400 ACTION OF WATER. [PART II. 



observed several kinds of Aralia, a very curious 

 form of plants. From the top of a very slender 

 stem the long narrow leaves spread in a circle, 

 hanging down in the shape of an umbrella. 



. The first part of the road was flat and very good, 

 and had been cut by the natives for the convenience 

 of the missionaries ; but it soon became a narrow 

 path, across which the luxuriant vegetation of creep- 

 ing plants often produced a perfect net-work. The 

 soil was pumiceous gravel, richly mixed with vege- 

 table mould. Afterwards the land became more 

 hilly, and its configuration and external outlines 

 were often extremely picturesque : the tufaceous 

 and pumiceous formation assumed castle-like forms, 

 and the rivulets had cut out deep channels in the 

 soft rock, leaving high banks. Notwithstanding 

 the softness of the material, it must have required 

 a great length of time to hollow out these rivulets, 

 which I am inclined to ascribe entirely, or almost 

 entirely, to the action of running water. Some 

 spots were devoid of trees, and were covered with 

 high fern (Pteris esculenta) ; they bore evident 

 marks of having been cleared of the forest by fire. 

 It is remarkable that this fern universally appears 

 wherever a spot has been thus cleared, even in the 

 middle of the primitive forest, where it never grows 

 so long as the trees are standing. I have no doubt 

 that the highly farinaceous root of this very com- 

 mon plant, which was formerly the principal food 



