HUMIDITY 
urnal curve of vapor pressure would follow the curve of 
mean hourly air temperature very closely. The Carne- 
gie data appear to bear out this conclusion (compare fig. 
31 with fig. 16). Although the curve of mean hourly va- 
por pressure is somewhat irregular, there definitely 
appears to be a well-defined maximum at 13h, as was 
the case with air temperature, and a less well-defined 
minimum at 05h. 
Variation of the Diurnal Amplitude of 
Vapor Pressure with Latitude 
As shown in table 5i, the diurnal amplitude of vapor 
pressure appears to vary with latitude in much the same 
manner as the diurnal amplitude of air temperature for 
the ranges of latitude north of the equator, with maxima 
at mean latitudes +10° and +30°, and with minima at 
mean latitudes +20° and at the equator. It may also be 
observed that the amplitude at 10° mean south latitudeis 
less than at mean latitude 10° north in both cases, sig- 
nifying that identical conditions tend to produce maxima 
and minima in the diurnal amplitudes of both air tem- 
perature and vapor pressure. 
Effect of Wind on the Diurnal Amplitude 
of Vapor Pressure 
The mean unperiodic amplitude of vapor pressure 
has been computed for fifty-two days iz tropical regions 
between latitudes +20° with a wind force equal to or 
greater than 4, Beaufort scale, and for fifty-three days 
within the same latitude range with wind force less than 
4. The results give an amplitude of 1.69 mm for days 
with wind force equal to or greater than 4, and one of 
2.02 mm for days with wind force less than 4. Obvious- 
ly the wind tends to reduce the daily range of vapor pres- 
sure, presumably because of the more thorough mixing 
of the surface layers of air. 
Harmonic Analysis 
of Vapor-Pressure Data 
As shown in table 52, the amplitudes and phase an- 
gles for the 24-hour, 12-hour, and 8-hour terms are ex- 
tremely variable between the various Groups. There 
appears to be some regularity, however, in the time of 
occurrence of the maximum, which generally appears 
about 13h. The average periodic amplitude of vapor 
pressure for all Groups is 0.2 mm. 
Table 53. Variation of vapor pressure 
with differences between sea and air 
temperature, Carnegie, 1928-29 
At Vapor No. 
Temperature days 
a 6 mm 
>+1.0 17.73 16 
+0.6 to +1.0 16.22 31 
<+0.6 17.44 41 
< -0.6 16.45 54 
-0.6 to -1.0 13.63 18 
>-1.0 12.69 14 
Mean and total 15.69 174 
Weighted mean 16.17 sista 
45 
Variation of Vapor Pressure with 
Sea- and Air-Temperature Differences 
In order to determine the effect of differences of 
sea and air temperatures (sea minus air) on vapor pres- 
sure, a sampling of the data has been made and the dis- 
tribution of vapor pressures for various ranges of sea- 
and air-temperature differences has been determined 
for 174 days of observations. An attempt was made to 
obtain a true sampling, care being exercised to secure 
equal numbers of days with given representative tem- 
peratures and vapor pressures. It is realized that such 
a method is faulty and might lead to erroneous interpre- 
tations. Two separate sets of data were first analyzed, 
however, (82 and 92 days respectively), and both sets of 
data, separately, produced essentially the same results, 
although in one case the difference between the mean 
values at ranges +0.6 to +1.0 °C and less than +0.6 °C 
was small, the two values being almost equal. 
It is obvious from figure 33 that vapor pressure in- 
creases as the sea temperature becomes higher than air 
temperature, and lower as the air temperature becomes 
successively higher than sea temperature. Helland- 
Hansen [32, p. 12], from observations on the Michael 
Sars, has reached the same conclusions. 
A plateau in the curve (fig. 33) might well be expect- 
ed between the ranges +0.6 to +1.0 °C and less than 
+0.6 °C, as here the differences would be slight and 
affected by chance variations. The authors, however, 
hesitate to present further interpretations of this curve 
in view of the possibilities of inaccuracy due to faulty 
sampling. 
Variation of Vapor Pressure with Latitude 
Data concerning the mean vapor pressures for the 
various ranges of latitude are presented in figure 32. 
Comparing the curve in figure 32 with the curves for 
mean sea and air temperatures (fig. 14), it may be noted 
that the profiles are identical except between mean lati- 
tudes 30° and 40° south. Evidently, either the mean va- 
por pressure within the range of latitude 25° to 35° south 
is too low, or the value for the range 35° to 45° south is 
too high. It may be remarked that there were only nine 
days of observations within the range 35° to 45° south, 
and thus it is quite possible that the values for this 
range are too high. An examination of the log abstract 
was made in order to determine the type of weather 
which prevailed during the period that the Carnegie spent 
in these latitudes (December 21 to 29, 1928). It was 
found that the entire period was cloudy or foggy, winds 
prevailed from a northerly direction, and there were fre- 
quent intermittent rains. No doubt a longer series of 
observations within this range of latitude made under less 
persistent meterological conditions would give a lower 
mean value for vapor pressure. 
A Comparison of Mean Vapor Pressure 
for Rain Days and Rainless Days 
Owing to the loss of the Carnegie precipitation data 
when the vessel was destroyed, it is impossible to cor- 
relate changes in vapor pressure with amounts of pre- 
cipitation. Data have been compared, however, for per - 
iods of thirty-one rain days and thirty-one fair days, all 
within the tropics between latitudes 20° north and 20° 
south. An attempt was made to secure a true sampling 
