'9I3-] OF GRANITES IN THE TROPICS. 171 



mation of these caldrons. In my opinion detlation may be important 

 in the early stages of excavation, when the depth does not prevent 

 the wind from blowing away small loose fragments of the minerals, 

 but when a pit is a meter or more in depth and a meter in width it 

 seems highly improbable that the wind is a factor of any importance 

 in the direct removal of materials, though it may be indirectly 

 important enough in the bringing of plant seeds or spores. 



The things that seem to me competent to account for the re- 

 moval of the disintegrated minerals are suggested by the dark lines 

 that run down the rocks from the notches in the lips of all of the 

 caldrons. These dark streaks are made by overflow^ing water. 



When water stands for some time in the pits it becomes covered 

 by low forms of plant life, apparently freshwater algae for the most 

 part. When in time the water dries up this material is left spread 

 over the bottom where it packs together, a sticky fibrous mass cling- 

 ing to and closely enwrapping whatever lies in the bottom of the pit. 

 In time the rains come again and the pit is filled to overflowing, the 

 matted masses of algae are floated to the top and carried over the 

 edge of the caldron, taking with them some of the mineral fragments 

 from the bottom of the pit. The rapid growth of vegetation in such 

 waters is greatly facilitated by the hot climate, while the hot dry 

 air and the high winds quickly dry up the waters when once the 

 pits are filled by the rains. The peculiar concentration of the rain- 

 fall at Quixada likewise hastens this natural process of excavation. 



Still more important in most instances is the stirring of the mate- 

 rials in the bottoms of the caldrons by the inflowing water. The mud 

 in the bottom being thus stirred by the water pouring in, a great deal 

 of the finer material is carried out when the pit is full and overflows 

 its lower lip. It must not be supposed, however, that the water flow- 

 ing into the caldrons is of sufficient volume to whirl the coarse mate- 

 rials and thus grind out the rock. This is far from being the case. 

 I found no evidence of any such mechanical wear on the insides of 

 the pits. They are all uniformly rough within. 



When the pits are shallow, however, and are so far down the 

 slopes that the water enters them with a rush it is readily imagined 

 that all or most of the finer materials must be swept over the rim. 

 Some of the forms suggest that old and deep caldrons have from 



