198 W. A. Norton onthe Variations of the Declination 
17h «618k = 19R ROK IK RRR BKK lA 2h 3h 4h 
Fig. 6 
oronte 
RS 
A 
oy 
eee) err 
17k =: 18h 19h 20h Qik 92h C 3h 
These curves show the deviations of the declination, at the ified hours, in the 
mon om the mean declination in ae the months and at all the hours, 7¢p- 
resented by the Horizontal line MM. An ordinate lying above MM shows that the dec- 
lination tc at that hour is less than the mean, one lying below that it 1s greater 
than th ay; 2, June; 3, July; 4, August 5, November ; 6, Decem- 
ber; 1, silane 8, February. Scale 0in5 to 1! of ar 
gram, (Fig. 6). It should disappear at noon, and recur in the 
afternoon, ,but we shall soon see that antagonistic causes come 
into operation in the afternoon, which prevent the same result 
‘om being realized. We shall see also that the ——. 
7 to 8 a.m, and the forenoon hours geverdly, 38 ssing from 
the summer to the winter months, is not wholly deete the shift- 
; of the radial current from the north to to the an ee e of the 
J vertical, as we have on sidered it. 
— 
eT. ad 
