HUMPHREYS: PRODUCTION OF CLIMATIC CHANGES 367 



The idea then that volcanic dust may be an important factor 

 in the production of climatic changes is not new, but hitherto 

 the idea has not been supported either by a clear explanation 

 of the process by which such dust can lo-v/er average tempera- 

 tures or by a discussion of all the available observational data. 

 In fact the rational or theoretical discussion has not long been 

 possible since, in almost every particular, it depends upon the 

 results of recent investigations. 



We shall now briefly consider, under separate heads, the sev- 

 eral points essential to an understanding of how volcanic dust may 

 influence, and of the extent to which it actually has influenced, 

 our average temperatures — by what process it may modify and 

 to what extent it actually has modified our climates. 



Atmospheric regions. At an elevation that in middle latitudes 

 averages about 11 kilometers the temperature of the atmosphere 

 becomes substantially constant, or, in general, ceases appreciably 

 to decrease with increase of elevation, this is, therefore, the upper 

 limit of distinct vertical convection and of cloud formation. 

 Hence, while volcanic or other dust in the lower or cloud region 

 of the atmosphere is quickly washed out by snow or rain, that 

 which by any process happens to get into the upper or isother- 

 mal region must continue to drift there until gravity can bring 

 it down to the level of passing storms. In other words, while 

 the lower atmosphere is quickly cleared of any given supply of 

 dust, the isothermal region retains such dust as it may have 

 for a time that depends upon the size and density of the indi- 

 vidual dust particles themselves, or upon the rate of fall. 



Size of volcanic dust particles. From the angular dimensions of 

 a reddish brown corona, known as Bishop's ring, seen around 

 the sun after the eruptions of Krakatoa, Pele and certain other 

 volcanoes, it has been calculated, by the aid of the laws of dif- 

 fraction, that the average diameter of the particles to which this 

 ring was due, assuming them spherical, was about 1.85 microns. 

 Hence, with this information, it becomes possible to calculate 

 the time of fall of volcanic dust. 



Tiine of fall. By using the excellent measurements recently 

 made by McKeehan^ on the terminal velocity of falling globules, 



■^ Phys. Rev., 33: 153. 1911. 



