a connection between the two phenomena which we have now 
considered each separately. 
Thoagh, after what we have seen of the different results for the 
pressure-effect, we can directly say that there is no question of a 
direct proportionality between the two effects, we may still make 
up these ratios, as in any case they will teach us something about 
the ratio of the magnitude of the two effects in the different groups. 
As far as there existed observations of both effects for the same 
spectrum line, I have therefore calculated these ratios and put them 
Ce Ae a 
in the columns : , — and —. For each line I have given in 
B’ B, D, Dp 
the last column the mean of the number occurring in each of the 
other columns separately. 
Even if we confine ourselves to one and the same column, a 
preliminary consideration shows already that these ratios have very 
different values. The import of the “Means” given below must there- 
fore be taken in a wide sense. Fig. 2 clearly shows how the ratios 
are divided in the different groups. 
he Ae TT eee Group I 
a Group II 
5 == GroupslV, 
mee =| Groups 
v 
a4 j 
E is 
= ij i \ 
en a 
1 - i \ 
PN ee \ ZE 
Se ee e se DE 
BSA SS 
Group L shows a maximum at + 150 with which smaller 
maxima are connected. In this group there occurs a maximum 
quite separately at + 450 where 4 lines are found which have 
nearly the same mean ratio between the Znrman-effect and the 
pressure-effect viz. 
2 3561,92 ; 3624,69 ; 3739.38 ; 3783,67 with a ratio 
456 451 446 450. 
With the exception of one value of a line which shows Pr. E., 
but no Z.E., the ratios of group II vary between 136 and 356, 
that is between narrower limits than those of group I. The highest 
maximum lies at 230. 
As of group IIL only two ratios could be calculated, this group 
has not been taken up in the figure. From these two means: 110 
and 59 we can infer little. 
