252 PROGRESS OF SCIENCE IN THE CENTURY. 



rocks to be chemical precipitates in water; and 

 Werner went the length of maintaining that volcanic 

 action was altogether a modern phenomenon. 



There was more progress in the work of Alexander 

 von Humboldt (published 1808-1823) who took a 

 world-wide survey of volcanoes, and concluded from 

 their distribution that they could not be due to 

 merely local causes (like coal-pits on fire), but must 

 be interpreted in reference to the state of the earth's 

 interior and clefts in the overlying crust. Hum- 

 boldt's position was strengthened by the work of his 

 friend Leopold von Buch, who began as a ISTeptun- 

 ist, but was soon led by observation in many coun- 

 tries to sounder views. Relying, like Hutton, on 

 the expansive power of the internal heat of the earth, 

 he made a point of distinguishing from true vol- 

 canoes what he called " craters of elevation." 

 These he supposed to be due to huge blister-like ele- 

 vations of the strata of the crust, which eventu- 

 ally collapsed, though without actual volcanic erup- 

 tion. 



In 1825, George Poulett-Scrope published the 

 first edition of his classic book on volcanoes, in which 

 he gave a careful description of the physical facts, 

 and sought to explain volcanic action both past and 

 present on a simple hypothesis. Supposing that 

 subterranean rock-masses were saturated with water, 

 and that this became heated from the interior, the 

 expansive force of the steam would account for erup- 

 tions. Like Lyell (1830), he entirely opposed von 

 Buch's theory of " craters of elevation " as con- 

 trasted with eruptive volcanoes. 



For many years a healthy conflict of opinions con- 

 tinued between supporters of von Buch, such as 

 Daubeny, Elie de Beaumont, and Dufrenoy, and 



