ATMOSPHERE IN RELATION TO HUMAN LIFE AND HEALTH. 219 



registered. The effect of a strong horizontal wind on a u chimney" of 

 hot vaporous air would be to increase greatly the force of the upward 

 torrent, as has been proved by anemoinetric experience with tall chimney 

 shafts and domestic fires. The effect of the violent wiud is exceedingly 

 destructive, especially when the tornado is of small diameter. Some 

 towns in the United States are particularly subject to these storms, 

 and, as they generally come from one direction, the effect of building a 

 perpendicular wall of 200 or 300 feet high on that quarter near the town, 

 in order to break or divert its course, would seem worth tryiug. 



Eeturning to the more normal conditions of the atmosphere, we may 

 imagine the vapor, whether from land or sea, to have mixed much but 

 not uniformly with the overlying air. The differences in the humidity 

 of different masses or parcels of air, and the viscosity, friction, or 

 resistance of the lower strata, where the pressure is 15 pounds to the 

 square inch, prevent the interaction from being continuous and uni- 

 form, and consequently the ascending currents are local and variable, 

 but when once fairly started, generally persist for a considerable time, 

 moving all the while with the prevailing wind. When the vapor 

 streams reach a certain height, they begin to condense, first and chiefly 

 because they expand, and in expanding cool themselves, according to 

 the laws of heat, and, secondly, because they mix with cooler strata. 

 If the vapor be supposed to have ascended to a height of 3,000 feet, 

 the pressure upon it has diminished from about 30 to 27 inches of 

 mercury, or by about one-tenth, so that it swells, allowing for contrac- 

 tion by cold, to a bulk nearly one-tenth more than it had at the sea 

 level. This is sufficient to produce a large diminution of temperature 

 and the molecules vibrate so much less rapidly that some of them cease 

 to maintain the condition of vapor. The vapor must condense, accord- 

 ing to recent discoveries, not in contact with mere air, but upon very 

 minute solid particles, motes, or dust, which may consist of ultramicro- 

 scopic sand, sea salt, or other material. So a cloud takes form. For 

 each amount of curvature of a liquid surface there is a definite vapor 

 pressure, and the pressure necessary for precipitation is greater as the 

 surface becomes more convex, so that precipitation takes place more 

 easily the larger the water globule in the presence of vapor. And so 

 great is the pressure required for the condensation of vapor in free air 

 that condensation can not take place except upon those small nuclei of 

 dust which, more or less, are present throughout the lower atmosphere. 

 Solid surfaces exposed to gases contract a film of gas upon their sur- 

 faces. Now, the dust of the air, owing to its minuteness, presents an 

 enormous surface, and is moreover largely hygroscopic, so that the 

 tendency to gather a film of vapor of water upon its surface becomes 

 very important and effective. Without this fine dust in the air the 

 world would hardly be tolerable or even habitable by the human race. 

 The vapor would condense, not in the sky and in the form of clouds, 

 but on the earth, on mountains, trees, houses, and clothes, so that the 



