NIGHT-SKY RADIATIONS FROM THE UPPER ATMOSPHERE 
By E. O. HULBURT 
Naval Research Laboratory, Washington, D. C. 
Introduction 
In places remote from artificial illumination at night, 
with no moon, the luminosity of the sky is due to several 
sources of light, all of which are at a considerable dis- 
tance. The sources are (1) radiations from the gases of 
the upper atmosphere, (2) polar aurorae, (3) zodiacal 
light, (4) comets and possibly scattered sunlight in 
interplanetary space (if there is something there to 
scatter the sunlight), and (5) stars and nebulae in 
interstellar space. 
If we omit polar aurorae from consideration, the 
radiations from the high atmospheric gases are the 
strongest source of night illumination, being four or 
five times as intense as all of the other sources combined. 
In more detail, the intensity of the mght-sky light 
averaged over the sky is divided as follows: For the 
photographic spectrum [5] region from 3500 to 4500 A 
about 4% is due to the stars, 46 to the high atmos- 
pheric emissions, and possibly: 1¢ to emissions from 
interplanetary space (the amount attributed to sources 
in interplanetary space has decreased as the measure- 
ments have improved; it is now down to about 1 of 
the whole and may eventually become a few hundredths 
or even less); for the visible spectrum [17] seen with 
the dark-adapted eye, about 14 is due to the stars and 
4¢ due to the high atmosphere; for the entire spectrum 
from about 3000 to 10440 A, because of its strong 
infrared nitrogen emissions, the high atmosphere con- 
tributes more than 9/9 of the nocturnal radiation and 
the stars less than 149. These fractions are average 
values; for areas in the sky such as the Milky Way 
where there are many stars the fractions of the sky 
luminosity due to the stars are greater than the average 
values, and for areas in the sky where there are few 
stars the fractions are less. 
Spectrum 
The night-sky light is feeble, and even with spectro- 
graphs of the highest light-gathering power, relatively 
low dispersion and long exposures are necessary to 
obtain spectra. We need to refer only to recent work and 
mention that the best spectra at present available were 
those obtained by Cabannes and Dufay [8] in 1933-34 
in France; by Elvey, Swings, and Linke [12] in 1939 at 
the McDonald Observatory, University of Texas; by 
Barbier [4] in 1942-44 at the Observatory of Haute 
Provence, France; and by Meinel [19] in 1948 at the 
Lick Observatory, University of California. 
The dispersion of the spectrograph of Barbier was 
150 and 630 A mm at 3200 and 5000 A, and the length 
of the spectrum from 3200 to 5000 A was about 7 mm; 
exposures from 100 to 200 hours were used. The Mc- 
Donald spectrograph had slightly higher dispersion and 
much greater light-gathering power; the dispersion was 
341 
115 and 500 A mm at 3200 and 5000 A, the length of 
the spectrum from 3200 to 5000 A was about 10 mm, 
and exposures of 9 hours were used. These were prism 
spectrographs and therefore had low dispersion for the 
longer wave lengths. To obtain better data for wave 
lengths longer than 6000 A, Meinel built a spectrograph 
with an f/1 camera and a 7500 line per inch transmission 
grating with a dispersion of 250 A mm im the first 
order; exposures of 30 hours were used. Night-sky 
spectra are shown in Fig. 1, in which A is a spectrum 
photographed by Barbier [18], B by Elvey, Swings, and 
Linke [12], and C by Meinel [19]. In Fig. 1C the narrow 
Vegard-Kaplan system 
Sno T HONDA So nm gown o> A 
A 
[iE 
a i Herzberg system 
of +. __41_. Unidentified | 
re) 
Be q 
P | 
{ 
B 
a a Bc 
‘ Pd 01 § 1 1°Q 
uel 8 
Fie. 1.—Spectra of the night sky, A by Barbier [18] 
(reproduced by permission of The University of Chicago Press) ; 
B by Elvey, Swings, and Linke [12]; and C by 
Meinel [19]. 
upper strip was the original spectrum from which the 
lower strip was obtained by spreading; Meinel re- 
marked that many of the features of the lower strip 
were spurious and were caused by grains of the photo- 
graphic plate. 
The night-sky spectrum is very complex and is com- 
posed of many lines and bands, for the most part in- 
