8 METEOROLOGICAL RESULTS OF LAST CRUISE OF CARNEGIE 
Table 9. Comparison of 12-hour waves of atmospheric pressure observed ou Carnegie, 
1928-29, and on Gauss, 1901-03 
(Values computed from mean of all data within the indicated latitude range) 

Gauss 





Latitude : 
Phase Ampli- 
ae: angle 
a :. . mm mm - mm mm 
55 N--5 N 147.9 0.129 OORT" Beeek Seas Re eeeees 
45 N-35 N 138.5 0.212 0.047 155.6 0.329 0.075 
35 N-25 N 147.8 0.420 0.039 156.0 0.511 0.058 
25 N-15 N 155.2 0.625 0.035 154.9 0.708 0.050 
15 N-15S 159.1 0.829 0.015 154.6 0.818 0.021 
15 $-25S8S 160.6 0.759 0.036 155.6 0.701 0.052 
25S-35S 150.0 0.479 0.051 160.5 0.496 0.038 
35 S-45S 138.2 0.278 OS09 Sie eee ee ee eos e ns eceeeee 

Table 10. Comparison of 12-hour waves of atmospheric pressure from observations at sea and 
as computed by Simpson 





Mean 
latitudes 
mm 
mm 
15 N and 15 S@ 159.1 0.829 0.015 
20 N and 20 s3 158.2 0.692 0.020 
30 N and 30S 148.9 0.450 0.026 
40 N and 40S 138.3 0.244 0.036 




mm mm mm 
156.3 0.852 0.011 154 0.924 
153.1 0.662 0.016 154 0.770 
150.3 0.501 0.023 154 0.609 
158.1 0.338 0.012 154 0.422 

2 Values determined from means of all data obtained within 5° north or south of indicated 
latitudes. 
© After Simpson. 
however, for the two sets of observations are not com- 
parable with respect to either season or longitude. 
A similar comparison has been made of the 12-hour 
wave as computed from the Carnegie data, and as com- 
puted from data averaged for various mean latitudes by 
Hann [10] from observations made on the Challenger, 
Novara, Saida, and Donau (table 10). In order to obtain 
means for latitudes comparable with the ranges of lati- 
tude selected for assembling the data of the Carnegie, 
Hann’s values have been averaged for each 10° of lati- 
tude by taking the mean Fourier coefficients, ag and bog, 
and computing new values of ¢2 and C2. The number of 
observations is large; therefore radii of the probable- 
error circles are small. Simpson’s values, on the other 
hand, were obtained by assuming the required latitudes 
for the equatorial part of the 12-hour vibration (equation 
2), and it was thus impossible to construct probable-er- 
ror circles for his data. 
Figure 7 emphasizes the fact that the amplitude of 
the semidiurnal pressure wave is smaller over the 
oceans than over land areas. Simpson’s formula was 
consiructed chiefly from pressure observations at land 
stations. At latitude 40° north, the amplitude obtained 
from Hann’s values is 80 per cent of that computedfrom 
Simpson’s formula. At this latitude the Carnegie values 
for the amplitude of the pressure wave indicate only 58 
per cent of the computed value. The amplitudes at other 
latitudes average around 85 per cent of Simpson’s values. 
The harmonic dials given in figures 6 and 7 show 
clearly that the amplitude of the 12-hour wave decreases 
with increasing latitude. Various investigators, notably 
Hann [11], Schmidt [6], Margules [12], Jaerisch [13], and 
Meinardus [4, p. 454], have attempted to evolve a mathe- 
b From observations on the vessels Novara, Saida, Donau, and Challenger. 
matical formula which would express this rate of de- 
crease in amplitude with latitude. The general form for 
all the suggested formulas has been to place c¢ equal to 
a constant multiplied by some power of the cosine of lat- 
itude. The constants were usually computed from pres- 
sure data obiained all over the world, irrespective of 
land or ocean position, and tended to be heavily over- 
weighted by data from land observatories. All these 
previously determined formulas, except that of Meinar- 
dus, obtained from observations on the Gauss, and that 
of Margules (who assigns no value to the constant fac- 
tor), give amplitudes much too large for purely oceanic 
areas. Moreover, none of these earlie' formulas as- 
sume the amplitude to be a function of longitude or sea- 
son. Simpson [5, p. 12], in 1913, by combining his equa- 
tions for the equatorial and polar vibrations (equations 
2 and 3), developed the following formula which sets 
forth Cg as a function of longitude as well as of latitude, 
wherein A is longitude east of Greenwich 
Co = [ {0.937 cos? sin 154° + 0.137 (sin2@ - 1/3) 
sin (105° - 2a )}2 + {0.937 cos%¢ cos 154° 
+ 0.137 (sin2@ - 1/3) cos (105° - 2a 2 1/2 4 
In order to determine how closely values for the 
amplitude of the pressuré wave, as computed from Simp- 
son’s formula, agree with the Carnegie values, the mean 
longitude and mean latitude positions corresponding to 
each of the Carnegie values for Cg were computed. Only 
the Carnegie values from the Pacific Ocean west of lon- 
gitude 180° and south of latitude 5° north have been used 
in these computations. As shown in table 11, the differ- 
