Xll 
General Results of the Makeestoun Observations. 
Mean Westerly Declination and its secular change. — Tlie mean declination for each month, in each year, 
from August 1841 till November 1849, is given in Table I. ; it diminishes generally from month to month ; 
in the mid-summer months, however, it frequently increases. 
4. The mean declination for each year, with the yearly value of secular change and its mean value for each 
four years are given in Table 2 : the mean for 1841 has been deduced in assuming the change from 1841 to 
1842 the same as that obtained from a comparison of the observations for four months of 1841 with the 
observations for the corresponding months of 1842 ; and the mean for 1849 has been found similarly fi'om 
the comparison of the observations in the first eleven months of the years 1848 and 1849. 
5. The mean yearly value of the secular change from the last column of Table 2 = 5''92. Some iiTegularity 
appears in the values of the secular change from year to year, especially in those for the years 1846-1849, com- 
pared with the values for the preceding years ; this marked difference, it is conceived, is not due to instnimental 
error, because no such amount of torsion in the suspension thread of the declination magnet has existed to pro- 
duce it ; and the observations of the bifilar and balance magnetometers indicate a similar variation in the 
value of the secular change for the year 1847-8. In that year, gTeat magnetic convulsions occurred, the effects 
of which seem to have extended into the years 1848-9. 
6. The last column of Table 2 appears to exhibit the variation of the yearly value of secular change ; its in- 
crease as the needle moves farther from its greatest westerly position. Between 1842 and 1847 the secular motion 
from year to year is moderately equable. This is not the case with the motion from month to month, which is 
occasionally x-etrogTade. We are induced to conclude, therefore, either that the secular motion varies from 
month to month, while neai'ly constant from year to year ; or that the secular motion being uniform from 
month to month other motions are superposed : in either case, by reducing the mean positions for the several 
months to one epoch, the residual variations will be more clearly exposed, and it may be determined whether 
they obey any law related to season. Since we are aware that the secular motion for the same place is sometimes 
eastwards and sometimes westwards, it does not appear necessary to form any other hypothesis than that the 
secular change is the excess of the motions in one direction over those in the other, and to determine whether 
the amounts and directions of motion have any relation to season. 
Table 2. — Yearly Means of Magnetic Declination and the Secular Change. 
Secular Change. 
Mean 
Year. 
W. Declination. 
Each 
Tear. 
Mean of 
4 Years. 
1841 
25 33-68 
1842 
25 28-45 
5-23 
1843 
25 22-85 
5-60 
1844 
25 17-06 
5-79 
1845 
25 11-32 
5-74 
5-59 
1846 
25 5-97 
5-35 
5-62 
1847 
24 59-65 
6-32 
5-80 
1848 
24 51-81 
7-84 
6-31 
1849 
\ 
24 45-12 
6-69 
6-30 
7. Annual Period of Magnetic Declination. — In the discussions for 1844 the apparent law of annual varia- 1 
tion has been oflered with some confidence, and tluxt chiefly because of the considerable agreement of four years' j 
observations where the variations were of the smallest order. In the moans for 1843-6, the proportional partsi 
of the yearly secular change being eliminated, the variation of the monthly means is tmder one minute: since 
the variations from month to month are so small, it is e-vddent that, in order to detect any relation to season, 
the greatest care must be taken to avoid all instrumental errors ; for this reason it appears proper to consider 
at iirst the results from those years only (1843-6), during which a sufficient number of daily observations were 
made to give tlie monthly means Avithout any considerable error. The moans for the iirst of these years ^^1843 
an! all'octed to some extent with torsion of tlio suspension tluvad, which broke gradually in June ; on whicb 
account tlie mean of May and July has been substituted for Juno in Table 3. 
