462 
MR. J. EVERSHED ON THE SOLAR ECLIPSE OF 1900, MAY 28. 
pointed out,* diffraction images more or less enlarged by jihotographic diffusion, and 
they appear to be as well adapted for bisection and wave-length determination as are 
the lines given by a slit-spectroscope. 
Exception may, perhaps, be taken in the case of the helium lines and the somewhat 
remarkable line at 4685'7. These do not increase in intensity towards the 
photosphere, and it is possible that they are very weak and even absent from the 
flash layer. A bisectio]i of these arcs may, therefore, represent a point in the 
chromosphere higher than the flash layer. 
No allowance has, however, been made for this, yet the values obtained indicate 
only a small displacement towards the red end, averaging T6 tenth-metre for the 
three lines 4713, 4471, and 4026, when compared with the jDiincipal components of the 
double lines as determined by Hunge and Paschen, and which they are assumed to 
represent. 
But it is noticeable that in these spectra the helium lines become broad and faint 
in the flash layer, although narrow strong lines outside; the measures are, therefore, 
somewhat uncertain, and it is possible that they may be partly affected by the less 
refrangible components of the double lines. In any case they serve to show what a 
small correction is needed, even for the lines of a sulDstance like helium, which is 
characteristic of the upper chromospliere rather than the flash layer. 
Estimates of Intensity. 
On account of the great range of intensity between the weakest and the strongest 
lines, a scale was adopted ranging from 0 to 100. Tins is practically equivalent to 
adopting two orders of intensity, 0 to 10 representing the weak lines, 10 to 100 the 
strong lines, the latter progressing by fives. 
The intensities of all the lines, with the excejjtion of those of hydrogen and H and 
K, were estimated while making the measures, two indejjendent estimates of each line 
Ijeing obtained from the two sets of measures of each spectrum. The mean of the 
two estimates is set down for each spectrum in Table I. 
The hydrogen and calcium lines were estimated separately. From to Hp they 
form a nearly uniformly diminishing series, giving a convenient scale of reference with 
which to compare the strong lines of the flash spectrum. 
At the ends of the spectra, where the density of the image falls oft considerably, the 
estimates are of course very rough and uncertain, and throughout the middle 
portions of the spectra the intensities are not perhaps strictly comjjarable, except over 
a limited range of wave-length. 
* ‘Phil. Trans.,’ A, vol. 197, p. 394. 
