[ 544- J

will be exactly equal in weight to the ten, how unequalsoever the arms of the balance may be ; and consequently,that any unequal-armed balance may, on this principle,have a set of weights adjusted to it, which being used al-ways in one scale, the instrument shall be of the same ac-curacy as if the arms were most exactly equal. The bestway of procuring equal small weights is, by cutting offequal lengths of the finest silver wire : the silver thread,kept equally stretched by a heavy body at the end, maybe coiled close round a thicker piece of brass wire, and allthe coils cut through at once by a sharp instrument appliedlengthwise. Silver wire is drawn to such fineness, and ofso uniform thickness, that weights, thus made by mensu-ration, are of greater nicety than it is possible for any ba-lance to weigh. A piece of the wire,whole length is verysensible and much further divisible, shall not have weightenough to give any sensible motion to the tenderest balance.These small pieces, or such as will but just move the ba-lance when empty, and which consequently will not moveit at all when loaded, I have found to be a very useful ap-pendage to the adjusted series of weights. Though abalance appears exactly in equilibrio, yet one side mayreally preponderate, by any quantity of force, less than thatwhich is sufficient to overcome the friction on the center:as less additional force will serve to depress this side thanthe other, one of the small weights, tried first in one andthen in the other scale, will enable us to judge whetherthe equipoise is exact, or on which side the preponderationlies.

The results of these experiments were published in thePhilosophical Transactions, together with the gravities ofthe several mixtures deduced from calculation ; from.which it appeared that the experimental gravities were al-most always , less than the CQmptited. b&ut anseftPt in.

those