Some Investigations of the Light of the Night Sky
Utsumi, M.
Abstract
Ⅰ. Uber den lagnwelligen Teil des sichtbaren Spektrums des Nachthimmellichtes.
Mit einem Lichtstarken Spektrograhen (F:1.9, f=5 cm, 1-Prisma)、wurden die Untersuchungen des Spektrums des Nachthimmelllichtes von Herbst 1937 bis Fruhling 1941 in Kakioka ausgefurhrt.
Auf diesen Platten aber bei der geringen Disprsion wurden nur die bekannte Linien mit den Wellenlangen λλ6300，λλ5893 und λλ5577 entgedeckt.
In dieser Arbeit werde uber ein Resultat der Beobachtung des langwelligen Teiles des Spektrums des Nachthimmellichtes mit eiuem schwachlichen Spektrogtaphen(F:4.5, f=25 cm，2- Prismen)
berichtet.
Von der im September und November 1940 erhaltenen Aufnahme mit Expositionszeiten von 147 Stunden，ist es nicht wert zu nachforschen.
Auf einer Platte mit Expositionszeiten vou 295 Stunden wahrend der Neumond und Viertel-stellungen des Mondes in den Monaten vom November 1940 bis Februar 1941 waren viel Linien und
Banden aufgenommen.
Es war etwa 30 Wellenlangen angegeben，die meisten mit Linien von Sauerstoff und Natrium, und Banden vou Stickstoff zu identifizieren waren.
λλ6364, 6300 und 5577 sind die bekannte metastabilen Linien des OI，und λ 5893 ist die Natrium D-Linien.
Fur die weitere Diskussion kommen also zunachst diese starkeren Banden λλ6580-6490，λλ6440-6260，λλ6170-5980，λλ5940-5810，λλ5780-5710 und λλ5710-5600 in Frage.
Sie liegen als die Fraunhoferschen Linien C.α ，D ，δ und Chappuis.
Es wird nun behauptet，dass die tellurischen Linien des Fraunhoferschen Spektrums mit den im Nachthimmelleuchten beobachten Banden identisch sind.
Aber es nicht ursprunglich Nachtimmelleuchtens bedeuten，wahrscheinlich werden sie unter den Blitzstrahlen gehoren.
Ⅱ. On the Altitude of the Luminescent Layer
As the light of the night sky has no such definite form as the polar aurora for which can be used the trigonometrical metod，it is not so easy to determine the altitude of tbe luminescent layer.
Cabannes and Dufay made an attempt to determine the height of the luminescent layer of the nightsky from the ratio of tbe light intensity near the horizon to that near the zenith. The ratio is 1.5， from which value they conclude that the layer must be heigher than 100 km, probably around 200 to 300 km.
Bernard has determined the height of the luminescent layer of sodium，radiation from which he assumed to occur under the action of the sun-light，to be about 60 km.
Some years ago，this author inferred，from the same assumption as Bernard, the altitude of the luminous layer of λ 6300 (O・+3P2-Do) must be heiher than 280 km.
But all these methods of determination of the altitude is not so rigorous.
The purpose of this investigation was to infer the most probable altitudes of the luminescent layer of OI radiations. On the radiations:λ6300 and λ5577，the author has determined the intensity ratios of the light of the zenith distance 58°to that of the genith by rigorous spectrophotometrical method.
The mean ratios of 15 measurements about the radiation λ6300 and λ5577 are respectively 1.407 and 1.383. From these values he determined the altitude by the formula，
Iz/Io =(Tz secα+I'z/Bo)/(T'o+I'o/Bo)
and
sinα＝（R/（R+h））sinz
to be 350 km for λ6300 and 340 km for λ5577.
These altitudes are coincident with that of F２-ionospbere.
He thought that the lumincscent layer of the nocturnal sky nnd F２-ionosphere would be one and the same in the upper atmosphere.
Ⅲ Measurements of the Relative Intensities of the Forbidden Radiations (λλ 6300, 6364 and 5577) in the Light of the Night Sky.
The determination of the relative intensities of the predominant forbidden radiations λλ6300，6364: OI(p4 3P2・1 - 1D9) and λ5577: OI (p4 1P2・1 - 1S0) in the light of the night sky seems to be useful from the standpoint of quantum mechanics and for the explanation of the conditions in the upper atmosphere.
