xlvi 



Geneeal Results of the Makerstoun Observations. 



83. The numbers in the last column of Table 39 have been obtained by comparing the mean of the first six 

 months of the year, for which the secular change is sought, with that for the corresponding six months of the 

 following year, and the mean for the last six months with the corresponding months of the preceding year ; the 

 mean of the two is taken for the secular change during the year : only one comparison could be made for 1842 

 and for 1849. 



84. The vertical component has diminished from year to year ; the value of the secular change has also 

 diminished since 1842 ; the greatest diminution occurring in the year 1846, after which year the value of the 

 secular change has not varied greatly : it was least in 1848, and appears as large in 1849 as in 1846. It is 

 probable that the apparent secular change is partially, if not wholly, due to loss of magnetism in the needle, 

 especially in the first four years. (See the Section Magnetic Dip.) 



85. When we deduce the yearly value of the vertical component of magnetic force from the days selected as 

 nearly free from irregular disturbance in 1844 and 1845 (see volume for 1844, p. 384, and Table XXXVII., 

 p. 21, of this volume), and compare these values with those obtained from all the ordinary observations in 

 each year, we find, that the yearly mean of the vertical component deduced from the nearly undisturbed days 



In 1844 is greater than that deduced from all the hourly observations of the year, by 0-000021 

 In 1845 0-000039 



The effect of disturbance in both years was to diminish the mean value of the vertical component on the 

 average by 0-000030. See Nos. 38 and 54. 



86. Annual Period of the Vertical Component. — In the discussion of the observations for 1844 it was found 

 that the result for the annual period differed considerably from that obtained for the previous year ; separate 

 discussions of the observations for the different years shewed so many differences, that it was concluded that 

 the instrument was incapable of exhibiting the law, if any such existed ; this conclusion was quite in accordance 

 with that previously made by those who had examined the instrument with the greatest care. A combination 

 of the results for several years, however, has rendered it probable that the errors of the instrument or irregu- 

 larities in the law have been eliminated to a considerable extent, and that the true law has been obtained ; this 

 will appear from the following discussion. 



Table 40. — Monthly Variations of the Vertical Component of Magnetic Force, free from Regular 



Secular Change. 



Years. 



Jiin. 



Feb. 



March. 



April. 



May. 



June. 



July. 



Aug. 



Sept. 



Oct. 



Nov. 



Dec. 



Prefix. 



0-000 



O'ono 



O-OOll 



0-000 



0-000 



0-000 



0-000 



0-000 



0-000 



o-ooo 



0-000 



0-000 



184.3-6 



+ 086 



-030 



-064 



-054 



-060 



+ 094 



+ 031 



+ 016 



-009 



-054 



+ 013 



+ 028 



1842-7-8-9 



-014 



i + 027 



-071 



-097 



-121 



-113 



+ 028 



-088 



-127 



+ 137 



+ 229 



+ 215 



1842-9 



+ 036 



j-001 



-068 



-075 



-090 



-010 



+ 030 



-036 



-068 



+ 041 



+ 121 



+ 122 



1842-7 



-007 



-046 



-099 



-052 



-040 



+ 065 



+ 011 



-005 



-002 



+ 008 



+ 088 



+ 079 



Table 40 has been formed in the following manner : The monthly means for the different groups of 

 years having been obtained from Table 38, the means for each group were con-ected for secular change in the 

 manner described No. 8, the secular change employed being the mean for the respective years in the last 

 column of Table 39 ; the numbers in Table 40 are the variations about the yearly mean for each con-ected group. 



87. The four years 1843-6 are those best fitted for exhibiting the annual law, the diui-nal series of observa- 

 tions being sufficiently large to give the monthly mean without any considerable error. The result from this 

 group is as follows : The vertical component is a maximum in June and January, and it is a minimum in 

 April and October. The remaining four years, especially the years 1848 and 1849, though but indifferently fitted 

 to exhibit a law liable to so many irregularities, (owing to the fewness of the observations made daily, and to 

 the great magnetic disturbances in the last two years,) yet they exhibit a rough approximation to the same 

 result : for this group, the vertical component is a maximum in November and December, a secondary maxi- 

 mum occurring in July ; and it is a minimum in May and September. The group of six years, 1842-7, includes 

 all the years during which a sufficient number of daily observations were made to give moderate approximations 

 to the monthly means, and this group indicates a law similar to that from the four years 1843-6. Neglecting 

 at present the considerations in favour of the law obtained from the four best years, the chief source of doubt as 

 to the value of the result, is to be found in the irregularity and great variation of the secular change to be elimi- 

 nated. In order to examine the monthly means free from this objection the following Table has been formed. 



