[ 399 ] 
the year 1750, Mr. Wargentin, fecretary of the Royal 
Academy of Sciences in Sweden, took notice, both 
of the regular diurnal variation of the needle, and alfo 
of its being difturbed at the time of an aurora borea- 
lis, as recorded in the 47th volume of the Pbilofo- 
phical Tranfadtions ; but is filent as to the caufe. I 
had n*o opportunity of making obfervations of this fort 
myfelf, till the latter end of the year 17-76; but, 
lince that time, I have made near four thoufand, 
with an excellent variation-compafs, of about nine 
inches in diameter. The number of days, on which 
thefe obfervations were taken, is 603 ; and the diur- 
nal variation on 574 of them was regular ; that is, 
the abfolute variation of the needle weftward, was in- 
creafing from about eight or nine o’clock in the morn- 
ing, till about one or two in the afternoon, when the 
needle became flationary for fome time; after that, 
the abfolute variation weftward was decreafing , and 
the needle came back again to its former fituation, or 
near it, in the night, or by the next morning. The 
diurnal variation is irregular , when the needle moves- 
llowly eaftward in the latter part of the morning, or 
weftward in the latter part of the afternoon ; alfo 
when it moves much either way after night, or fud- 
denly both ways within a fhort time. Thefe irregu- 
larities feldom happen more than once or twice in a 
month, and are always accompanied (fo far as I have 
been able to obferve) with an aurora borealis. Thus 
having explained what I mean by the regular and ir- 
regular diurnal variation, and fhewn, that this varia- 
tion is generally regular; I fhall now, in the firft 
place, endeavour to account experimentally for the 
regular variation ; then offer a conjecture concerning 
