TEERESTRIAL MAGNETISM AT KEW OBSERVATORY. 
107 
iucoiispiciiousuess of the crest in column n + 4 of Tables Y. and YII. in the case of 
the group of years 1891, 1895, 1896, and the great development of the crest occurring 
14 or 15 days earlier. There is, it is true, in Table XI. and XII. some indication of a 
period of about 14 days for the years 1891, 1895, and 1896, which is not apparent in 
the other groups of years, and this would help to explain an increased prominence 
in the crest about column n—W in Tables V. and YIL, if the crest in column n + 4 
had its usual prominence. But it cannot by itself account for the earlier and 
presumably secondary pulse being more prominent than the later. 
Table XIII. aims at throwing further light on the phenomena of Tables Y. and 
YII. A given mean H range, or a given mean character figure, may arise in many 
different ways. Of two collections of equal numbers of days which have the same 
mean range or character figure, one will contain a larger number of days of character 
“ 2 ” than the other. It appeared desirable to ascertain whether increase in a mean 
character figure in Table YII. arose from similar increases in the number of days of 
character “ 1” and “ 2,” or whether it denoted a special development of highly 
disturbed days. 
Each mean value in Table YII. for the 11-year period was derived from 660 days. 
Table XIII. shows how many of these were of character “ 2,” and how many were of 
character “ 1.” The days of character “ 2 ” and “ 1 ” combined give the total of 
disturbed days, and the difference between this total and 660 gives the number of 
days of character “ 0,” i.e., quiet days. The last column in Table XIII. gives the mean 
of the corresponding entries in columns n—15 to ri+15, showing that on the average 
out of the 660 days in each column 82 were of character “ 2,” 300 of character “ 1,” 
and 278 of character “0.” Figures which exceed the means in the last column are in 
heavy type. 
Table XIII. makes it clear that the secondary pulse in Table YII., with crest about 
column n—11, is due almost entirely to an excess of days of character “ 2,” while the 
primary pulse, with crest about column ^^ + 4, is mainly due to an excess of days of 
character “ 1.” The number of days of character “2” in columns n—12, n—ll, and 
10 notably exceeds that in any other column. 
Quiet days are markedly in excess of the mean from 7 to 1 day previous to the day 
of largest spot area, and they are even in slight excess on the representative day 
itself. They are also decidedly in excess in columns r;,—15 to ri—13, though their 
excess in column n—13 is neutralised by the co-existing excess in the number of days 
of character “2.” 
On the whole, it may he said that the primary pulse is due rather to the absence of 
quiet conditions than to the presence of large disturbance. 
§ 27. Table XIY. serves the same purpose relative to Table XI. that Table XIII. 
served relative to Table YII. It proceeds, however, on a slightly different plan, and 
gives some data for the 3-year periods in addition to those for the whole 11 years. The 
first four lines give the total number of days of character “ 2.” Consider, for example, 
p 2 
