is evident even for this frequency range. What is more important is that 
the orders of magnitude of the power spectra from these two experiments 
are similar, The data of 50 m is puzzling because it is so much lower in 
level than that of 500 m. This difference was explicitly pointed out in 
Reference 2. 
The power spectra of the temperature at 150-ft depth as obtained 
by BT’s taken during the course of the three days, every half-hour, is 
also plotted in this figure. Here again the order of magnitude and the 
rate of falloff are similar to the data of 500 m. The rise at 1 cycle 
per hour is due to aliasing, which shows that the BT responds to temper- 
ature changes at a greater rate than it cought to for observations taken 
at this slow rate. 
A second important result shown in Fig. 8 is the absence of lines 
of any sort in the power spectra. Essentially this means that this data 
is not susceptible to interpretation as being composed of a si:ple har- 
monic temperature wave or even as a combination of a few simple harmonic 
waves. This is quite different from the results obtained by early experi- 
menters who reported the frequency and wavelength of the simple harmonic 
waves with which they reconstructed the experimental temperature record. 
Perhaps the temperature records are analogous to sea surface records, which 
are describable only by their power spectra and by statistical measures. 
CONCLUSIONS 
For the data obtained in this set of one-day observations, spread 
over 24 miles horizontally and 120 ft vertically, the power spectra of the 
temperature “hojimtinns have been found to obey the fe law in the fre- 
quency range from 0.5 to 12 cycles per hour. No lines have been observed 
in the spectra. These results differ from those of early experimenters. 
However, they cannot be considered to be at varia:ce with the data reported 
in the references cited in Footnotes 1 and 2 bocause the frequency range 
covcred differs so widely. 
SiGe 
