240 
corresponds with the mean variation for the whole period ; the 
nocturnal oscillation vanishing, and the course of the repre- 
sentative curve, from 10 P.M. to 9 a.M., being in all nearly 
a straight line. Thus the curves for the separate months ap- 
pear to distribute themselves into two groups, depending upon 
the position of the sun to the north or south of the equator ; 
and it is remarkable that the transition from one to the other 
system appears to take place abruptly, and almost per saltum. 
The whole range is nearly the same in each of the six 
summer months. The maximum range occurs in April, and 
its amount is 13/.3 ; this maximum is followed by a secondary 
minimum in July, which is succeeded by a secondary maxi- 
mum in August. There is a sudden change in the mag- 
nitude of the range from February to March, and again 
from October to November. The minimum range occurs in 
December, and its amount is 7’.0. 
The physical dependence of the phenomena of the changes 
of the declination upon the sun is evident from the fact, that 
they observe a diurnal and an annual period. In addition to 
this fundamental fact, it has been long ago observed, with 
respect to the diurnal change, that the time of the maximum 
of westerly declination follows the sun’s meridian passage at 
a nearly constant interval, and that the morning and evening 
minima are in like manner connected (although not so closely) 
with the hours of sunrise and sunset; and another point of 
connexion between the cause and effect has been established 
by the fact, long since observed, of the greater magnitude of 
the range in summer than in winter. 
Dr. Lloyd proposes to show that the sun acts by means of 
its heating power (as, in fact, is assumed both in the hypo- 
thesis of Canton and that of Christie) ; and that the connexion 
between the changes of declination and those of temperature 
is more intimate than has been hitherto supposed. 
The force which produces the deviation of the magnet 
from its mean position, at any moment of the day, is mea- 
