October 18, 1895.] 



SCIENCE. 



511 



Furthermore, the fact that this field ex- 

 hibits three compensating couples, fulfilling 

 the laws of refraction when a permeable 

 shell is placed within an external magnetic 

 field, renders it certain that we have at last 

 secured the basis of the complete solution 

 of the ancient problem of the distribution 

 of the earth's quasi permanent magnetism 

 and its variation in short and long periods. 



In order to distinguish the field of force 

 that was supposed to produce the aurora 

 and the other phenomena above mentioned, 

 it was proposed to call it 'coronal,' or 

 'polar' radiation, in distinction from the 

 sunlight, or equatorial radiation. It is radi- 

 ation of some kind, if there is any trans- 

 mission of energy through the ether from 

 the sun to the earth, and it may be simplj' 

 magnetic, or curved radiations, as opposed 

 to rectilinear, or electro-magnetic radiation, 

 the latter having been practically estab- 

 lished as natural by the work of Maxwell 

 and Hertz. 



It was evident that if a solar-polar mag- 

 netic field existed and extended to the dis- 

 tance of the earth, its presence would be 

 revealed by periodic variations, the period 

 being determined by the synodic i-otation of 

 the sun, and the variations by the impressed 

 energy due to the magnetic output on the 

 several meridians. Also for the mainten- 

 ance of such a normal field, whatever fluc- 

 tuations it might undergo in itself, it was 

 necessary to suppose that the nucleus of the 

 sun is to some extent rigid, or at least non- 

 vaporous. The detection of the synodic 

 period and the approximate form of the curve 

 representing the solar field at the distance of 

 the earth followed, the period being 26.67- 

 928 days, and the curve the one many 

 times published. The period was found 

 from the years J 878 to 1889, these contain- 

 ing the available European modern obser- 

 vations; since that time an application of 

 the same period cari-ied back from the 

 epoch 1887 to the British Colonial Stations, 



1841 to 1848, gives back the same curve, as 

 if begu.n about one-tenth of a day later. 

 Thus my first period is sufficient for a half 

 century's work, and it is plain that a redis- 

 cussion of all the data will enable us to de- 

 termine the rotation of the sun with ex- 

 treme accuracy. Since we recognize the 

 fact that the magnetic curves are a true 

 and delicate register of solar action through 

 at least 800 revolutions, it is clear that few 

 natural phenomena have been so contin- 

 ously recorded as the solar motion and in 

 such detail. The same remark applies to 

 the other physical manifestations of the 

 energy of solar nucleus, if we learn to cor- 

 rectly interpret the changes in the mag- 

 netic curves. This unconscious contribution 

 to solar physics by magneticians, thi'ough 

 more than fifty years, is abundant justifica- 

 tion of the faith in science that has in- 

 spired their work, and a sufficient answer 

 to the cavilling question, eid bono. 



The securing of the solar period was, of 

 course, the foundation of progress in the 

 classification of large masses of hitherto un- 

 workable data. The illustration of its 

 power is contained in the series of results 

 thru.st upon us by using it. At first the ef- 

 foi't was made to detect a similar curve in 

 meteorological and in solar phenomena by 

 simply massing the observations in this 

 period. The results were tantalizing, if not 

 discoui'aging, for, while it was evident that 

 a similar synchronous beat existed in the at- 

 mosi^heric elements, yet the residuals were 

 so small, and the curves exhibiting them so 

 rough, when compared with one another, 

 that it seemed for a long while as if further 

 progress might be impossible. However, 

 by persistent study of successive periods 

 during which no little practical skill was 

 developed in detecting the underlying har- 

 mony in apparently unrelated curves, it was 

 discovered that the normal curve was sub- 

 ject to inversion. That is to say, the curve 

 was workable for a season, say a few weeks, 



