TERRESTRIAL MAGNETISM AT KEW OBSERVATORY. 
113 
and decidedly the smallest daily range in both H and D. August and June, however, 
had a smaller number of clays of character “ 2.” 
On the average of the 12 months, the D range is about 20 per cent, greater than 
the H range ; but the proportional excess is notably less in June, July, and August 
than in the midwinter months. 
§ 30. Table XVII. represents an attempt to reach comparative results for the 132' 
months of the 11 years. The figures in ordinary type were obtained from the mean 
values of the absolute daily H ranges in individual months. The value for each of the 
11 Januarys was expressed as a percentage of the mean of the 11 January values, and 
the same was done for the other months of the year. This was intended to eliminate 
the annual variation in the amplitude of the daily range. The figures in heavy type 
were derived from the character figures, the value for each month being again expressed 
as a percentage of the arithmetic mean character figure for the 11 months of the 
same name. The data in the two columns headed “ year ” represent arithmetic means 
of the corresponding 12 percentage values in the monthly columns. 
The difficulty of maintaining a uniform standard of magnetic character in different 
years should be remembered, as it introduces greater uncertainty in the character 
data of Table XVII. than in those of any previous table. Consider, for example, 
what would happen if the standard for characters “ 1 ” and “ 2 ” were lower for 1896 
than other years. In January 1896 the character figure was remarkably large, as 
compared to that of the other months of the year. The consequence of the hypothetical 
low standard would be not merely to unduly exalt the character figure for January 1896 
in Table XVII., but also to depress the character figures in all the other Januarys. 
The data in the second last column of Table XVII. put the years in the same 
order as the corresponding data in Table XV. 
The daily range remains very considerable in the quietest times when character 
figures are nearly all “ 0.” Thus the range figures in Table XVII. naturally fluctuate 
within narrower limits than the character figures. The extreme smallness of the 
latter towards the end of 1900 is particularly striking. 
While there are many marked differences in the order in which the two sets of 
figures place the months, a conspicuously high value in the one set of figures is nearly 
always associated with a high value in the other. January, 1896, March, 1892, 
June, 1894, November, 1894, and December, 1892, stand first in both lists for months 
of the same name ; while February, 1892, July, 1894, September, 1894, October, 1895, 
November, 1895, and December, 1895, stand either first or second in both lists. 
If we take a mean between the two sets of figures, November, 1894, was relatively 
the most disturbed, and November, 1900, was relatively the quietest month of all. 
February, 1892, from the same standpoint, only just fell short of November, 1894, 
and was followed after a slight interval by January, 1896, July, 1892, July, 1894, and 
June, 1894. 
§ 31. An attempt was made to utilise the figures of Table XVII. in a similar way 
VOL. CCXII.-A. Q 
