164 
DR. T. E. MERTON AND PROF. J. W. NICHOLSON ON 
But the important fact for our present enquiry is that the role played by a large 
quantity of Hydrogen is directly contrary to that played by a small trace, and we 
may argue that the mechanism of inter-action of the two gases is quite different in 
the two cases. A definite phenomenon has been quantitatively isolated which demands 
for its appearance only a spectroscopic “ trace ” of one of the acting gases. 
This reversal of the effect of a trace of Hydrogen, by the admixture of more 
Hydrogen is, however, interesting in another way, for it introduces us to a striking 
difference of behaviour between Diffuse and Sharp series. Inspection of Table XHI. 
indicates that the line X7065 shows very little tendency to change in relation to 
A.4713 by the action of this Hydrogen—or at least that the change in the Diffuse 
series is of quite another order. Moreover, the change among the lines XX7065, 4713, 
4121, though comparatively small, is quite definitely present as a combination of two 
effects. For X4121 is reduced relatively to X4713, as by the effect of the trace 
of Hydrogen, while X4713, as against X7065, is quite definitely enhanced. It 
seems that the Sharp series under these circumstances is just ceasing to show the 
first phenomenon, due to the trace of Hydrogen, and commencing to show the second, 
so that if the quantity of Hydrogen were increased yet further, the second might 
predominate. In other words, the essential difference between the Diffuse and Sharp 
series is that in the latter case a more considerable admixture of impurity is needed 
to produce the effects observed in the Diffuse series. Sharp series are in fact sensitive, 
to an equal extent with Diffuse series, to the infiuence of a trace of Hydrogen, but 
not to a comparable degree to the different mechanism of interaction with large 
quantities of Hydrogen. We feel no doubt that the available data can be summarised 
in this way, for the phenomena shown by the Parhelium spectrum follow the same 
course throughout. 
The Principal series of Helium, showing only one member X3888 is not, of course, 
capable of test in this manner by the present experiments. 
Princi'pal Seo^ies under the same Conditions. —The Diffuse series of Parhelium 
contains five members on some of our plates, and we can therefore make by its use a 
much more exhaustive test of the conclusions outlined in the preceding section. It 
is also possible to obtain information relating to Principal series, and this will be our 
first object in the present section. Since there are only two visible members, in these 
experiments, the Table (XV.) is very short. 
Table XV. 
X. 
(a). 
(b). 
(c). 
(4). 
5015 
10 
10 
10 
10 
3965 
1-92 
2-08 
1-85 
2-4 
