258 
ME. J. NORMAN LOCKYER ON SPECTRUM-ANALYSIS 
1st observation, when the lines 4924 and 4911 were 
gone the gauge stood at ...... 30 millimetres. 
2nd 
55 
55 
55 
29 
3rd 
55 
55 
55 
29 
4th 
55 
55 
55 
31 
A rise to 34 millimetres was sufficient to restore the lost lines. 
Experiments with Chemical Compounds. 
Since it appeared that the purest and densest vapour alone gave the greatest number 
of lines, or, in other words, that the truly complete spectrum of a body is alone to be 
obtained upon the metallic pole itself where the vapour is densest and purest, it 
became of interest to examine the spectrum of a compound consisting of a metal com- 
bined with a non-metallic element. 
To this end a number of experiments were made, in which the metallic spectra were 
compared with those given by the same metals when combined with chlorine under the 
same conditions as in the former experiments. 
The compounds thus experimented on were as follows, the jar being used : — Li Cl, Na 
Cl, Mg Cl 2 , Zn Cl 2 , Sr Cl 2 , Cd Cl 2 , Ba Cl 2 , Pb Cl 2 , and Al 2 Cl 6 . It was found in all cases 
that the difference between the spectrum of the chloride and the spectrum of the metal 
was : — That under the same spark-conditions the short lines were obliterated, while the 
air lines remained unchanged in thickness. 
Changing the spark-conditions by throwing the jar out of the circuit, this change was 
shown in its strongest form, the final results being that only the very longest lines in 
the spectrum of the metal remained. 
The following are the details of the experiments made under these conditions : — 
Method of Observation. — Some pieces of stout aluminium wire 10 millims. long and 
3 millims. in diameter were taken; one end was flattened for about one third of the length 
for the purpose of inserting it in the spark-holder, and the other was drilled down in the 
direction of the axis for from 2 to 3 millims., and thus formed into a small conical cup ; a 
very fine hole was then drilled through the side of this cup at the bottom and the flat- 
tened end carefully split. Through the lateral hole a piece of platinum wire 0-5 millim. 
in diameter was passed and one end brought round through the split end of the aluminium, 
while the other was brought up the centre of the cup. The split was now closed by strong 
pressure in a vice, and the ends of the platinum wire cut off. The whole now presented 
the appearance of a small candle, the platinum wire representing the wick: the accompany- 
ing figures (figs. 4, 5) will render the preceding statement clear*. Round this wick the 
chloride in fine powder was tightly rammed down. [A similar cup, without the wick, was 
used for the examination of the spectra of metallic barium, strontium, and lithium, 
* The object of the wick was to confine the spark to the centre of the dry chloride. Before it was adopted 
the spark was very unsteady, leaping about from side to side of the cup. 
