of fufpending Magnetic al Needles. 541 
magnet did not move fo freely. The firft difficulty could 
be eafily diminifhed by letting the working needle pafs 
through a fmall and fmooth hole in a gial's or metallic 
plate, to controul its vibrating motion, and by adapting to 
it a fhoulder, by which it fhould flop upon the glafs or 
metallic plate, and from which it fhould be taken up 
again, by Hiding the vertical magnet a little down fo as 
to touch the point of the working needle, and to lift it 
up again into its former place. The fecond difficulty 
could, I think, alfo be in a great meafure obviated by em- 
ploying, inftead of a common magnetical needle, a thin 
if eel tube (of which I will fpeak by and by.) which is only 
a little fpecifically heavier than an equal bulkof the liquid; 
By this means a fmall vertical magnet could be made to 
fupport a large one, even of a coniiderable abfolute 
weight, of which weight, however, the vertical magnet 
fhould fupport almoft nothing, it being counter-balanced 
by the fluid in which it js immerfed. 
In the experiment, as I had made it, I could not per- 
ceive, that the horizontal magnetical needle was influ- 
enced by the vertical magnet, which would probably be 
the cafe if the vertical one was much out of the vertical 
line, and too near the horizontal needle. 
exp. v. I adapted to a flat ftrong magnetical needle 
two caps, turned one againft another, fo that the needle 
Vol, LXIX. 4 B could 
