Conservation of Solar Energy. 63 
and in the solar photosphere as in our atmosphere, and for 
the same reason no more hydrogen in those regions than we 
can detect in the atmosphere. 
While admitting the uniform distribution of C0 2 in our 
atmosphere, which is subject to powerful circulating currents, 
I cannot agree with Mr. Cook in ascribing to gaseous diffu- 
sion any considerable influence upon the constitution of a 
great solar inflow and outflow current. Mr. Cook points 
out very properly that the density of some of the metallic 
vapours known to exist in the sun is either inferior to, or 
does not materially exceed that of, carbon dioxide; and these, 
he concludes, would mix rapidly by diffusive action with the 
photospheric current, doing away with its distinctive character; 
regarding stellar material, he says, towards the end of his article, 
"any difference in composition would be rapidly removed 
by the action of gaseous diffusion." Mr. Cook here appears 
to fall into a common error of interpreting Dalton's expression, 
that " one gas diffuses into the space occupied by another as 
though the latter had no existence," into an action comparable 
to the rush of a gas into a vacuum; whereas in reality the rate 
of the diffusion of gases at equal pressures does not exceed a 
few feet per hour, and that of fluids not a few feet per annum, 
as proved by Sir William Thomson's experiments now going 
on at Glasgow University. The effect of such slow action 
must be inappreciable upon gaseous currents depending upon 
solar rotation, amounting to 23,760,000 feet per second, or 
4500 miles an hour. The metallic vapours in the photosphere 
must therefore be attributable chiefly to mechanical intermix- 
ture, which no doubt is considerable, not indeed on the polar 
surfaces, where the tangential motion is not great, nor at the 
equator, where the photosphere has acquired the solar rotation, 
but in the intermediate zones of great differential velocity, 
rendered visible to us by the occurrence of sun-spots, the 
result, as is now largely admitted, of cyclonic action. The 
metallic vapours thus introduced into the photospheric current 
will be burnt and projected outwards into space, not indeed 
in the vaporous condition as supposed by Mr. Cook, but as a 
metallic dust (several thousand times denser than the sur- 
rounding medium), the greater portion of which will pro- 
bably soon return to the sun by virtue of solar gravitation, 
there to be dissociated on reaching the great gaseous metallic 
sea underlying the photosphere. 
It is of course important to consider to what extent the 
density of this metallic sea is likely to exceed that of the pho- 
tosphere. Assuming the latter to contain, besides the light 
combustible gases, a large proportion of oxygen and nitrogen, 
