io A4K Cavallo's MagnaUcal 
partly owing to moifture, or to very fine duft that adheres to 
the quickfilver when expofed to the atmoiphere. 
In performing fuch experiments care fhould be had to keep 
the air and the quickfilver as much undifturbed as poflible, and 
alfo to prefent the magnet in a proper manner, viz. fo as not to 
touch the fu rface of the mercury, nor to ftrike again ft the 
floating body, efpecially when that is in motion ; for after 
the impulfe, though that may be very flight, the floating 
body will be impelled backwards, which may be often miftaketi 
for magnetic repulfion. The leaft exceptionable method is the 
following : firft, if the floating body be in motion let it flop, 
then hold a ftrong artificial magnet nearly in a perpendicular 
direction, and with one pole juft over one fide of the floating 
body, or rather fo that the perpendicular, let fall from the 
pole of the magnet to the furfiice of the quickfilver, may be 
about one- tenth of an inch diftant from the body to be tried. 
The height of the magnet above the quickfilver fhould be juft 
fufficient to let the floating body pafs under it without touching 
it. In this fituation the magnet muft be held fteady ; and if the 
floating body has any magnetifm, it will be foon drawn diredlly 
under the magnet. 
In thefe experiments it will be generally found, that one 
part of the floating body is more magnetic than the reft, which 
appears from that particular part being conftantly drawn 
dire&ly under the pole of the magnet ; whereas, when the 
magnetifm is diftufed equably, the center of gravity of the 
body (provided its fhape be not very irregular,' becomes fta- 
tionary juft under the pole of the magnet. 
It is not every magnet that will difcover the very weak mag- 
netifm of certain fubftances ; for fometimes a powerful mag- 
