ATMOSPHERIC TIDES AND OSCILLATIONS 
is reduced near the Pacific Coast [16]. This was noted by 
Hann, who also found that sz is less on the east Adriatic 
Coast than in Italy, and in the West Indies as compared 
with the East Indies, where indeed sx, like J, is abnor- 
mally large. 
The Annual Variation of S.. The solar semidiurnal 
barometric variation also shows, as Hann indicated 
[67], considerable regularity in its change throughout 
the calendar year. This is illustrated in Fig. 16 by dial 
diagrams for four widely spaced stations in temperate 
latitudes (N or S), namely, (a) Washington, D. C. [16]; 
(6) Kumamoto (33°N, 131°E) [43]; (©) the mean of 
a 
) {e) 
COIMBRA-LISBON- SAN FERNANDO (c) 
o 
523 
@2 has a mean value of 311° from 0° to 40°S latitude, 
and 299° from 0° to 40°N. The corresponding means of 
o and #; are 0.020, 0.078 and 134°, 94°. He discussed 
in much detail the regional irregularities in the distribu- 
tion of a; and ;. 
As regards o2, Hann concluded that from 14°S to 
50°N it has its maximum in January, and south of 14°S, 
in July. Figure 16 does not altogether confirm this. The 
many other details of Hann’s discussion of S». cannot be 
summarized here. The data now available for S. and 
L, call for a more comprehensive comparative discussion 
than has yet been attempted. 
MONTEVIDEO (d) 
Fic. 16—Harmonic dials indicating the annual change in the solar semidiurnal barometric variation (S2) at four widely spaced 
points in middle latitudes, (a) Washington, D. C., (b) Kumamoto (33°N, 131°E), (c) mean of Coimbra, Lisbon, and San Fernando, 
(d) Montevideo, Uruguay. 
Coimbra, Lisbon, and San Fernando [57] (as in Fig. 
10f for Lz); and (d) Montevideo, Uruguay [67]. 
The annual paths of the S, monthly dial points are 
much better determined than the ZL, paths in Fig. 10, 
yet they seem to differ more from one another, suggest- 
ing greater regularity of annual variation for L». than 
for S2, which if established would be very remarkable. 
Hann [67] considered separately the annual variations 
of s. and o», though it would certainly be better to 
treat them together, that is, to discuss the annual move- 
ment of the S, dial point. He expressed the yearly varia- 
tion of s2 as the sum of a twelve-monthly and semian- 
nual term: 
So = a + a sin (wv + B,) + ae sin (2x + Bs), 
reckoning « from mid-January at the rate 360° per 
year. He found that a. decreases polewards from the 
equator rather regularly, from 0.075 mm at the equator 
to 0.035 at 30° and 0.026 at 60° latitude, and that 
THE THEORY OF THE ATMOSPHERIC 
OSCILLATIONS 
Newton to Kelvin. Newton, in his de Mundi Systemate 
(the third part of his Principia) remarked that universal 
gravitation implies a tidal ebb and flow in the atmos- 
phere as well as in the oceans. He rightly considered that 
it would be inappreciable—though it has been seen 
(on p. 519) that this flux and reflux can be determined 
by extensive computation from sufficient wind data. 
Laplace in his dynamical theory of tides laid the 
foundation for all subsequent work on the subject. He 
first determined the tides of a liquid ocean completely 
covering a spherical rotating earth. In most of his work 
he took the ocean to be of uniform depth. Later he 
showed that he could associate the theory for such an 
ocean with the tidal oscillation of the (compressible) 
atmosphere, provided that the vertical accelerations are 
neglected (which we now know is permissible), that the 
