VALLEY OF THE PARAHYBA. 319 



A general survey of the South American flora 

 shows, along with elements derived from distant 

 regions, a large number of types either absolutely 

 peculiar to that continent, or which, in some cases, 

 appear to have spread from that centre to other areas. 

 Of these peculiar types some may probably have 

 originated in the Andean chain, but as to the majority, 

 it seems far more probable that their primitive home 

 was in Brazil ; and it is precisely on the ancient 

 mountains of this region that I should look for the 

 ancestors of many forms of vegetation which have 

 stamped their character on the vegetation of the con- 

 tinent. 



I should be the first to admit that the views here 

 expressed have no claim to rank as more than pro- 

 bable conjectures ; but I hold that these, when resting 

 on some positive basis of facts, are often serviceable 

 to the progress of science, by stimulating inquiry and 

 leading observers to co-ordinate facts whose connec- 

 tion had not previously been apparent. 



In following the valley, in places where the siliceous 

 soil supported only a scanty vegetation, I was struck 

 by the singular appearance of scattered piles, usually 

 about four feet in height, having much the appearance 

 of rude milestones, occurring here and there in some 

 abundance, but never very near each other. I was 

 often able to avail myself of the short halts of the 

 train at wayside stations to- secure specimens of in- 

 teresting plants, but I was not able to approach near 

 to these unknown objects. I have no doubt, how- 

 ever, that they were habitations of termites, or, as they 

 are commonly called, white ants. I have never been 



