[ 547 ]

free from some os the causes of error which attend theother metals. It appears however from the foregoing ac-count, that of twelve mixtures of platina and gold, therewas not one so heavy as the gold itself, whereas onMr. Scheffers principle they ought all to have been heavier..It is plain therefore that either platina is not so heavy asgold, or that the principle of induction does not obtain inthe mixtures of gold and platina.

From the two last mixtures, the gravity of platina comesout between 18 and 19; but on these no dependence canbe had, the difference between the experimental and com-puted gravities being so inconsiderable, that it may be at-tributed to the unavoidable imperfections of the instru-ments used for the weighing; for an error of less than athirty-thoufandth part of the weight makes a difference of,012 in the specific gravity of the mixture, and of 1,000in that of the platina deduced from it. The case is thesame in the mixtures with other metals where the platinais in small proportion.

The other compositions give the platinas gravity lessthan 17; and as the platina is found by itself to be 17 ormore, it seems to follow, that there must necessarily be adiminution of gravity produced by the union of the twometals with one another. A phenomenon observed in thefusion appeared to confirm this. Most metallic bodies,made fluid by fire, shrink and assume a concave surface intheir return to solidity: pure gold shrinks perhaps rathermore than any of the others. But mixtures of gold andplatina, where the platina was in considerable proportion,were observed to shrink little; some of them even ex-panded and became convex. Of this expansion or dila-tation of volume, a decrease of specific gravity is the ne-cessary consequence.

B b b b

As