322 
Agricultural Meteorology. 
London gave a clifierence of not less than 51 "5 on one tlay in 
June from 9 a.m. to 6 p.m., amounting indeed at 2 p.ini. to as 
much as 67°"5. 
Captain Parry remarked at Melville Island, N. lat. 74^ 25', 
on the 16lh of March, during more than three hours, a difference 
of 50° Fahr. between a thermometer on the shady and one on 
the sunny side of the ship — the former being 22° below zero, the 
latter 28° above. Captain Scoresby observes, in his ' Account of 
the Arctic Regions,' " that the sun's rays frequently melt the 
pitch out of the seams of the ship on one side, while ice is 
forming on the other side, in the shade." Mr. Daniell ascer- 
tained, by covering the bulb of a thermometer with pitch, and 
exposing it to a gentle heat, that it required at least 120° to melt 
it. Here then the solar radiation could not have been less than 
90° Fahr. or 50° of Centigrade. 
These and other instances which might be adduced serve to 
show that M. Gasparin's proposition in its present form is inap- 
plicable,* and that it is not practicable to determine the aptitude 
* This theory, which Gasparin adopts from Boussingault, originated 
with Reaumur, and was subsequently followed out by the Pere Cotte (see 
'Traito de Meteorologie du P. C.,' p. 424, and the Mem. de I'Acad. 
des Sciences for 1735, p. 5r)9). Quetelet, however, considers that in esti- 
mating the action of heat on plants, the more correct metliod is to take 
the sum of the squares of the temperatures — since temperature is with 
reference to plants a vital force. In the notation of the calculus of finite 
differences, Quetelet's method is expressed by 2 tl. 
Reaumurs by S t„. 
(the respective diurnal temperatures being supposed equal to t, t\ f , Sec, 
while n represents the number of days reckoned from the first observed 
flow of the sap after the winter frosts, to the first expansion of the petals.) 
Judging from Quetelet's account (pp. 243, 244, in the first edition of 
his ' Lettres sur la Theorie des Probabilites,), of his observations upon the 
lilac, and his compared calculations according to the two methods re- 
spectively, that of the smn of iJie sqtinres accords the most nearly with 
facts, and at any rate reduces the deviations within the limits of probable 
error (mathematically speaking). It is possible that M. Gaspavin"s data 
would lead to more consistent results if they were treated according to 
Quetelet's theory, and it would be interesting if M. Gasparin could be in- 
duced to revise them with this view. Quetelet's formula ^t'i, when 
analyzed, leads to a remarkable result stated in a note of his appendix 
to the above work. If T be the mean of the diurnal temperatures ob- 
served during ?i days, and A, A', &c,, the respective daily differences 
between the observed temperature and the mean T, then by a very simple 
algebraical development we deduce 
Thus 1 tl is obviously a function of T and of A, and is of such a nature 
that (lor any specified value of T) the first side of the equation will be 
greater the greater A i.s, and will he least when A is zero. That is to say, 
that departures from the mean are more favourable for developing vpgelation 
than uniform temperatiues : a most interesting deduction, should it be veri- 
fied by facts ; and it appears that Humboldt considers this corollary to be 
