TEERESTRIAL MAGNETISM AT KEW OBSERVATORY. 
91 
for some years at de Bilt in presenting the results obtained by international co-operation 
for the “ magnetic character ” of individual days. 
Each co-operating station assigns a “ character ” figure “ 0,” “ 1,” or “ 2 ” to each 
day, according as it is quiet, moderately disturbed, or highly disturbed. The character 
figures assigned by the co-operating stations are summed at de Bilt, as if they were 
ordinary numerical quantities, and the final order assigned to the days in the scale of 
disturbance is based on these sums (or their arithmetic means). If, for instance, 
30 stations send in results, and all assign character “ 2 ” to one particular day, while 
half the stations assign “ 2,” and half assign “ 1 ” to a second day, the total character 
sums obtained at de Bilt for the two days would be respectively 60 and 45, giving as 
the mean estimates 2 and 1'5. 
The principles followed in assigning character figures at different stations vary, and 
the fact that the scale of values is so narrow necessitates grouping together at any 
one station days which differ widely in disturbance. It is also very difficult to 
maintain even a roughly uniform standard throughout a series of years. One naturally 
wishes to discriminate between the days of each year, and if one adopts a standard 
which gives fairly similar numbers of days of characters “ 0,” “ 1,” and “ 2 ” in a highly 
disturbed year, one has in a quiet year hardly any “ 2’s,” and a wholly extravagant 
number of “ O’s.” The natural consequence is a tendency to lower the standard for a 
“ 1 ” or a “ 2 ” in a quiet year, and to raise it in a disturbed year. This renders 
character figures a somewhat uncertain basis for the comparison of one year or one 
group of years with another, but it militates only slightly against their use when 
comparing days of the same month, or even days of different months of the same year, 
unless the year is much more disturbed in some months than others. 
The international scheme came into operation in 1906, and since that time character 
figures at Kew have always been assigned by myself, so that I have had considerable 
practice. The choice at Kew has been based not so much on the absolute size of the 
changes shown as on the more or less oscillatory nature of the curves. 
§ 15. The second line of attack differs from the first only in substituting the character 
figures of individual days for their H ranges. As the period to be considered preceded 
the introduction of the international scheme, it was first necessary to assign character 
figures to the days of the 11 years. This was not so formidable a task as might 
appear at first sight. A single glance at the curves usually enables one to assign the 
character figure to about half the days of a month. My own practice is to consider 
the D and H curves of the month separately in the first instance, assigning to each 
character figures 0, 0, 0 ; 1^, 1, 1 ; 2, 2, and 2. By “ 0 ” is meant a very quiet curve, 
by “ 0 ” a curve one inclines to assign “ 0 ” to but hesitates, regarding the assignment 
of “ 1 ” as at least a possibility, and so on. At Kew V (vertical force) curves need 
hardly be considered, as the element is so much less disturbed than D and H. • If the 
D and H curves for the same day both get a “ 0 ” or a “ 0,” then usually character 
0 ” is assigned without further enquiry. If both get 
N 2 
0,” or one a “ 0 ” and the 
