

DIURNAL OSCILLATIONS OF THE BAROMETER. 401 
In no case have the means at the hours observed been reduced to sea-level 
in calculating the amount of the oscillation, because the present state of 
meteorology as regards the methods of observation pursued does not warrant 
this being done; but the oscillations are taken directly from the means as 
observed at the place. The question of the influence of elevation on the 
diurnal oscillation will be afterwards dealt with. 
In order to show the geographical distribution of the amplitude of this 
oscillation in different regions of the globe and in different months, the amounts 
at the 335 places enumerated in Table I. and at the 49 places enumerated in 
Table II., or at 384 places in all, were entered on 13 charts for the months 
and the year (Plate XX XIV.)* Lines were then drawn representing equal 
amplitudes of 0:100, 0:080, 0:060, 0:040, 0°020, and 0-010 inch. It has not 
been judged necessary to draw lines representing a larger amplitude than 
0100 ; all such regions, however, over which in any month a larger amplitude 
obtains are shaded. The lines representing 0-020 and 0-010 inch oscillation 
are hatched, for the sake of distinction. 
The general results, broadly stated, are these :—This oscillation is greatest 
in the tropics, and diminishes on advancing into higher latitudes ; greater over 
the land than the sea, and rapidly increases on proceeding inland ; nearly 
always greater with a dry than with a moist atmosphere ; and generally, but by 
no means always, greatest in the month of highest temperature and greatest 
dryness combined. The regions characterised by largest amplitude of oscilla- 
tion include the East India Islands, Eastern Peninsula, India, Arabia, tropical 
Africa, tropical South America, and Central America, over which it either 
exceeds, or at least closely approaches, 0:100 inch. At Sibsagar, Assam, it 
amounts to 0°133 inch. It is also large in the interior of continents in equa- 
torial regions. Thus, at Mexico City it is 0111 inch; at Mafiaos, 0:126 inch; 
and at Gondokoro, if correctly observed, 0°145 inch. In the tropical parts 
of the ocean, the oscillation is from 0:020 to 0-030 inch less than on land. 
In January and July, the two extreme months as regards temperature and 
humidity, the more marked seasonal changes at a few individual stations are 
these :—For the two months the oscillations are respectively 0:120 and 0-067 
inch at Bombay ; 0°133 and 0-059 inch at Poonah; and 0°132 and 0:091 inch 
at Calcutta, thus showing the influence of the dry and wet seasons. At 
Madras, where the rain-bringing character of the monsoons is reversed, the 
numbers are 0°114 and 0:115, being nearly equal, and at Roorkee, in the 
* Some of the mean oscillations given in Karmrz’s “ Meteorology” have also been used in con- 
‘structing the charts, the figures, however, being treated merely as rough approximations, seeing that 
Kaemtz’s “mean oscillations” represent only the mean difference of the two maxima and two minima. 
Weight has also been given in drawing the lines to what appears to be, in a few cases, an undue 
slugsishness in the movements of the barometer, which the returns for Seftenberg, Bologna, Berne, 
Port Said, Ajmere, Jhansie, False Point, Nassau and Caraccas, seem to suggest. 
VOL. XXVII. PART III. 5.N 
