350 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vor Vis, 
It seems to be cone-shaped, driven outward from the sun, not drawn like 
a beam by the earth. And the intensity of the action may vary in 
different parts of the conical pencil. 
By taking the average interval between magnetic storm repeats, the 
rotation of the sun ought to be accurately ascertainable, even better than 
by observing sun-spots, which are constantly changing their form and 
often their position, being so controlled by movements in the solar 
atmosphere that they go round in different times in different latitudes, 
By applying the approximate measure, 274 days, to a long series, to aid 
in the detection and identification of repeats, I found the exact synodical 
rotation to be 27.24575 days. My calculation is in very close accord 
with the best result of sun-spot observers, who differ considerably among 
themselves. The sidereal rotation period corresponding to the above 
figures is 25.354 days. Mr. Carrington gives 25.38 days. 
Solar protuberances and prominences are disturbances in the solar 
atmosphere, and by no means so deeply-seated in their origin as spots 
In charting them upon a map of the sun, which by means of meridians 
dividing the equator into 271%4 parts one can do, they arrange themselves 
in belts. The record of the Italian Spectroscopical Society, edited by 
Tacchini, is that which I have followed. True, it is an exaggeration to 
mark down the.prominences of a whole year, as if they were all in play 
at one time, but it is the best way that occurs to me to get a synoptic 
view. The resemblance to the belts of Jupiter is striking, and for that 
matter, the earth would not present a very different appearance in this 
respect, if looked at from the moon, for there are fairly regular zones of 
moisture and clouds on both its northern and southern hemispheres, as 
well as regions which clouds seldom obscure. If then the belts of Jupi- 
ter, Saturn and Earth are atmospheric, we may reasonably conclude that 
those on the sun are of similar nature. The magnetic curves supply 
another reason; they are not affected by prominences, so far as I can 
discover, though as prominences are only noted at the edge of 
the sun, and not after they have been brought on the disc, the infer- 
ence as to their want of effect when central is open to objection. 
The most important prominence of 1897 was that seen on the western 
limb, at latitude + 8° 2, on November 23rd. The locality whence it was 
ejected had therefore been central on the 16th. There was no magnetic 
dip on either date, nor was there any spot near the place where it was 
seen. 
Prominence belts change their latitude from year to year. I have 
, some reason to believe that the foci of the disturbances which produce 
