[ 259 ]

with so much water, as made them just soft enough toadmit of being formed into long narrow plates : A par-ticular account of these experiments will be given here-after ; at present it is sufficient to observe, that the plateof pure clay shortened in drying one part in eighteen,while a mixture of the clay with twice its weight of sand,shortened but one part in thirty. It is not known that anykind of earth shrinks so much as clay, and hence the purityof clay may be judged from the degree of its contraction.

Plaster-of-paris on the contrary, diluted with water intothe consistence of a soft or thin paste, quickly sets orgrows firm, and at the instant of its setting has its bulkincreased, as appears from the pretty experiment men-tioned somewhere in Boyle’s writings, and which I haveoften tried : A glass vessel being filled with the fluidmixture and closely stopt, the glass bursts while the mix-ture sets, and sometimes a quantity of water issues throughthe cracks. This expansion of plaster-of-paris, in passingfrom a soft to a firm state, is one of its valuable proper-ties ; rendering it an excellent matter for filling cavitiesin sundry works, where other earthy mixtures wouldshrink and leave vacuities, or entirely separate from theadjoining parts. It is probable also that this expansionof the plaster might be made to contribute not a little tothe elegance of the impressions which it receives frommedals, Lee. by properly confining the soft matter, thatits expansion may force it into the minutest traces of thefigure ; the expansion of the matter doing the same officeas the pressure by which wax is forced into the cavitiesof a seal.

There are grounds to believe, that differences of thefame kind obtain in melted metals at the instant of their fix-ing or becoming solid ; that at this period they do not ob-serve the same laws as before or after it; and that, while

L1 2 some