CAUSES OF FORMATION. 
207 
by volcanoes) ; while the other rests on the operation of 
small powers, which produce effects almost insensibly by 
progressive action. Those who love to indulge in geological 
hypotheses must not, however, forget the hori/.ontality so 
often remarked amidst gypseous and calcareous moun- 
tains, in the position of grottoes communicating with eacli 
other by passages. This almost perfect horizontality, this 
gentle and uniform slope, appears to he the result of a 
long abode of the waters, which enlarge by erosion clefts 
already existing, and carry oil - the softer parts the more 
easily, as clay or muriate of soda is found mixed with the 
gypsum and fetid limestone. These cfiects are the same, 
whether the caverns form one long and continued range, or 
several of these ranges lie one over another, as happens 
almost exclusively in gypseous mountains. 
That which in sheliy or Neptuneau rocks is caused by 
the action of the waters, appears sometimes to he in the 
volcanic rocks the effect of gaseous emanations* acting in 
the direction where they find the least resistance. When 
melted matter moves on a very gentle slope, the great axis 
of the 'cavity formed by the elastic fluids is nearly horizontal, 
or parallel to the plane on which the movement of transition 
takes place. A similar disengagement of vapours, joined 
to the elastic force of the gases, which penetrate strata soft- 
ened and raised up, appears sometimes to have given great 
extent to the caverns found in trachytes or trappean por- 
phyries. These porphyritie caverns, in the Cordilleras of 
Quito and Peru, bear the Indian name of Mcichays . f They 
are in general of little depth. They are lined with sulphur, 
and differ by the enormous size of their openings from those 
observed in volcanic tufas J in Italy, at Tenerife, and in 
* At Vesuvius, the Duke de la Torre showed me, in 1805, in currents 
of recent lava, cavities extending in the direction of the current, six or 
seven feet long and three feet high. These little volcanic caverns were 
lined with specular iron, which cannot be called oligistc iron, since M. 
Gay- (.tissue’s last experiments on the oxides of iron. 
1* Macbay is a word of the Quichua language, commonly called by 
the Spaniards ‘ the Incas' language.' Callancamachay means ‘‘‘a cavern 
as large as a house,” a cavern that serves as a tambo or earavansarai. 
I Sometimes fire acts like water in carrying oil’ masses, and thus the 
cavities may be caused by an igneous, though more freouently by au 
aeneous erosion or solution. 
