AIR TEMPERATURE 
Obviously this apparent condition is unreal and arises 
from the fact that the data have not been corrected for 
noncyclic changes in this case. In a vessel which is 
moving from warmer to colder latitudes there is a pos- 
sibility that the variation of air temperature may be 
greater, owing to its motion, than the usual diurnal var- 
iation of temperature. This would tend to place the max- 
imum temperature for the 24-hour period at 00h, or the 
first observation of the day. Similarly, on a vessel 
which is moving from colder to warmer regions, there 
would be a tendency to record the maximum tempera- 
ture at 23h, or the last observation of the 24-hour peri- 
ed. This fact demonstrates the necessity for properly 
evaluating meteorological data obtained on shipboard be- 
fore attempting to interpret such data. 
Taking these data on the whole, the results do not 
agree well with those drawn by Visser [24] and Braak 
[25] for several tropical regions. From observations 
made during three cruises of the Snellius inthe Nether- 
lands East Indies during 1929-1930, Visser has made 
note of the fact that the maximum air temperatures in 
this region occur between 18h and 20h, and the minima 
at 06h. These results apply to data recorded in areas 
more than 100 km from the coast, and should thus be 
comparable with the Carnegie results. Braak, using air- 
temperature data obtained between Ambon and Batavia, 
found the highest air temperatures occurring between 
16h and 20h. Visser does mention the fact, however, that 
this retardation of maximum temperature does not occur 
to such a degree on the open ocean as is obvious from 
different instances quoted in Hann’s Handbook [26]. No 
doubt the extreme retardation of maximum temperature 
in this region is due to excessive rainfall. 
Diurnal Variation of Air Temperature 
7 
General Remarks 
A study of the frequency distribution of the unperiod- 
ic diurnal amplitude of air temperature indicates that 
the daily range over the oceans is usually small when 
compared with ranges in continental or insular areas. 
Table 26 shows that the diurnal variation of temperature 
on the Carnegie was less than 3° on 71 per cent of the 
days. This result is not surprising when we consider the 
efficiency of the ocean as an energy-absorbing and stor- 
ing unit. 
Diurnal Variation of 
Mean Hourly Air Temperature for all Days 
As shown in figure 16, the mean hourly air tempera- 
tures for all days of the cruise present a fairly smooth 
curve with a definite maximum at 13h and the minimum 
at 05h. These results compare well with the frequencies 
of hours of occurrence of maximum and minimum air 
temperatures (tables 24, 25). 
Variation of the Diurnal Amplitude of 
Air Temperature with Latitude 
Assuming all heat-transport factors equal, we should 
expect the diurnal amplitude of air temperature over the 
ocean to be greatest within the ranges of latitude where- 
in air temperatures are highest, and conversely. Com- 
paring figure 14 with figure 18, however, it is found that 
this is not the case with the Carnegie air-temperature 
Table 26. Frequency distribution of the unperiodic 
diurnal amplitude of air temperature, 
Carnegie, 1928-29 

Temperature | No. {| Percentage {| Cumulative 
range days of total | percentage 
= 
<i 14 5 5 100 
1-2 105 34 34. 9S 
2-3 98 32 71 61 
3-4 57 19 $0 4629 
>4 32 10 i00. =: 110 
Total 306 100 

data. Figure 18 shows that the diurnal amplitude of air 
temperature on the Carnegie varies inversely with mean 
wind velocity. Obviously, large diurnal amplitudes are 
due in great measure to insufficient ventilation of the 
thermometer screens during periods with low wind ve- 
locities. For this reason it is impossible to determine 
with accuracy the comparative amplitudes of the diurnal 
variations in air temperature for the various latitude 
ranges over the ocean from the Carnegie data. 
Effect of Wind on the Diurnal Variation 
of Air Temperature 
As has just been indicated, wind appears to be the 
most important single factor in determining the ampli- 
tude of the diurnal variation of air temperature over the 
sea; that is, a smaller amplitude appears with the higher 
wind velocities. The reasons for such effects (in the 
case of air temperature) are two: one the mixing of sur- 
face layers of air due to greater mechanical turbulence, 
and the other the result of better ventilation of the ther- 
mometers. Although the wind data are not available in 
detail, it has been possible (from data in the log ab- 
stract) to select fifty-three days in tropical regions with 
an average wind force equal to or greater than 4 on the 
Beaufort scale, and fifty-three days within the same gen- 
eral regions, with wind force less than 4. The results 
give an amplitude of 1°79 for days with a wind force 
equal to or greater than 4, and one of 3°05 for days with 
wind force less than 4. 
An attempt was made to undertake a similar study 
of the effect of cloudiness on the diurnal variation of air 
temperature, but it was found that the records of cloudi- 
ness in the log abstract were not complete enough to al- 
low the division of a sufficient number of days into ap- 
propriate groups. 
Diurnal Waves of Air Temperature 
and Pressure Compared 
Investigators appear to agree that the diurnal varia- 
bilities of pressure are directly related to the rhythmic 
heating and cooling of the atmosphere. In this connection, 
therefore, it has seemed valuable to summarize the Car- 
negie air-temperature data in the same manner as the 
pressure data. 
The hourly values of air temperature, as given in 
table 78 of appendix II, were collected for each ten-de- 
gree range of latitude as was done for the pressure data. 
The departures of the mean 24-hourly air temperatures 
from the mean daily temperatures were determined for 
each range of latitude, and the mean diurnal variation 
