192 8S. P. Langley—Selective Absorption of Solar Energy. 
form as to show from the mean of a year’s observations, the 
percentage of absorption suffered by each ray in the entire 
spectrum, visible and invisible. 
The reader who may desire still fuller details as to the ap- 
paratus, the original observations and their treatment, is referred 
to the forthcoming official publication already mentioned. In 
the later memoir will be found a description of the method used 
for determining the wave-lengths corresponding to measured 
deviations, and the formule for deducing from the prismatic 
spectrum, the distribution of the energy and the extent of the 
spectrum on the normal scale. 
SUMMARY. 
As one result of this present research, the chart of the pris- 
matic spectrum as observed at Allegheny with the bolometer 
is now presented (Plate IIT). The abscissz are proportional to 
deviations and the ordinates to measured energies. The second 
chart now given, (Plate IV), represents the normal spectrum as 
deduced from the prismatic, as it has been thought advisable 
to present it here for the reader's convenience, in advance of a 
description of the means used for making it. The abscisse on 
this are proportional to actually measured wave-lengths, and 
the ordinates to measured energies. In both charts, the area, 
between ordinates cotresponding to like wave-lengths is the 
same, and hence the total areas are the same. Their very 
as to the nature of the new absorption bands, may be give 
hereafter, together with tables (already prepared) of the absorp- 
tive action of the solar atmosphere for each spectral ray. These 
will, it is hoped, give with a satisfactory approximation the 
distribution of the energy, before any absorption whatever; a 
the source, that is, of the energy, in the photosphere itself. 
The extent of the newly observed region may be most clearly 
en by reference to the map of the normal or diffraction spe 
trum. (Plate IV.) Previous maps end at or near wave-length 
1*-2. Beyond this point (with the exception of the single 
band near wave-length 1-4) every line, and every ordinate 
representing heat, is believed to be new. The extent of the 
region here newly mapped, is then considerably larger, on the 
normal scale, than the whole of that (both visible and invisible) 
e observe that the prismatic spectrum is enormously 
expanded at the violet end. To carry this on the prismatic 
