1816.] during the Year 1815. 15 
tinued without interruption for two years and six months. A 
description and engraving of the instrument employed is given in 
the Annals of Philosophy, vol. ii. p.96. Three observations were made 
every day; one about half past eight in the morning, another at 
noon, and the third in the evening about seven o’clock. The re- 
sults obviously deducible from these experiments are the following. 
(1.) The variation was least in the morning and greatest at noon. 
The mean variation at the three periods of observing for two years 
was as follows: 
SDOVAAMIGS sw feta oti i= ons > EM SPE i I apie bs 2 
< LETS py Rs DIO Re: tate oso, oA gel OF 
PENCRIB Es an cin'e ings 50 ahs eemeve LG sy ao 
So that at noon the declination was 7’ 15” greater than at half 
past eight in the morning ; and 5’ 49°5” greater than at seven in 
the evening. 
(2.) The law laid down by Mr. Canton, that the variation is 
greatest in summer and least in winter; and that it varies with the 
temperature, does not hold. The variation was greatest indeed 
during both years about the month of August ; but the next great- 
est variation to that occurred in the month of March. ‘The noon 
variation for 1813, placing the months in the order of its intensity, 
was as follows : 
August .... 24° 23’: 32” ApHh wri, est Diy het 
March.....:24 23 8 February .. 24 20 58 
Daly. fit) .. 2E) 128.0 4 November.. 24 29 54 
October ... 24 22 53 May .....: 24 20 54 
September . 24 22 32 December.., 24 20 30 
tes ist!) S422 17 January ... 24 19 3 
Their order for the year 1814 is as follows : 
April... ..,.24° 28’ 53” Maes ainn sn tr 22 13 
PUCUSE 5. ok. 2S. 48 February... 24 21 5) 
De denisir0in 24,20. 44 October ... 24 21 45 
Marech..... 24 23 40 November.. 24 20 $7 
September.. 24 23 17 December.. 24 20 36 
June...... 24 22 48 January ....24 20 12 
January always gives the least variation, and it may be con- 
sidered as the coldest month of the year; but the other months 
present anomalies which cannot well be ascribed to the temperature. 
Thus March, though usually a very cold month, exhibits a very 
high variation ; while February and May very nearly agree in the 
quantity of variation. 
(3.) The variation between two contiguous days often varies four 
or five minutes, and the needles sometimes vibrate 7’, or even 14’, 
without any apparent cause. 
(4.) A south west wind seems to increase the variation and the 
unsteadiness of the needles. 
{ am inclined to ascribe the diurnal variation to changes pro- 
