( 437 ) 
give the graphical representation of 4 as a function of d for each 
of these three cases. 
The result of the observations is represented in figs. 4—8. 
These drawings are made with the aid of photographic negatives. 
We have not yet succeeded in getting negatives that showed all 
details simultaneously and equally well. Hence ocular observations 
had to supply the imperfections of the photographic records. 
Figs. 1, 4, 5 refer to the triplet (type line D, in feeble fields) ; 
figs. 2, 6, 7 to the quartet (type line D,); figs. 3, 8 to the sextet 
‘type line JD,). 
When comparing the results of observation with theory, it should 
be taken into account that the theoretical curve indicates the distortion 
which one single interference band would undergo. With the method 
of observation used, the central part of the field of view contained 
also parts originating from bands lying higher and lower than the 
one considered. The theoretical figure must therefore be completed 
with parts of theoretical curves lying above and below the one 
represented. 
We will first of all consider the quartet. We indicate the bands 
by a, 6, c, a being the superior one, and by 1, 2, 3, 4, we indicate 
the positions in the spectrum which would be occupied by the com- 
ponents. The double curved line between 2 and 3 shows entirely 
the same character in both figures. This sinuous line (figs. 6 and 7) 
thickens out at the extremities into more intense parts (where the 
double refraction is at a maximum or at a minimum) turning their 
concave side towards band 6. These intense parts correspond to the 
loop of the theoretical curve, the loop between 1 and 2 belonging 
to band ce, and the one between 3 and + to band a. It was not to 
be expected that the two branches which asymptotically approach 
the components, would be seen separated from the loops. The distance 
is too small by far to allow that. The vertical central line in the 
figure is the reversed sodium line due to the are. With increased 
vapour density the loops increase their distance from their band. 
Fig. 7 relates to this case, which is also in accordance with theory. 
As the vapour deusity increases, fewer details become visible, but 
we will not go further into this point now. 
The observations concerning the sextet are very difficult on account 
of the extremely small distance of the components. It is already 
difficult to observe the inverse sextet, and hence so much the more to 
observe phenomena occurring between its components. Only under 
very favourable circumstances could the phenomenon be observed 
as it is represented in fig. 8. The other phenomena observed with 
