54 'the force of

the lower cavity in the same state. Observe the heightof the water in the tube V W, measuring it from the sur-face & deducting the augmentation proceeding from theattraction of the glass. The effects of this attraction areknown previous to the experiment, by noting how much the■water contained in the tube, is higher than the surface &This experiment should be repeated several times in thefame manner, with equal quantities of powder of the samequality, in precisely the same state of the atmosphere, observedby a very accurate thermometer; and the elevation ofthe water in the tube at each time remarked. Neither therecipient F G L nor the cylinder E D E should be evertouched with the naked hand, but with a folded napkin, forthe contact even of a finger would in a short time excite adegree of heat sufficient to raise the water in the tube. Af-ter several repetitions, take the mean height; thus the va-rieties occasioned by the absorption of a part of the fluid, andthe heat remaining in the machine may be accounted for.Call this mean height -7, and the mean height of a baro-meter filled with the same kind of liquor as the glass vessela

A ; the fraction — will express the ratio between the elasti-A

city of the fluid produced from the powder, and that of theatmospheric air.

120. In small quantities of powder the proportion betweenthe saltpetre and the other ingredients may not be exactly thesame as in a larger mass ; which in a series of experimentsmay cause a considerable variation in the elevation of thewater in the tube. To remedy this inconvenience, grindsome saltpetre, sulphur, and charcoal separately ; weigh thesaltpetre and mix it with such a quantity of the other twoingredients as will most quickly consume it. The mixtureneed not be granulated, as the elasticity and density of thefluid are the only objects of research in this experiment j thephlogiston being only employed as the means of decomposingthe nitre.

121. The elasticity of the fluid, though very much di-lated in the two cavities is exactly proportionate to its den-

a

stty : thug the fraction — expressing the elasticity = n will alsoA

express the density. Now supposing the contents of the two

cavities