SOLIDS, LIQUIDS, AND GASES. 375 



In the first place, they are both fluids, as already agreed. 

 What, then, is the essential difference between liquid flu- 

 idity and gaseous fluidity? The expert in molecular math- 

 ematics, discoursing to his kinematical brethren, would 

 produce a tremendous reply to this question. He would 

 describe the oscillations, gyrations, collisions, mean free 

 paths, and mutual obstructions of atoms and molecules, 

 and, by the aid of a maddening array of symbols, arrive at 

 the conclusion that gases, unless restrained, expand of 

 their own accord, while liquids retain definite limits or di- 

 mensions. 



The matter-of-fact experimentalist demonstrates the 

 same by methods that are easily understood by anybody. I 

 shall, therefore, both for my own sake and my readers', 

 describe some of the latter. 



In the first place, we all see plainly that liquids have a 

 surface, i.e., a well-defined boundary, and also that gases, 

 unless enclosed, have not. But as this may be due to the 

 invisibility of the gas, we must question it further. The 

 air we breathe may be taken as a type of gases, as water 

 may of liquids. It has weight, as we may prove by weigh- 



squeezed by all that rests above it; thus the air around 

 us is constrained air. It is very compressible, and is ac- 

 cordingly compressed by the weight of all the air above it. 

 This being understood, let us take a bottle full of water 

 and another full of air, and carry them both to the summit 

 of Mont Blanc, or to a similar height in a balloon. We 

 shall then have left nearly half of the atmosphere below, 

 and thus both liquid and gas will be under little more than 

 half of the ordinary pressure. What will happen if we un- 

 cork them both? The liquid will still display its definite 

 surface, and remain in the bottle, but not so the gas. It 

 will overflow upwards, downwards, or sideways, no matter 

 how the bottle is held, and if we had tied an empty blad- 

 der over the neck before uncorking, we should fiud this 

 overflow or expansion pf the gas exactly proportionate to 

 the removal of pressure, provided the temperature re- 



