242 
DR. C. CHREE: MAGNETIC DECLINATION 
however, the mean sunspot areas from all days were not known for individual months, 
absolute sunspot areas are given in the several columns. The corresponding magnetic 
ranges are shown. 
Table XXVI. shows that contradictions exist between different months, just as 
between different years. In several months the left-hand side of the table (i.e. the 
half dealing with the days of largest range) supports the view that magnetic 
disturbance is even largely dependent on the sunspot area three or four days 
previously. February and July are outstanding examples, the increase in sunspot 
area from day n + 1 to day n — 4 being most marked. Again, in the majority of 
months— e.g. February, March, June, July, August, September and December—the 
areas in the several columns of the left-hand side exceed those in any of the columns 
of the right-hand side. But diametrically opposite phenomena appear in more than 
one month. In May and November, taking corresponding columns from the two 
halves of the table, the sunspot area on the right-hand side is the larger from day n — l 
to day n — 4. In January and October the smallest sunspot area on the right-hand 
side is larger—and notably larger—than the largest area on the left-hand side. 
A seasonal change in the amplitude of a phenomenon causes no surprise, but a seasonal 
change which absolutely alters the sign of a phenomenon, so that January is opposed 
to December, and October to September, is, I think, without precedent. 
As regards the right-hand side of Table XXVI., we know already, through 
Table XXV., that the means from the months combined are all below the mean 
sunspot area for the 11 years. This is almost the only feature favourable to an 
association of sunspots and magnetic ranges of the kind suggested by the mean values 
on the left-hand side. In April, indeed, the sunspot area shows a progressive though 
small decline from day n- f 1 to day n— 4, but May shows the opposite phenomenon, 
and so do June, July, and September, from day n to day n — 4. 
§ 31. An independent investigation was made on the lines of § 27, taking the ten 
days of largest and the ten days of least sunspot area in each month as the contrasted 
groups, and comparing the absolute magnetic ranges on these days and the three 
following days. The results obtained were generally similar to those in § 27, but the 
apparent association of sunspot frequency and magnetic range in the final means from 
the 11 years was only about half that shown by Table XXIII. 
An investigation was also made on the lines of § 27, but employing Greenwich 
projected areas of faculag for those of sunspots. This was limited to a single year, 
1892, the results obtained being at least as contradictory and as unfavourable to any 
intimate relationship of cause and effect as the corresponding results from sunspot 
area. 
Discussion of Remarkable Special Cases. 
§ 32. During the investigation into sunspot areas, some individual instances were 
noticed of a highly suggestive character. In August, 1890, the largest range occurred 
