ON THE REDUCTION OF METEOROLOGICAL OBSERVATIONS. 91] 
and 4 hours, the almost complete barometric development of the storm. I 
have accordingly projected the two curves on an additional sheet scale 1+ of 
the larger divisions =*i. The immense and rapid fall at Quebec, at the 
middle of the storm, as compared with the phenomena presented at New 
York at the same time, is interesting; also the shoulders at New York at 
10 a.m. of the 20th, and at 10 a.m. of the 22nd, exhibiting a wave at the 
circumference of the funnel. 
(B.) In comparing the ranges, I have taken the range at New York from 
the shoulders only, as the observations at Quebec do not include them. 
March 1837.—Boston, Flushing, Albany, Montreal, Bahama, Magnetic 
Island. 
(H.) The observations at Magnetic Island break off at midnight of the 
Q1st. All the other curves, Bahama included, descend, but with different 
rapidities and degrees of irregularity. 
(H.) The Boston curve is very smooth (hourly observations). It con- 
tinues nearly level till 11 p.m., then dips gracefully to a minimum at 25 hours 
(1 hour, March 22nd). Its range within the 26 projected hours is 0°192. 
(H.) The Flushing and Albany curves are much and abruptly broken, 
but exhibit no correspondence in their zigzags. Both attain minima at the 
25th hour. ‘Their ranges, as above, are 0°387 and 0:225. 
(H.) The Montreal curve exhibits the diurnal oscillations superposed on 
a regular line of descending slope, the evening maximums being somewhat 
broken. In other respects the curve is very smooth. The general slope is 
rectilinear and less steep than in the others, viz. at the average rate of 0°145 
in 24 hours. 
(H.) The Bahama curve is still less steep, sloping at the average rate of 
0°13 in 24 hours. Its diurnal fluctuations, though very visible, are not nor- 
mally developed. 
(B.) Among the curves on this sheet that were obtained from the eastern 
ee of the United States, there is a general similarity. It is probable they 
ormed part of a system of oscillation having a point of greatest oscillation. 
The curve obtained at Flushing exhibits the greatest oscillation, but as there 
are no observations south or south-west of this station it cannot be considered 
as the point of greatest oscillation. Observations from the south-east and 
south on the Atlantic would have been very interesting at this term. 
June 1837.—Quebec, Montreal, Burlington, Gardiner, Boston, Middletown, 
New Haven, Wiilliam’s College, Flushing, Natchez, Washington (incom- 
plete). 
ait) The curves of this term may be classed in two systems, the first con- 
sisting of those of Quebec, Montreal and Burlington; the second, of all the 
rest. The character of the first class is a general tendency to ascent, inter- 
rupted somewhat irregularly by the diurnal phases, in which the morning 
minimum is anticipated by two or three hours, a particular, however, which 
cannot be traced at Quebec, the observations having been discontinued from 
midnight till 6 a.m. 
The other stations may be divided into subclasses, graduating into one 
another; those which exhibit a strong maximum in the earlier and a strong 
minimum in the later hours of the projected series (Gardiner and Boston) ; 
those in which this maximum merges in the maximum preceding the noon of 
June 21, and is separated from the strong characteristic minimum by 8 or 10 
hours of level, the said minimum itself beginning at an earlier epoch (at about 
2 or 3 a.m., as Middletown, New Haven, William’s College); and lastly, those 
in which no such characteristic maximum can at all be traced, and in which 
