[ 554 1

The first of these mixtures with tin is that from whichMr. Scheffer endeavours to obtain the true gravity of pla-,tina, and from this it comes out 21,649. He observes 1that tin is not variable, as iron is, in its gravity, or quan-tity of phlogiston, so long as it preserves its metallic form;and hence concludes, that when platina and tin are meltedtogether, the excess of the specific weight of the mixtureabove that of the tin, must give the true specific weight ofthe platina. As the experiment on equal parts of tin andplatina, makes the gravity of platina, on this principle,above 21, he seems to think that all the mixtures, whosegravity was found such as to make platinas gravity lessthan this, must have been porous, and are therefore to bedisregarded in the present enquiry. He remarks, fromthe whole, that though the specific weights of fluids maybe determined accurately enough by hydrostatics experi-ments, we cannot be so certain about that of solids, onaccount of cavities, incompactness, and air bubbles ad-hering ; that the experiments on the foregoing mixturesafford a proof of this, mixtures of platina with one andthe same metal being sometimes heavier, and sometimeslighter, than they ought to be by calculation ; and thatthe same thing happens also in the pure unmixed metals,according as they are cast in a weaker or stronger heat.

The gravities of metals are doubtless influenced not alittle by circumstances cf this kind ; and it must be added,that in the mixtures with platina, there is another causeof variation, which has not yet been attended to. Whenplatina is melted with other metals in any considerableproportion, a part of the platina, unless the mixture iscooled hastily, is apt to separate before the fluid sets, sothat unless the whole mass be weighed in the hydrostaticbalance, which was not the case in some of the foregoingexperiments, we cannot be sure but the part weighed may

have