618 
LIEUT.-GENERAL R. STRACHEY ON HARMONIC ANALYSIS OF 
com})onents, extending to the fourth order, have been made from the Greenwich data 
the method proposed by myself (and explained in the ‘ Proceedings of the Pioyal 
Society,’ vol. 42, pp. 61-79), with the help of the Tables to be found in the Appendix 
to the “ Hourly Headings of the Self-recording Instruments at the Observatories of 
the Meteorological Council ” for 1884. 
The values of the constants for the seven observatories of the Meteorological OflSce, 
extending to the third order, were obtained from the photographic records by means 
of the mechanical analyser designed by Sir W. Thomson, a description of which will 
be found in the ‘Proceedings of the Royal Society,’ vol. 27, p. 371, and as to the 
use of which and the degree of dependence to be placed on its indications reference 
may be made to the ‘ Annual Reports of the Meteorological Office,’ for the years 
1881, 1883, 1885, and 1890, and to the ‘Proceedings of the Royal Society,’ vol. 40, 
p. 382. 
In these Tables, following the usual formula expressing the hourl}^ value A„ in terms 
of its harmonic components, 
A„ = 2^0 + Pi cos n 15° + sin 15° + cos 2 n 15°, &c., 
2Jq represents the mean value for the whole twenty-four hours ; the coefficients of 
the cosines are designated by the letters q)vP- 2 ’ Ps’ ‘^c., for the several orders of com- 
])onents ; and those of the sines by the letter q^, qo, q^, &c. The values are expressed 
in degrees Fahrenheit for temperature, and in inches for pressure. 
The total amplitudes of the several components in the corresponding formula, 
involving sines only, are designated by the letter P^, Pg, &c., where Pi-= \/(pi~ + q\^), 
and 
A,I = Pq + Pj sin (n. 15° -J- Tj) -f- Pj sin (2?<.15° + T 3 ) -|- P 3 sin (3n.l5° -|- Tg), &c. 
The Tables, however, instead of giving the values of the subsidiary angles T, show 
the epochs of the first maximum reckoned from midnight in angular measure, the 
letters pg, pg, &c., indicating the epochs for the several components. This method 
of stating the phase or epoch of the components has the advantage of directly indi¬ 
cating their relation to the hour of the day, and, therefore, to the position of the Sun, 
to which the diurnal variations of temperature and pressure should obviously be 
referred. 
If it be understood that tan x = 2 Jlq, disregarding the signs of the coefficients, then 
the quadrants in which T will fall when regard is had to the signs of p and q, and 
the relation of p to T and x, for the several components, are shown in the following 
Table. 
