604 LAMONT: MAGNETIC RESULTS 
sixty days (two months), of which the results are exhibited gra- 
phically in the accompanying Plate, in which. the hours of the 
day are indicated by dotted, and those of the night by full lines. 
We see that the diurnal movement differs considerably in 
summer and in winter. In summer the march towards the west 
begins between 7 and 8 a.m., and attains its maximum at | P.m., 
or very soon afterwards. Thence it retrogrades until about mid- 
night, when a small westerly movement takes place until be- 
tween 2 and 3 A.M., when it again retrogrades until between 7 
and 8 A.M. 
These leading characteristics may also be recognised in the 
winter months, but the movement during the day is less at that 
season, and the nightly movement is greater ; the turning points 
are also earlier in the morning and later at midday and in the 
night. 
Besides the principal turning points there are other smaller 
inflections, particularly during the night, which may be recog- 
nised in several and sometimes in all the months ; a remarkable 
outward inflection of the curves usually occurs at sunset. 
It does not however appear desirable to found further con- 
clusions or calculations on the data at present possessed ; we 
should rather await the results of subsequent observations, and 
avail ourselves, in respect to these, of the indications afforded 
by the numbers before us. It should be noticed, in particular, 
that it is not sufficient to note the declination at every even hour; 
it appears almost necessary to take every hour, except perhaps 
5, 9, 11 A.M. 
If observations are made at every even hour, and at part only of 
the intermediate hours, the former only can be employed in the 
calculation of the mean declination. The mean of the even 
hours departs a little from the true mean (given by the quadra- 
ture of the daily curve), but never materially, as may be seen by 
the following table of corrections of the means of the even hours 
to the true means. 
Jan. 1. + 0°03 May 1. 0:00 Sept.1. 0:00 
Feb. 1. + 0°01 Junel. 0°00 Oct. 1. 0°00 
March 1. — 0°02 July 1. — 0°02 Noy. 1.+ 0°02 
April 1. — 0°02 Aung.1. 0:01 Dec. 1. + 0°01 
We may hence infer that for obtaining mean values only two- 
hourly observations are quite sufficient. 
