LTEUT.-GEN'f]RAL R. STRACHEY ON HAR:\royrC AXALYSTS OF 
r.38 
Note. —In 1879 Sir G. Stokes communicated to the Meteorological Council, of 
which he was tlien a member, copies of letters addressed by him to Sir G. Airy, 
containing the results of computations of the harmonic components of the Greenwich 
diurnal temperature curve, made in the ordinary manner, which differs somewhat 
from that adopted by me. These results supply the coefficients up to the fourth 
order for each of the 20 years of the Greenwich series, separately, and those of the 
fifth and sixth orders for the mean of the 20 years taken together. They are almost 
identical with the values given in the volume of tables published by the Meteorological 
Council, and the comparison of the means of the two sets of computations may be of 
some interest. 
Vv 
Iv 
Pi- 
A- 
ft- 
ft- 
Pi- 
ft- 
Po- ' Qy 
Pe- ?«• 
Sir G. Stokes 
-4-280 
~2-722 
+ -883 
+ -563 
+ •063 
+ •119 
-•043 
--014 
-■001 --026 
-■003 --008 
Meteoroloo'ical 
1 
Tables . . 
-4-28-2 
-•2-738 
+ 
CO 
00 
+ •557 
+ -066 
+ •118 
-•046 
-•015 
1 
P 
1- 
P 
3- 
Pfi- 
Sii’ G. Stokes 
5-072 
1-047 
•135 
•045 
•026 
•009 
Meteorological 
1 
Tables . 
.5-08.3 
1-046 
•1.34 
•047 
It has not been thought worth while to repeat the detailed figures for the separate 
years, as they are virtually identical with those given in the Meteorological Office 
Tables. 
Referring to a comparison that he had made between the mean observed hourly 
temperatures and the values obtained by computation from the daily means and the 
harmonic coefficients, Sir G. Stokes remarked that “ the minute fourth order is 
swallowed up, even in a mean of 20 years’ series, by the uncertainty in the amount of 
the diurnal fluctuation, not to speak of the much larger uncertainty in the mean 
temperature for a day. Nevertheless, it appears from the figures, that though this 
order is so minute, and though its effect on the mean temperature at any particular 
hour is trifling compared with the average variation from year to year of the sapie 
hour, it still emerges roughly with tolerable certainty from the mean of a good many 
years. ” 
This conclusion is fully confirmed by the more complete examination of the results 
obtained from the mean monthly values, much of the strong evidence of the periodicity 
represented by the minute fourth order of components being lost by dealing with the 
mean yearly values. 
