388 
:vlR. H. H. TURNS ll ON THE 
“ (2) The Sun’s atmosphere is not composed of strata which thin out, all substances 
being represented at the bottom; but of true strata, like the skin of an onion, each 
ditferent in composition from the one either above or below. 
“ (3) In the lower strata we have, not elementary substances of a high atomic 
weight, b'Ut those constituents of all the elementary bodies which can resist the greater 
heat of these regions. 
“ It was stated in (6): While discussing the conditions of observation, that whether 
we were dealing with strata of substances extending down to the Sun, or lunited to 
certain heights, the spectral lines would always appear to rest on the Solar spectrum, 
and that the phenomena would in the main be the same. This, however, is true in 
the main only; there must be a difference, and this supplies us with a test between 
the rival hypotheses of the greatest stringency. 
“ For take three concentric envelopes. A, B, C, so that only A rests on the photo¬ 
sphere, B rests on A, and C on B. The stratum B, being further removed from the 
photosphere than the stratum A, will be cooler, its lines will be dimmer, and the lines 
of C will be dimmer than the lines of B, and so on. So, if we could really observe the 
strata, the longer a line is, i.e., the greater the height at which the stratum which 
gives rise to it lies, the dimmer the line will he. 
“ Now, our best chance of making such an observation as this is during a total 
eclipse. We do not see the lines ordinarily, in consequence of the illumination of our 
air. As during an ecli^^se, before totality, the intensity of this illumination is rapidly 
diminishing, the lines first visible should be short and bright, and should remain 
short, while the new lines which become visible as the darkness increases should be of 
gradually increasing length. 
“ Further, the sliort lines which first appear should be lines seen in prominences 
and not in spots, and relatively brighter in the spark than in the arc, while the longer 
lines added should be lines affected in spots and not in prominences.” 
The manner in which these expectations were realised is shown at once by the 
subjoined diagram from the same paper (fig. 1). 
For convenience of reference I have used small letters to denote the lines shown in 
the lowest line of the diagram, beginning v/ith a on the left. Thus— 
two lines at 4870-71 
are called a and h. 
„ 4890 
,, ,, c and d. 
three ,, 
„ 4918-23 
„ „ e,/andg. 
two ,, 
„ 4932-33 
,, „ h and i. 
one „ 
„ 4956 
>) 
5 5 55 
„ 4970 
In this nomenclatui'e, then, lines g and I are seen by Tacchtni in prominences. 
