METEORS AS PROBES OF THE UPPER ATMOSPHERE 
Fig. 2 these residuals A logy p are plotted against the 
mean average ground temperature at Boston (lat. 
42.5°N) at the date of each meteor. Three velocity 
6 4 
22°C 
X v< 25 km/sec 
© 25<v<40 km/sec 
A v>40 km/sec 
Se — 
—0.6 
Fig. 2—Seasonal density variations from photographic 
meteors. 
groups among the meteors exhibit the same correlation. 
A least-square solution shows that logi p increases 
0.019 + 0.0013 (P.E.) per degree centigrade of ground 
temperature. This result applies at a mean height of 78 
km. The total seasonal range would be 0.46 in logo p, 
corresponding to a height variation of 8.6 km. 
The seasonal correlation is less marked when the 
solar declination is used in place of the ground temper- 
ature. The correlation is not improved by comparison 
with the actual ground temperature at the date rather 
than the general average; hence, the correlation is 
truly a seasonal one which does not measure local 
variations. No effects associated with synoptic weather 
fronts, deviant temperatures in the lower stratosphere, 
sunspot numbers, lunar-hour angle or solar-hour angle 
are conspicuous. From further meteor investigations, 
including also atmospheric data from the beginning and 
end points of photographic meteors, Jacchia [36] finds 
evidence that the seasonal effect decreases with increas- 
ing height, becoming small and uncertain around the 
100-km level. 
Combining the photographic meteor data from de- 
celerations and beginning-point data, Jacchia [36] has 
derived the upper-atmospheric density distribution 
given in Table I. Heights are given in kilometers above 
sea level and density in grams per cubic centimeter. 
The deviations of logy) p from the N.A.C.A. Tentative 
Atmosphere are given in the third column; the N.A.C.A. 
values have been adjusted slightly to avoid discontinu- 
ities in the temperature gradient, which are awkward 
physically. The values of atmospheric temperature in 
Table I are calculated with a constant molecular weight 
but include the decrease of gravity with height. The 
temperatures may be varied greatly within the range 
of solution, particularly the minimum values near 83 
km and the values above 100 km. 
The observational basis of Table I is shown in Fig. 3. 
The deceleration data are considerably more reliable 
than the beginning-point data. The results given in 
Fig. 3 and Table I have been derived purely from 
361 
meteoric data without reference to other methods except 
for the zero point in logio p. The simple theory has been 
used, the power of the velocity having been varied to 
TaBLe I. Aporrep Drnsity PROFILE 
H A lo; Derived 
(km) log p MePR CAD (deg K) 
70 —6.801 0.000 281 
72 —6.883 0.001 266 
74 —6.971 0.001 252 
76 —7.067 —(0.001 240 
78 —7.173 —0.004 231 
80 —7.287 —0.006 224 
82 —7.412 —(0.009 221 
84 —7.543 —0.012 221 
86 —7.679 —(0.022 224 
88 —7.816 —0.033 229 
90 —7.951 —0.044 235 
92 —8.083 —0.057 241 
94 —8.211 —0.070 246 
96 —8.334 —0.080 250 
98 —8.453 —0.090 253 
100 —8.569 —0.099 255 
102 —8.683 —(.107 256 
104 —8.795 —0.115 256 
106 —8.908 —0.124 256 
108 —9.019 —0.132 256 
110 —9.131 —0.141 256 
produce the best interval agreement among the meteoric 
data. 
The determination of atmospheric densities and 
seasonal effects from the photographic observations of 
HEIGHT (km) 
50 60 70 80 90 100 ike) 120 
T 
T T T T 
LOG Pore 
Oe * DECELERATION 2 \\ 
© BEGINNING POINTS O 
1 TABLE I 
— A ° 
Eon NAC 
=|0)0)'——= 
Fre. 3.—Atmospherie densities from photographic meteors. 
meteors can be improved by at least four methods of 
approach: 
1. More numerous and accurate observations of me- 
