456 
NEW OBSERVATIONS OF THE PLANET MERCURY. 
ing, and 
on the same side of its theoretic value, being behind time when the phase is increas- 
ahead of it when it is decreasing. This proves, then, that the discrepancy 
is due to a relative visibility or n on- visibility of the cusps, as compared with the centre 
of the terminator. 
Now, if the illumination at the cusps was relatively less than that at the terminator, 
dichotomy would 
before its predicted time 
geometrical consideration 
show 
omy 
We are therefore left with the alternative ; namely, that retardation of dichot 
d 
to relative prolong 
of the 
cusp 
And th 
can 
ly b 
geometrically to be the fact, as in the following diagram, where A B, Fig. 1, and B (7; 
Figs. 2 and 3, represent the loss at their respective phases. 
J^.r- 
tj 
^sS-IT. 
J'. 
37. Loss at the Cusps. — On the other hand, it is evident that there is some loss 
of visibility at the cusps. This is evident, both theoretically — otherwise there would 
be no loss at the centre of the terminator, since the cusps cannot be infinitely 
brighter than the centre 
and practically 
witnessed to by the measures of 
the polar diameter, which (Table IX.) show an increase steadily from the time of 
dichotomy to the time of coni unction. 
38 
with 
It varies with the ph 
The relative loss at the cusps 
This is shown by the following tabl 
