SUPPLEMENTARY REPORT ON METEOROLOGY. 53 
Enormous as both these ranges are, we cannot but observe that 
the Asiatic extreme warmth does not approach within 30° of 
the American one! and yet the mean temperature of both sta- 
tions is nearly alike. 
41. The determination of the diurnal curve of temperature 
and its variations may be almost called complete, in comparison 
with that of the annual curve, which, notwithstanding, is a 
matter in some respects of still more important concern. Even 
the mean temperature of a place, at any period, is very difficult 
to fix, owing to the great discrepancies of successive seasons. 
Thus 29 years, from 1806 to 1834, during which comparable 
observations have been statedly made at the Observatory at 
Paris, give a mean of 10822 cent., with a variation between 
one year and another from 12%10c. in 1822, to 9°35c. in 1829, 
the difference being almost 5° of Fahrenheit*. It is evident 
that many years must be required to fix even this comparatively 
simple datum; how much more to determine the form of the 
annual curve which, in every region, is doubtless characterized 
by peculiar inflections ! 
42. The study of the annual curve is very much retarded by 
the want of simple, permanent, and comparable meteorological 
observations. These, which require little expense and little time, 
pre-suppose, from the extent of years over which they must 
extend, a kind of co-operation to which too little attention has 
been directed. The Paris observations are perhaps the only 
ones to which we can look with any degree of confidence in 
this respect, even for the comparatively short period of thirty 
years. Professor Brandes, of Breslau, has, with praiseworthy 
industry, collected and reduced a vast number of observations, 
continued at various stations for eight, ten, or more years, and 
has compared them and projected the annual curves, which he 
obtains by finding the mean temperature of successive periods of 
fivedayst. The chief stations are Petersburg, Stockholm, Cux- 
haven, Zwanenburg, London, Mannheim, Vienna, St. Gotthard, 
La Rochelle, and Rome. In these curves, imperfect as they are, 
are well shown the peculiarities of insular and continental 
climates, and those of plains and mountains. Some points of 
agreement may probably be found in all, for which sufficient 
reasons may be assigned; such as the more rapid increase of 
temperature than its decline, which applies equally to the diur- 
nal curve. This is, I conceive, chiefly owing to two causes, 
* Poisson, Théorie de la Chaleur, p. 463. 
+ Beitrage zur Witterungskunde, von H. W. Brandes. 8vo. 1820. Unter- 
metuungen uber den mittleren Gang der Warme-Aenderungen durchs ganze 
ahr. 
