HOURLY OBSERVATIONS OF AIR TEMPERATURE AND PRESSURE. 
G29 
The relation of the epoch of the first maximum in the third component to the time 
of sunrise—which, though directly dependent on the Sun’s declination, may, of course, 
equally be regarded as a function of the Sun’s longitude—seems to be decidedly 
marked; the former occurring, on the average of the whole year, at an interval of 4S'’, 
or 3 h. 12 m. earlier than the latter, or, what is the same thing, the first minimum 
occurring 12° or 48 minutes after sunrise, the mean deviation of the interval from 
that average being only 7°, or 28 minutes of time. A reference to Table B will 
illustrate this. 
The periodical variation in the value of pj leads to the third component having a 
'positive maximum about 1 p.m. during the winter months, from October to February, 
which will be accompanied by maximum negative values four hours earlier and foin* ' 
hours later, corresponding wdth the reduced temperature in the mornings and after¬ 
noons of the shorter days. (Plate 23.) 
In the summer months, from April to August, pg has a maximum negative value 
about 1 P.M., instead of o, loositive maximum as in winter; and this, being accompanied 
by two positive maxima, one four hours earlier and the other four hours later, will in 
like manner correspond to the higher temperature in the mornings and afternoons of 
the longer day. 
It will be seen that the positions of the midsummer and mid-winter maximum 
phases correspond respectively to days of IG hours with nights of 8 hours, or days of 
8 hours with nights of 16 hours, and that at these seasons, when the variations of 
temperature due to these differences are greatest, the values of the amplitude of the 
component are also at a maximum. At the equinoxes, with a 12 hours day aird 
night, the amplitude of the component has a minimum value, and the transition of 
the position of the maximum phase takes place as already described. 
It might, perhaps, be expected that the change in the position of the maximum 
phase of the fourth component, would take place in an analogous manner, in con¬ 
nection with days and nights of 6 and 18 hours duration respectively, corresponding 
to the epoch of the component. But such days and nights will only occur in higher 
latitudes than those of our observatories, and no data are available by which to test 
such a suggestion. 
Although the several harmonic components of the temperature curve cannot be 
regarded as indication of specific physical efiicient causes, acting at definite periodical 
harmonic intervals, the graphical representations of the series of monthly temperature 
curves and their components, which are given in the plates accompanying this j)aper, 
present some points to which attention may usefully be drawn. 
As before noticed, the deviation of the day temperature from the first component is 
extremely small in the months of May, June, and July, during which between 8 A.M. 
and 8 p.m. the other components nearly cancel one another. From September to 
March howmver, the temperature during the day hours rises considerably above the 
first component, the second and third component either being both positive and 
