Fuly 15, 1886] 
Peters, are much more in harmony with the figures in the second 
table than the first. 
I may mention, in connection with this part of the subject, 
that the view of the great mobility of the photospnere which 
this hypothesis demands, so soon as we regard metallic promi- 
nences as direct effects of the fall of spot material, is further 
justified by the fact that, if we assume the solar atmosphere, that 
is the part of the sun outside the photosphere, to be about 
500,000 miles high, which I regard as a moderate estimate, the 
real average density of the sun is very nearly equal to one-tenth 
that of water, instead of being slightly greater than that of 
water, as stated in the text-books.} 
We can then only regard the photosphere asa cloudy stratum 
existing in a region of not very high pressure. It is spherical 
because it depends upon equal temperatures. 
_ The second direction in which I have attempted to develop 
the hypothesis has relation to the circulation in the sun’s atmo- 
sphere. I have taken the facts of the solar atmosphere as a 
whole, as they are recorded for us in the various photographs 
taken during eclipses since 1871, and also in drawings made 
before that time, the drawings being read in the light afforded 
by these photographs, 
£ Fic. 2.—Minimum. 
\ 
| | These currents, as they exist in the higher regions of the 
) atmosphere, carry and gather the condensing and condensed 
materials till at last they meet over the equator. 
There is evidence to show that they probably extend as solar 
meteoric masses far beyond the limits of the true atmosphere, 
and form a ring, the section of which widens towards the sun, 
)-and the base of which lies well within the boundary of the 
atmosphere (Fig. 2). 
| _ If we assume such a ring under absolutely stable conditions, 
| there will be no disturbance, no fall of material, therefore there 
| will be no spots, and therefore again there will be no promin- 
€nces. Such was the state of things on the southern surface of 
the ring from December 1877 to April 1879, during which 
period there was not a single spot observed the umbra of which 
) Was over 15-millionths of the sun’s visible hemisphere. 
| Assume ‘a disturbance. This may arise from collisions, and 
these collisions would be most likely to happen among the 
particles where the surface of the ring meets the current from 
| the poles. These particles will fall towards the sun, thereby 
* The density referred to water = 1°444 and to the earth o'255, 
to Newcomb. 
according 
NATORE 
Tracing of Newcom)'s observation of 1878, the brighter portion of corona being hidden bya screen. 
| extension and concentric atmospheres. 
2 
55 
I find that the working hypothesis at once suggests to us that 
the sunspot period is a direct effect of the atmospheric circula 
tion, and that the latitudes at which the spots commence to form 
at the minimum, which they occupy chiefly at the maximum, 
| and at which they die out at the end of one period in one hemi- 
sphere, probably at the moment they commence to form a second 
one in the other (as happened in 1878-79), are a direct result of 
the local heating produced by the fall of matter from above de- 
scending to the photosphere, and perhaps piercing it. The 
results of this piercing are the liberation of heat from below 
and various explosive effects due to increase of volume, which, 
acting along the line of least resistance, give, as areturn current, 
incandescent vapours ascending at a rate which may be taken as 
a maximum of 250 miles a second, a velocity sufficient to carry 
them to very considerable heights. 
The view of the solar circulation at which I have arrived may 
be briefly stated as follows :— 
There are upper outflows from the poles towards the equa- 
torial regions. In these outflows a particle constantly travels, 
s» that its latitude decreases and its height increases, so that the 
true solar atmosphere resembles the flattened globecinsPlateaw’s 
experiment (see photographs, 1878, and Fig. 3). 
Shows the equatorial 
disturbing and arresting the motion of other particles nearer the 
photosphere, and finally they will descend with a crash on to 
the photosphere, from that point where the surface of the ring: 
enters the atmosphere some distance further down. 
The American photographs in 1878 supply us with ample 
evidence that this will be somewhere about lat. 30°, and here 
alone will the first spots be formed for two reasons. 
(1) In the central plane of the ring over the equator, the 
particles will be more numerous ; a rapid descent, therefore, in 
this central plane will be impossible, for the reason that the 
condensed matter has to fall perhaps a million of miles through 
strata of increasing temperature ; there will, therefore, be no 
spots ; and practically speaking, as is known, there are no spots 
at the equator, though there are many small spots without 
umbrz between latitudes 3° and 6° N. and S. 
Above lat. 30°, as a rule, we have no spots, because there is 
no ring, and further the atmosphere is of lower elevation, so that 
there is not sufficient height of fall to give the velocities require 
to bring down the material in the solid form. a @ 
The lower corona, where the corona is high, and it is highest 
over the equator, acts as a shield or buffer ; volatilisation and 
