Why the definition is not satisfactory, is that it does not tell us what there is in the spaces between the molecules of what is called the gas. If there is room for them to move in every direction there must be spaces between them, and these spaces must either be absolutely empty, or filled with something. If they are supposed to be empty, then the molecules being actually small pellets, like diminutive marbles, or snipe-shot, we immediately begin to think why gravitation does not make them, being ponderable bodies, fall down to the bottom of the bulb; and seeing that, by the definition, they are evidently considered not to do so, we think of what can keep them from falling, and of how they can be pumped out of a bulb or any sort of vessel. If we fill a vessel with marbles, snipe-shot, wheat flour, or dust, and set a pump to work on it, we shall find that we make very little progress in pumping them out of it. At first we might extract a puff or two of flour or dust--marbles or snipe-shot by no means--carried into the pump by any air there might be mixed with them, but that would very soon come to an end; besides, there would be air, gas, something, in the interstices--if any--of the flour or dust to drive them into the pump when a vacuum was formed in it, and the puffs would cease when the air, which would be in exceedingly small quantity, was all extracted. But independently of all this, we have supposed the spaces between the gas pellets or molecules to be absolutely empty, and there would be nothing to push them into the vacuum created in the pump. There is no possibility of pumping marbles, sand, flour, or dust out of a vessel without the assistance of a fluid agent of some kind, water, gas, or air; and even then it would be done with much difficulty.