OBSERVED DURING THE RORAIMA EXPEDITION. 253 
certain grasses, dwarf shrubs, and herb-like plants, form the dominant vegetation. Yet 
a few remaining parts are marked by the occurrence of certain distinct and, so to speak, 
localized species, which are scattered more or less widely among the more ordinary 
forms. Again, a very few other parts are still more distinctly marked, and made very 
distinct areas, by the more or less complete absence of the ordinary forms, and the 
substitution there of an entirely new and generally very distinct set of species. These 
areas with a few localized species, several of which were passed by us on our way to 
Roraima, and still more these areas of distinct vegetation, of which the Kaieteur 
savannah which we traversed, and especially Roraima itself, are remarkably fine 
examples of the utmost botanical interest. 
A few notes must first be given of the species here described as localized. It is to be 
remembered that these notes were made during a single walk, long as it was, through a 
country otherwise almost absolutely unknown ; so that though these species were noticed 
by me because I saw them either only in one spot, or at least in very few spots—i. е. I 
passed through either only one distinct group or through very few such groups of them— 
yet it is, of course, impossible to assert that many other such distinct groups do not occur 
wherever the requisite soil and other circumstances permit. 
A considerable number of such localized species occur on tracts where the soil is of so 
peculiar a nature as to have earned a special name for such places from the natives, who 
call them Jppellings. This name is applied by the Arekoonas to certain tracts in 
which the underlying very soft sandstone is overlaid by a coating of hard dense and 
dry mud, or, in some other cases, of hard conglomerate. Wherever, as is often the 
case, this hard-mud surface is unbroken, it resembles an asphalt pavement, or, perhaps, 
rather a floor made of hard-beaten earth. But this curious earth-surface overlies hill and 
dale alike—is, therefore, not often level. Wherever, then, there has been the slightest 
crack in its surface, rain-water gathers, and, having once obtained a lodgment, eats away 
and enlarges the crack. The result is an eppelling surface, which, instead of being like 
an asphalt pavement, is like a pavement formed of irregularly-shaped and scattered flag- 
stones. But, again, the mud-layer which overlies the eppelling being by no means thick, 
whenever this has once been indented, as just described, by many cracks enlarged by 
water, these cracks are soon engraved through the mud-layer down to the soft sandstone 
below ; and, when this has once occurred, the sandstone thus exposed, which yields to the 
action of the water even more readily than does the hard mud, is rapidly worked out. In 
this way the eppelling is made to assume the form of a number of blocks of sandstone, 
often pillar-like. Each of such blocks is capped and protected by a patch of the original 
hard earth, or, in other cases, of the original conglomerate. (See woodeut, fig. 1, p. 254.) 
Now, where the original eppelling surface is unbroken, in which state we have com- 
pared it to an asphalt pavement, it is as entirely devoid of vegetation as such an artificia, 
pavement would be. But where the surface of the eppelling has reached its furrowea 
stage, a few plants find lodgment, chiefly certain orchids and other such plants, of which 
the roots are of such a nature that, in the dry season, when the furrows are waterless, the 
whole plant shrinks into complete rest, and even in some cases loses its roothold, and is 