Hitherto, no one has attempted to do this important measurement, because the spectrophotometry of the light of the night sky, which is very feeble in intensity is accompanied by technical difficulties.
The author observed the feeble light by the F/l. 9， 1-prism spectrograph in 10 nights from March to April, 1941.
The relative intensities I6380+6364 / I5577 of these 18 exposures were measured spectrophtometically compared with the intensities of a Hefner lamp at the rspective wave lenghth.
As the dispersion of the mean spectrograph was very small he could not separate the line λ6364 from that ofλ6300，
To determine the mean relative intensities I6364/I5577 and I6300/I5577 he examined the spectrogram which was thaken by the F/4.5， 2-prism spectrograph with an exposure of about 150 hours. From this spcctrum he estimated the relative intensities,
I6364/I6300=0.5 and I6364/I5577=3.01
and so, I6300+6364/I5577=4.52
The intensity ratio 4.5 is in satisfactory agreement with that(4.3) obtained by the F/l.9. 1-prism spectrograph.
These values are given in Table 9.
Ⅳ. Measurements of Diurnal Intensity Variations of λ6300: OI(1D2-3P2), λ5577:OI(1S0-1D2) and the Relation between these Lines and the Electron Density of F2-Ionosphere.
The observations of many investigators have proved beyond doubt the occurrence of the diurnal variations in the intensity of the green light λ5577 in the night sky.
Some years ago, the author also observed the variation of the green light by the filter method. He noticed two kinds of variations; one of which was a gradual increase in intensity with a maximum at half an hour after midnight，followed by a slow decrease until the end of the night(be denominated such a variation as W-type)，while the other showed a gradual decrease with its minimum at the midnight (he denominated such a variation as V -type).
Rayleigh, McLennan & others have observed only the W-type, and Karandikar & Ramanathan have observed only the V-type.
The purpose of this investigation was to study the diurnal variations for both the red (λ6300) and the green (λ5577) light in the night sky by the methods of filters and a spectrograph，and to make reference of the relation between these emission lines and the electron density of F２-ionosphere.
The observations by the filter method were made by both red and green filters at the same time in the directions of the zenith and the horizon from 1938 to 1941, and the F/I.9，1- prism spectrograph was occasionally used by sliding the plate，each exposure time of which was about 2 hours.
From these materials he noticed the following facts.
1. As to the green light the W-type variation occurred about 70% and V-type about 30% (total number of these data was 58)，while on the red light the W-type variation occurred about 46%，and V-type 54% (total number19).
2. The intensity variations of the zenithal and horizontal directions showed almost similar appearance for both red and green light.
3. The Variations of the intensity of the green light were compared with the nocturnal F２-layer ionization (data 10).
It is seen that relatively large variation of the nocturnal electron density occurred in F２-ionosphere and the markedly negative correlations would appear in such cases as those the variations of the green light were W-type，and when the variation of the green light indicated the V-type，the electron densities of F２-layer were nearly const or progressive decreasing.
However，sometimes both variations seemed to have positive correlation.
4. The author examined the intensity variations ofλ6300 and λ5577 obtained by the spectrograph and F２-ionization，and the noticed that the variations of the light of the night sky and F２-electron density were always closely correlated positively if the intensities of λ6300 and λ5577 were reasonably added according to the following assumptions.
The radiation λ6300 is produced by two diiferent processes，optical forbidden transition and electron impact，with probabilities of the same order，and the former transition has no effect on the electron density of the F２-ionosphere. The radiation λ5577 is mainly produced by attachment process. Accordingly，the transition λ6300 is predominant from sunset to midnight，and the transition 5577 is prominent from midnigh to sunrise，and both transitions are interruptd by each other.
Such assumptions seem to be by nomeans unreasonable.
By the fact that the intensity of the red OI-triplet，relative to that of the green line increases strikingly towards twilight and decreases around midnight, it is possible to consider that the colisional forbidden transitions 1D2 - 3P2 of the neutral oxygen atom play an important role in the sunlit portion of the upper atmosphere and that the detachment proces １S o- 1D2 of O-，which may be produced by attachment or some other mechanisms are the main source in the dark upper atmosphere. From the idea the assumption mentioned above may be fairly confirmed.
