HOURLY observations: OF AIR TEMPERATURE AND PRESSURE. 
623 
Part I.— Temperature. 
Greemvich Temperature. 
The series of Tables dealing with Greenwich temperature shows that, with consider¬ 
able variations in the absolute values of all the p, 5 coefficients for the several months, 
and from year to year, as well as of the total amplitudes P, and the epochs of 
maximum p, there is still a strongly marked general consistency in the characteristics 
of the various elements. Reference to Table B annexed to this paper (which is 
Table VIT. of the Greenwich series above described) shows very clearly that, with 
relatively few exceptions, chiefly to be found in the winter months, and among those ' 
components the absolute values of which are very small, the mean monthly epochs of 
maximum in the several five-year periods differ among one another by very small 
amounts, corresponding to intervals of time seldom as large as half-an-hour. As, 
however, the weight of tlie conclusions that may be drawn from these computations 
depends greatly on the internal evidence supplied by their greater or less consistency 
among themselves, it is important to dwell somewhat more on this point. 
The requisite test wall best be supplied by the values of p which indicate the times 
of maximum phase, and are, therefore, immediately dependent on the Sun’s action, 
and should follow his position ; and this being subject to regular periodical change in 
the course of the year, it would follow^ that if the values of p show corresponding- 
regular, or nearly regular, periodical variations, they may be regarded as affording 
valid internal evidence of being trustworthy. The values of P, the total amplitude 
or variation of temperature, which are certainly subject to influences of a much more 
local and irregular character, will be liable to much greater and more irregular 
changes. 
A convenient and succinct summary of the facts bearing on this point will be found 
in the following Table, which shows for the series of 20 years the number of times 
(reckoned as a percentage) in which the computed values of p for single months fall 
in the several quadrants, the hours corresponding to which are as follows :— 
Time of Maximum according to Value of p falling in successive Quadrants. 
Components. 
1st quadrant. 
2nd quadrant. 
3rd quadrant. 
4th quadrant. 
First . 
Second . 
Third. . . 
Fourth . . 
Midniglit to 6 a.m. 
) 1 )) 
9 
„ „ 1.30 „ 
G A.M. to noon 
3 ,, 6 A.M. 
?! 4 ,, 
1.30 3 „ 
Noon to 6 P.M. 
6 A.Jl. to 9 A.M. 
4 „ 6 
3 „ 4.30 „ 
6 P.M. to midnight 
9 A.M. to n;/_-n 
G ,, 8 A.M. 
4..30 „ G ., 
