XVI 
Geneeal Results of the Makeestoun Obseevations. 
extent ; the smaller disturbances not having been observed with the same completeness in that year. The 
mean diurnal ranges for each month from these two Tables are as follow : — 
Jan. Feb. J[arch. April. May. June. July. Aug. Sept. Oct. Ifov. Dec. i Year. 
1844, ll'-63 13'-63 19'-36 19'-10 14'-83 12'-60 13'-36 16'-58 17'-74 19'-26 19'-66 12'-95 15'-89 
1845, 17'-81 15'-31 16'-52 17'-01 14'-67 13'-82 13'-72 17'-79 18'-20 14'-48 12'-64 15'-01 [ 15'-58 
Mean, 14'-72 14'-47 17'-94 18'-05 14'-75 13'-21 13'-54 17'-18 17'-97 16'-87 16'-1.5 13'-98 | 15'-74 
These means give the same law of variation as that already found fi'om Table 5, but the values are con- 
siderably higher. From the means for both years we may conclude, that the mean angle, including the diurnal 
oscillations in years of moderate disturbance at Makerstoun, is about 18' at the equinoxes, about 14' at the 
solstices, and about 16' for the whole year. 
Annual Variation of the Ranges of the Monthly Mean Diurnal Variation. — We have considered above 
the annual variation of the mean ranges for each day, we now give in Table 6 the ranges of the mean diurnal 
variation for each month. The first four lines contain the diurnal ranges of the means of all the regular daily 
observations made in each month ; and, as in Table 5, only 1844 and 1845 are comparable with each other: 
the last line contains the range of the hourly means for each month, as deduced from the observations for the 
4 years given in Table 12. 
13. From the last line of Table 6 it appears that when a sufficient number of observations is employed, the 
range of the mean diurnal variation is neai'ly constant for the six months, April to September, being on the 
whole rather larger for the first three than for the last three of that half-year : the mean range for the whole 
six months is about ll'-O. The range is nearly of equal value for pairs of the remaining six months, namely, 
for March and October ( = 9' -2) ; for February and November ( = 7'" 3) ; and for January and December ( = 5'-9l 
Table 6. — Rano-es of the Mean Diurnal Vai'iation of Mao'netic Declination. 
Year. 
Jan. 
Feb. 
March. 
April. 
May. 
June. 
July. 
Aug. 
Sept. 
Oct. 
Nov. 
Dec. 
Mean. 
1843 
1844 
1845 
1846 
4- 92 
5- 26 
6- 95 
6-54 
8-35 
6- 36 
7- 31 
6-05 
7-93 
9-94 
9-92 
10-56 
10-52 
10-20 
13-08 
12-58 
9-93 
8-96 
12-42 
12-79 
11-14 
11- 05 
12- 52 
11-41 
9-95 
10-06 
10- 86 
11- 94 
10-83 
10-31 
12-67 
10-62 
10-75 
9-95 
10- 53 
11- 99 
7-90 
10-94 
9-42 
9-24 
4-63 
9-28 
7- 50 
8- 67 
5-73 
5- 96 
6- 43 
5-77 
7-54 
7- 67 
8- 40 
8-52 
Mean|^ 
of all j 
5-85 
711 
9-28 
11-29 
11-02 
11-61 
10-60 
10-83 
10-45 
9-50 
7-55 
6-02 
8-03 
14. When we examine the range for each month of the mean diui-nal variation as deduced from any single 
year's observations, we observe a similar result to that obtained from the means of the diurnal ranoes (Table 5) ; 
namely, that the range is, on the whole, greater for the mouths near the equinoxes than for those near the 
summer solstice ; but as we combine a lai-ger number of observations the diiference gradually ilisappears, till 
(as we see in the mean of four years' observations, Table 6) the only difl'erence appears in the slightly greater | 
range for the months immediately before the summer solstice than for those immediately after it. The difier- I 
once betwixt tlic two results is, therefore, in all probability, due to irregular causes, which shift the epochs of J 
the extreme positions of the magnet. That this is the case will be rendered nearly evident by the ranges of the ! 
diurnal variation as deduced from the 17 days' observations with least irregular disturbance (see 1844, i 
]). 339, and p. 5 of this volume); they are as follow, ; 
.Ian. Feb. March. April. May. June. -fuly. Aug. Sept. Oct. Xov. I'ei. 
3'-43 4'-54 7'-81 ll'-55 9'-66 ll'-41 ll' ll ll'-lS 9'-64 7'-70 5'-31 4'-ll J 
These, with the exception of the range for May, indicate generally the constancy of the regular dimnial | 
range in the montlis from April till August. The exception of ^lay is easily explained by tlie fewness of the ' 
observations, and the method by wliich the observations were selected (sec volinno for 1844. p. 339). 
15. Since, t hen, the means of the diunial ranges differ from the diurnal ranges of the means, chiefly because of 
irregular disturbing causes which shift the epochs of the extremes, the differences of the results will give some 
measure of these disturbances. Taking the dilFerences betwixt the last line of Table 5, and the last of Table 6. 
we have, 
Jan. Fob. March. April. May. Juno. July. Aug. Sept. Oct. Nov. Dec. 
3'-58 3'-53 4'-27 3'-87 2'-29 1-36 2'-51 3' 50 4'-94 3'-66 3'-21 3'-46 
