553 



the southern bank of the Oroonoko. This hill 

 is scarcely elevated one hundred toises above 

 the neighbouring plains. If the assertions of 

 the natives be true, it is probable, that some 

 subterraneous cause exists in Duida and Gua- 

 raco, that produces these flames ; for they never 

 appear in the lofty neighbouring mountains of 

 Jao and Maraguaca, so often wrapped in elec- 

 tric storms. The granite of the Cerro Duida is 

 full of veins, partly open, and partly filled with 

 crystals of quartz and pyrites. Gaseous and 

 inflammable emanations, either of hydrogen, or 

 of naphtha, may pass through these veins. Of 

 this the mountains of Caramania, of Hindoo-kho, 

 and of Himalaya, furnish frequent examples. 

 We saw the appearance of flames in many parts 

 of eastern America subject to earthquakes, 

 even from secondary rocks, as at Cuchivero 

 near Cumanacoa*. The fire shows itself when 

 the ground, strongly heated by the sun, re- 

 ceives the first rains; or when, after violent 

 showers, the earth begins to dry. The first 

 cause of these igneous phenomena is at immense 

 . depths below the secondary rocks, in the primi- 

 tive formations : the rains, and the decomposi- 

 tion of atmospheric water, act only a secondary 

 part. The hottest springs of the Globe issue 

 immediately from granite-^. Petroleum gushes 



* See vol. iii. chap. 6, p. 82. 

 t See vol. iv, chap. 16, pp. 171 and 195. 



