OBSERVED DURING THE RORAIMA EXPEDITION. 255 
entirely covered by a gravelly layer of shattered conglomerate, a very beautiful herb, 
with flowers of an intense violet-blue—a very rare colour in Guiana—was common, and 
pleasantly reminded me of an English * viper's bugloss." It was Stachytarpheta mutabilis, 
Vahl [No. 1], which seems to me to correspond to my description of a localized species. 
Again, between the Ireng and the Cotinga rivers, there grew in abundance, and 
evidently аз a native, a plant | Furcrea gigantea| which, common enough near the coast 
of Guiana in cultivation, and even as an evident escape from cultivation, is nowhere else, 
as far as I have seen in many wanderings, wild in that colony. 
Lastly, as regards localized species, I would mention several dwarf bamboos, none of 
which, unfortunately, did I succeed in finding in flower. One of these, a wonderfully 
graceful species, appears to me peculiar, in that it grows in dense thickets on the open 
savannah. This was on Ше Ireng river, and more sparingly onward from there toward 
the Cotinga. Another of these bamboos (Chusquea [sp.?], No. 18), I think the most 
graceful plant I ever saw, occurred sparingly, and only in one spot, on the Arapoo river 
close to the village of Tooroiking. A third bamboo, a climbing form ( Guadua) | No. 859), 
occurred to me first on the same river, but is much more common on Roraima itself, and 
should perhaps be spoken of in connection with the vegetation of that mountain. 
Turning next to the areas of distinet vegetation, the first to be mentioned is that of 
the Kaieteur savannah *. ‘This is certainly a very remarkable place, with an equally 
remarkable vegetation. It is an open space, some two miles long by one across, in the 
heart of the ordinary dense forest, and some four days journey on foot from the nearest 
open country. It has been said that the descent from the tableland of the interior 
toward the sea is not a gradual slope, but occurs chiefly in a series of step-like 
descents. These descents are generally of no great individual height ; but that 
of the Kaieteur takes the form of an almost abrupt cliff—at the Kaieteur fall itself it 
is an actual cliff—of between seven and eight hundred feet in height. The Potaro river, 
rising apparently from the neighbourhood of, but not actually on, Horaima, after an 
unknown upper course of considerable length, runs along one side of the almost 
perfectly level Kaieteur savannah, and precipitates itself, at the east end of that savannah, 
down the sheer descent of 800 feet. The savannah itself is virtually a flat exposed rock, 
many parts of which are as absolutely bare as a London pavement. ‘This rock is 
sandstone, which, as in the eppellings (indeed it probably is one, but of unusually 
unbroken surface) is capped by a harder material, a layer of conglomerate. Just as 
the hard surface of the eppellings cracks, and eventually affords roothold in the fissures 
thus made for plants, so the hard conglomerate covering of the Kaieteur savannah has 
cracked, and in many of the fissures thus produced has given harbourage for plants. Some 
of these latter fissures have gradually been filled up by the accumulation of vegetable 
matter; others remain still open. On this savannah, however, the fissures are larger 
than is commonly the case in the eppellings—are, in fact, often very long but generally 
narrow fissures. Many of these are now entirely occupied by shrubs and dwarf trees. 
The lines of these masses of vegetation, necessarily following the direction of the fissures, 
* Some excellent * Remarks on the aspect and flora of the Kaieteur Savannah” were published by my friend 
Mr. б. 8. Jenman in * ТтмЕнвт” vol. i. (1882) p. 229. 
