200 A. D. Bache on Declinometer Observations. 
end of the magnet at that time of the day, is deflected to the 
east in the northern hemisphere and to the west in the southern; 
In other words, in regard to direction, the law of the annual va- 
riation is the same as that of the diurnal variation, but opposite 
in soa from the northern to the southern magnetic hemi- 
sphere. 
Some interesting conclusions as to the law of change of the 
annual variation, will flow from a closer discussion of the obser- 
vations at the two hours of maximum, namely 6 to 7 A. M. and 
1to2p.Mm. The general table, by subtracting the annual mean 
from each monthly value, at the respective hours, gives the fol- 
lowing values for the annual variations at or near the hours of 
principal and secondary maxima of range, the signs + and — 
indicating as usual westerly, or easterly deflection from the an- 
nual mean position. 
6to7 lto2 
A. M. P.M, I 
‘ é 
+201 -0°98 || July, 
February, 41:31 —112 ucust, 
March, +047 -— 047 September, 
April, +015 | +106 : 
May, — 1:38 -++0:93 || November, 
June, — 1:90 +0°89 December, 
The greatest range at 6 to 7 A. M. is 5’-0 the easterly deflection 
being greater than the westerly by 04. That from 1 to 2 P. M. 
is 2'*7, the eastern and western deflections being equal. A gen- 
eral inspection of the columns shows that the solstices are approx 
imately the turning epochs of this annual variation, and that the 
signs change at the time of the equinoxes. ‘T'o determine these 
points with the more precision the numbers of the table were 
irene by an analytic formula. According to this, January 1st 
nd July 1st (ten days after the solstices) are the dates of the 
greatest values, and the transition from positive to negative val- 
ues, and the reverse will occur on the first of April and first of 
October (ten days after the equinoxes). . 
_A table is given in the memoir showing the satisfactory coit- 
cidence of the observed and computed values. This result agrees 
with that deduced by a different method by General Sabine. — 
To give a definite determination of the law of the phenom- 
enon, so as to embrace the whole progress shown by the series, 
the regular solar diurnal variation has been expressed as a func 
tion of the time, by four terms of Bessel’s formula, the equation 
being found for each month, and also for each half year from 
April to September and from October to March and also for 
the whole year. Allowance was made in determining the ¢0- 
efficients for the different weight of the readings at the even and 
at the odd hours. nota 
