520 
Figures 13a, b [47] show the actual wind-speed vec- 
tors due to the S. and L,. wind variations at each hour 
of either half of the solar or lunar day (this diagram, of 
course, is not a dial diagram like Fig. 12). These wind 
velocities are superposed on any other local winds pres- 
MAURITIUS WIND (16 YEARS DATA) 
LUNAR TIDE 
270° (a) 
MAURITIUS WIND 
SOLAR TIDE 
10 20 cm sec 
°° 
(b) 
Fig. 12—(a) Harmonie dial (with probable-error circles) 
for the annual mean lunar semidiurnal variations in the north- 
ward and eastward components of wind velocity at Mauritius, 
from about 16 years’ bihourly data. (6) Harmonic dial for the 
corresponding annual mean solar semidiurnal variations. Note 
the tenfold scale difference between the two diagrams. 
ent. The speed scale in Fig. 13a (lunar) is ten times 
more open than that for Fig. 13b (solar). The diagram 
Fig. 13a is, of course, not well determined. 
The corresponding semidiurnal paths of Mauritius 
air particles due to these oscillations are similar in form 
and orientation to the ellipses of Fig. 13, which will rep- 
resent these paths if all the time marks are advanced 
by three hours, and if the speed-scale is changed to a 
length-scale 7/27 times as great, where 7’ denotes the 
duration in seconds of the (solar or lunar) half day. 
This factor is 6876 for the solar diagram 13b, and 7114 
for the lunar diagram Fig. 13a. The distance scales are 
indicated on the left of each diagram. The extreme de- 
parture of any air particle from its mean position, at 
DYNAMICS OF THE ATMOSPHERE 
Mauritius, owing to these oscillations, is about 23 km 
for S; and about 1 km for Lp. 
The Lunar Tidal Rise and Fall of the Ionospheric 
Layers. It has long been inferred from the evidence of 
the geomagnetic variations that there are horizontal 
lunar tidal currents in the ionosphere, the ionized elec- 
trically conducting region of the high atmosphere, con- 
taining at least two distinct layers, the E-layer at about 
100 km height, and the F-layer at about 250 km. The 
first direct determination of the lunar tidal rise and fall 
of the high atmosphere was made in 1939 by Appleton 
MAURITIUS 
WIND 
(16 YEARS" DATA) 
Fig. 13.—Diagrams based on Fig. 12, showing in plan the 
wind velocities at each lunar or solar hour (morning and 
afternoon) associated with the lunar and solar semidiurnal 
variations of wind at Mauritius. The velocity at each hour is 
represented (on the scale shown at the right) by the line (not 
drawn) from the center of the diagram to the numbered point 
for that hour. The diagrams also illustrate the corresponding 
paths of an air particle at Mauritius due to these wind var- 
lations, if the hour-numbers are increased by three; the dis- 
tance scales are shown on the left of each diagram. Note the 
tenfold scale difference between the two diagrams. 
and Weekes [6], who found from hourly radio measure- 
ments of the height of the E-layer above Cambridge, 
England, throughout several weeks, a lunar semidiurnal 
(Le) variation of height of the H-layer amounting to 
1 km above and below the mean level. This remarkable 
result is illustrated in Fig. 14, which shows eleven dial 
points each representing a determination from a period 
of twelve to fourteen days, between August 1937 and 
July 1938. The cross shows the mean dial point, and 
the probable error circle indicates the uncertainty of 
any one of the eleven points. The radius r for the mean 
point is 1/1/11 times less than this uncertainty, so 
that the determination is a good one. 
Martyn [76-79] has lately examined the lunar tidal 
variations in the heights of the H- and F-regions, using 
the less accurate data available from routine recording 
at various ionospheric observatories throughout the 
