348 ME. E. SPrUCE OF INSECT-MIGEATIONS IN SOUTH AMEEICA. 



thinly-set trees and bushes, with scarcely any lianas — the palms 

 few, but peculiar, and often odd-looking, — on a near view by the 

 greater abundance of Ferns, especially on the trees, and sometimes 

 of terrestrial Aroids and Cyclanths ; the Recent Forests by their 

 low irregular tangled growth, paler foliage, and general weedy as- 

 pect ; the Eiparial Forests, even where the water is not^ visible, 

 by the varied tints of the foliage, and by the trees rarely equalling 

 • those of the Virgin Forest in height — sometimes, indeed, begin- 

 ning on the water's edge as low bushes, thence gradually growing 

 higher as they advance inland, until at the limit of inundations 

 they mingle with the primeval woods, and are almost equally lofty, — 

 by the greater proportion of herbaceous lianas which drape the 

 trees and often form a curtaiu-Iike frontage, — and by the abund- 

 ance of Palms, whereof the taller kinds usually surpass the exoge- 

 nous trees in height, and (the Fan-Palms especially) often stretch 

 in long avenue-like lines along, or parallel to, the shore. On some 

 black-water rivers, such as the Pacimoni, the Atabapo, and the 

 Eio Negro in some parts of its course, the breadth of inundated 

 land is entirely clad with bushes and arbuscules of very equable 

 height, on the skirts of which the Virgin Fores b rises abruptly to 

 a height more than twice as great. This is called by the natives 

 " Caatinga-gapo." 



Besides these differences of aspect, the natives will tell you there 

 are other more intrinsic ones, — for instance, that the riparial 

 trees have softer and more perishable timber, as well as inferior 

 fruits ; while the Caatingas, with a far greater show of blossom, 

 have hardly any edible fruit at all, and very few indeed of the 

 trees rise to the magnitude of timber-trees. And yet, when the 

 constituent plants of the different classes of forest come to be 

 compared together, they are found to correspond to a degree quite 

 unexpected ; for, although the species are almost entirely diverse, 

 the differences are rarely more than specific. It is only in the 

 Caatingas that a few genera, each including several species, seem 

 to have taken up their exclusive abode : such are Commianthus 

 among Eubiaceae, Pagamea among Loganiacese, Myrmidone and 

 Majeta among Melastomacese ; and there are a few other peculiar 

 genera, chiefly monotypic. But, of the riparial plants, nearly 

 every species has its congener on terra firma, to which it stands 

 so near, that, although the two must of right bear different names, 

 the differences of structure are precisely such as might have been 

 brought about by long exposure even to the existing state of 

 things, without supposing them to date from widely different con- 



