FROM ORDINARY DAYS OF THE ELEVEN YEARS 1890 TO 1900. 
275 
oscillations, the V curve shows practically no trace of disturbance, or merely a slight 
exaggeration of the afternoon maximum. At times, however, V exhibits disturbances 
rivalling those in D and H. If we confine ourselves to days when the range 
exceeded 250y, there were 24 occurrences in Y as compared with 28 in D, and on 
4 occasions the V range exceeded 500y, while the largest D range observed was 
457y. .During some of these very large disturbances the V trace—which has usually 
a gently rounded contour—-showed large rapid oscillations just like the D and H 
traces. 
The great majority of the largest storms occurred in 1892 or 1894. In the group 
of years of sunspot minimum a range as high as 250y appeared only once in V, and 
on no single occasion in D or H. 
There is a certain resemblance between the frequency curves for winter and for the 
sunspot minimum year, between those for equinox and for the average year, and 
again between those for summer and for the sunspot maximum year. The curves for 
the sunspot maximum year show a less smooth rise to the maximum ordinate than 
the others. 
§ 30. Fig. 14 presents the range distribution data from a different point of view. 
Unit abscissa represents the arithmetic mean range for the element and season con¬ 
sidered. Thus, in the case of the year from the 11 years, it represents 72‘2y in D, 
60‘7y in H, and 32‘9y in V. This eliminates the effect of mere size, and so helps to 
bring out the degree of resemblance between the laws of distribution followed in 
different elements, or in the same element at different seasons. In the case of the 
* 
whole year, the range of greatest frequency occurs very nearly at the same place— 
i.e., bears very nearly the same ratio to the mean range—in the H and V curves ; 
in D it answers to a decidedly larger abscissa. 
In the case of H, the range of greatest frequency answers to a higher and higher 
fraction of the mean range as we pass from winter to equinox, and from equinox to 
summer. There is not much difference between the curves in fig. 14 answering to 
sunspot maximum and minimum, though the mean ranges are respectively 75'8y and 
43 *9y. 
All the frequency curves have the general aspect of the ordinary diagram of 
intensity of visible and invisible radiation from a solid at high temperature. This, 
however, may mean no more than that the number of classes in which the ranges 
were grouped was too limited to show the existence of bands of reduced frequency or 
of no frequency. The existence of such bands, at least towards the side of the very 
high ranges, is by no means improbable, especially if storms should prove to have 
more than one origin. 
For the careful measurement of the curves and the calculation of the hourly means 
I am indebted to various members, past and present, of the Kew Observatory staff. 
Amongst these I would particularly mention Messrs. G. Badderly and C. Cooper, 
2 p 
VOL. CCXVI.-A. 
