” Phil. Trans., 18&4. 
