101 



Prof. B. Stewart on the Daily 



[Mar. 22, 



Table I. (continued). 



Date. 



1858-61. 



1862-65. 



1866-9. 



1870-3. 



Mean. 



Sept. (1) 



•609 



•540 



•514 



•646 



•577 



» (2) 



•581 



•523 



•494 



•621 



•554 



(3) 



•559 



•493 



•481 



•595 



•532 



Oct. (0) 



•537 



•483 



•458 



•573 



•513 



» (1) 



•522 



•464 



•445 



•552 



•496 



» (2) 



•504 



•448 



•437 



•522 



•478 



» (3) 



•486 



•445 



•418 



•503 



•463 



Nov. (0) 



•465 



•427 



•408 



•480 



•445 



» (1) 



•420 



•402 



•389 



•462 



•418 



» (2) 



*d89 



•37b 



•obi 



. ,| OA 



•430 



•389 



„ (3) 



•363 



•354 



•333 



•390 



•360 



Dec. (0) 



•341 



•337 



•309 



•371 



•340 



„ (1) 



•341 



•321 



•279 



•345 



•322 



m (2) 



•323 



•311 



•259 



•339 



•308 



„ (3) 



•325 



•305 



•254 



•349 



•308 



3. It will be seen from Table I. that while there is a maximum of 

 declination-range in June about the time of the summer solstice, there are 

 also maxima in April and August, and that a behaviour of this kind is 

 indicated in each four years' observations. Comparing this result with that 

 embodying the annual variation of temperature-range at Kew (Proc. Roy. 

 Soc. 1877, vol. xxv. p. 578), it will be seen that the latter variation has only 

 one maximum in July. Perhaps there is a reference to the equinoxes as well 

 as to the solstices in the annual variation of the declination-range. A 

 comparison of the two is exhibited in Pigs. IX., X., p. 120 (Pig. IX. giving 

 declination- and Pig. X. temperature-ranges). 



B. Variations of Long Period. 



4. It is well known that the range of the magnetic declination has a 

 long-period variation, apparently connected with the physical state of the 

 sun's surface. In order to investigate the nature and closeness of this 

 connexion the following plan has been adopted ; — Let us assume as the 

 most probable hypothesis that the cause which exalts or depresses the 

 mean annual declination-range exalts or depresses also in a similar 

 manner the variations of this from one month to another. This is what 

 would take place if we could imagine the effect to be produced by some 

 influence emanating from the sun, which acted more powerfully on some 

 years than on others, while the variations of this effect due to the sun's 

 position in the ecliptic were also altered in the same proportion. On the 

 whole this is borne out by Table I. Constructing, now, a Table for each 

 year, and for 48 points in each year, and reckoning the mean of the 16 

 years' ranges for each of these points (as exhibited in the last column of 

 Table I.) equal to 1000, we find in Table II. a series of values exhibiting 

 the proportion between the observed range for any point of any one year, 



