308 
The goal of the expedition was to obtain measures of 
meteor velocities, heights, radiants, and statistical data 
visually by the use of the “rocking mirror” technique. 
In applying this technique the observer determines the 
angular velocity of a meteor by looking at its reflection 
‘ in a plane mirror, the mirror being made to oscillate 
in such a fashion that a normal to its surface describes 
a conical surface of small apex angle, with the apex 
near the center of the mirror. Trajectories and heights 
were measured by simultaneous visual observations 
from two stations separated by 36 km. 
The active direction of the expedition and the anal- 
ysis of the results were conducted by Opik. From Octo- 
ber 1931 to the end of July 1933, some 22,000 individual 
meteors were observed, and heights were determined 
for 3540. Most of the latter were sporadic meteors 
(about 80 per cent) and only 7 per cent came from the 
major showers. In analyzing the heights, Opilx [65, 66] 
made statistical corrections for all effects that he 
thought might possibly introduce systematic errors into 
a seasonal phenomenon, if present. He found that an 
average meteor corrected to standard conditions of 
brightness, velocity (actually elongation of radiant from 
apex of earth’s motion) etc., appears at a greater alti- 
tude in summer than in winter. From this he concluded 
that the data “suggest an annual fluctuation of the 
height of the atmosphere, more or less corresponding 
to the annual temperature curve, of an (total) ampli- 
tude of 3.7 + 0.7 km,” applying at a mean altitude of 
about 88 km above sea level. 
From his theory [67, 69] of the meteoric process he 
concluded that an atmospheric temperature of +100C 
at an altitude of 90 km and a mean molecular weight 
as at sea level would be consistent with the meteor 
observations by the Arizona Expedition. 
Opik’s further conclusions as to a “night effect,” in 
which the height of the ‘normal’? meteor decreases 
during the course of the night, has not yet been con- 
firmed or disproved. His conclusion [68] as to the 
appreciable percentage of ‘‘hyperbolic” velocities among 
visual meteors appears not to be consistent with the 
radio observations of meteors mentioned earlier. 
Long-Enduring Meteor Trains and Winds in the 
Upper Atmosphere. The most comprehensive study of 
atmospheric winds and turbulence as reflected in the 
motions of long-enduring meteor trains has been made 
by Olivier [62, 63], who gives data for nearly 1500 
trains. Of great importance also is the work by Fedyn- 
sky [20], who reports on 41 night trains observed sys- 
tematically for the purpose. 
Olivier finds from more than fifty cases that the 
average beginning height of night trains is 104 km, 
with average end heights of 80 km, while the corre- 
sponding data for twilight trains are 77 and 45 km 
respectively (twenty-six cases) and for daylight trains, 
45 and 27 km (nine and fourteen cases). He calculates 
from the observed motions that the average wind 
velocity is 203 km hr~ for the fifty best-observed night 
trains, 182 km hr for night trains whose heights have 
been assumed, and 173 km hr for day or twilight 
trains. A more detailed analysis suggests that the winds 
THE UPPER ATMOSPHERE 
increase with altitude, that is, from daylight to night 
trains. It is clear, from both photographs and drawings, 
that “layers of the atmosphere, lying quite close to- 
gether, have winds blowing in different directions and 
at quite different velocities.” In very brief intervals of 
time after the appearance of trains ‘‘the photographs 
show that the trains have become zigzag in shape... . 
There seem to be inescapable proofs, in certain cases, 
of vertical components.” 
As for the directions of the train motions, Olivier 
finds that for America the greater number drift north 
with strong east-west components. See Fig. 1. For all 
BEFORE MIDNIGHT 
-——--AFTER MIDNIGHT 
Fic. 1.—Frequency of night wind vectors over America. 
Europe an eastern tendency is strong, while other 
areas of the earth show various preferential drifts. Un- 
fortunately, Olivier has not searched for seasonal effects 
—a search that almost certainly must lead to important 
conclusions concerning upper-atmospheric circulation. 
The results given by Fedynsky include a number of 
unusually high velocities, of the order of 1000 km hr~, 
and it is difficult to appraise the accuracy of the ob- 
servations. His average deduced velocity is 391 km hr+. 
Additional confirmation would be desirable before ac- 
cepting completely his separation of the winds into two 
groups of low and high velocity. 
PHOTOGRAPHIC METEOR STUDIES 
Methodology. Since photographic meteors have been 
used as a tool for research on the upper atmosphere 
almost solely at the Harvard Observatory, the discus- 
sion of this section will be confined largely to a short 
résumé of that work. 
Systematic photography of meteors by the use of 
two cameras equipped with rotating shutters was begun 
by Elkin [16, 17] at the Yale Observatory in 1893 and 
continued until 1909. The astronomical results of this 
