
ON THE REDUCTION OF METEOROLOGICAL OBSERVATIONS. 79 
In all the six curves enumerated, abstracting these and other more trifling 
inequalities, the rise is nearly at the same rate, and they form a group in de- 
cided accordance. Halifax and Edinburgh deviate much from their type, 
being nearly level for the first 9 or 10 hours of the 21st, and then suddenly 
and irregularly rising. The Markree curve also, from 0 hour to 4 hours, runs 
nearly level or with a slight descent, then rises by gentle successive swells 
(through 0°066 inch) to a slight maximum at 19 hours, gently dips to a feeble 
minimum at 21 hours, and then suddenly starts up with a bold rise for the 
remainder of the term, 
Kremsmiinster and Cadiz are exceptional to the general character, and for 
once they offer considerable agreement with each other. Both descend toa 
decided minimum between the 6th and 7th hour, rise to about their original 
level at midnight, dip to a slight minimum between 14 and 15 hours, and 
then rise again, the rise being sustained at Cadiz to the end, but at Krems- 
miinster only to 19 hours, when another fall commences. The total range 
-* Cadiz is 0°119, at Kremsmiinster 0°086, within the limits of the projected 
ours. 
Sept. 1838.—Markree, Halifax, Bristol Channel, Cambridge, London, 
Greenwich, Ghent, Alost, Brussels, Louvain, Cadiz. 
Markree stands in bold contrast with all the other curves of this term, It 
sweeps down over a range of 0°264 from a maximum at 0 hour toa minimum 
at 22 hours, with a very regular and free curve, while all the other curves, 
except Cadiz, rise with a gentle ascent. None of the diurnal movements are 
seen except the afternoon minimum, which is pretty conspicuous in most of 
them, and in some exaggerated into an extensive depression extending over 
the six or eight first hours of the afternoon (a feature, indeed, of no uncom- 
mon occurrence). At Brussels the forenoon maximum of the 22nd is also 
sensible. 
The curves for Ghent and Louvain are not continued through the night. 
So far as they go they preserve their parallelism with that of Brussels, and 
offer the same eminently smooth character. At Alost this character and par- 
allelism are again broken, precisely as in the June term, by an unexpected 
descent of the barometer during the three hours before midnight. All these 
eurves, as well as those for London, Greenwich and Cambridge, begin with 
the depression already noticed, subsequently to which they reascend during 
the rest of the series attaining a higher level, the total range in all being 
nearly alike (0°165 inch). The curve for Cambridge, however, is materially 
more irregular and fluctuating. 
The curves for the Bristol Channel and Halifax manifest the same gene- 
rally ascending character, the former throughout; the latter up to the 18th 
hour, after which it redescends. Both are smooth curves and their total range 
nearly alike, and somewhat less than in the cases of London, Brussels, &c., 
viz, 0°099. 
_ Cadiz is again exceptional. Its curve offers on the whole a slight descent, 
and a full and. somewhat violent development of the effects of diurnal oscil- 
lation, in both maxima and both minima; beyond this no features worth re- 
mark. In this term then the movements of the European atmosphere seem 
to have affected three distinct and independent systems, Cadiz and Markree 
being types of the two exterior, and the rest of the stations of the interior 
stem. 
ee 1838.—Halifax, London, Greenwich, Ghent, Alost, Brussels, Lou- 
vain, Cadiz. 
_ A generally descending, much undulated curve, for each station except 
Cadiz. The undulations, however, are rather numerous and small than ab- 

