Savannah Vegetation. 
that gro>v in isolated places here iu the ueiglibourhood of the Pirara soon 
lay behind us. With the crossing of this river the rolling ground stretch- 
ing towards the south completely disappeared.. Our course was a north- 
westerly one. The change in surface formation was accompanied with 
an alteration in soil. The clay that generally constitutes the substratum 
lost its red colouring, while those roundel .sliiuy bits of qiur. tz ami cJay, 
stained red-brown Avith oxide of iron, tliat liad covered the undulating 
savannah for miles and miles, Avere no longer visible. The ant-hills were 
also entirely Avantiug in tlie plain. As the whole ot tliis stretcli lies 
something like a linndred feet beloAv Pirara Village it forms a lake, 
l^ake Amueu or Pa;-iiiia,* dnriiig the rainy season. In its vegedition 
tills flat corresponded entirely with what was already known to me, 
except that the predominating Jtifrsoiiiiiia verba scifolia covering the 
savannah in all directions Avith its silveiy felt-like leaves and yelloAV 
blossoms at least caused a break in the general uniformity: the latter, at 
all events, had reached their culm|inating point, owing to the di-y season 
■ noAv coming to an end. Ci/peracrnc such as Cupenis ainentacciis Kudge, 
If^olcpis junciforiiiks Tluni'J). Por.]»., 7. rapi1J((ris Eoem. et Schult., 
Hypolytrum pungrns Clilorulra'^ aM(\ Fcstucavcac, i\\ between 
which one nevertheless found plenty Eriocauloncae, formed the usual 
herbage-covering. Among the lattin* the l\trpaIantJuis capilhiceiis 
has a specially peculiar interest in that the Indian by burning off the 
saA^annali apparently helps it to blossom. When the fire destroys all the 
leaves the innumerable sAveet-scented knob-like buds start developing 
within the next two or three days from out of the thick, short, leafless, 
blackened stalks: if the flowering season is over, then the leaves alone 
appear. I have only in exceptiolially rare cases found specimens AAiiich, 
escaping the fire, possessed leaA'es and blossoms at one and the same time. 
A species of grass, though I never saAA' it in flower, often covered entire 
areas to the exclusion of others, and Avas speciallA' interesting on acoonnt 
of its name: the Macusi called it vaiinali and such flats A'andai. Whether 
the term Savannah might be derived from this Avord. T must of course leave 
to etymologists to decide. f The ground AA^as so cracked Avitb the intense 
heat that it AAas covered witli a regular network of clefts over three to four 
inches wide . 
6. Our journey Avould have proA cd extremely monotonous today had 
it not been lightened by the Canuku and Pacaraima Ranges in the south 
and north respectively, combined Avith the innumerable blue blossoms of 
Eichhornid a.:iirea and Hricranfhcra Vuiiom Vahl. tliat had gathered 
near the iioav almost Avaterless Lake Venturu and tlie large flocks of duck 
which on our near approach rose from it with a piping cry. That the 
'Anas vidnata and hrasiUensis had to supply a numerous contingent for 
our next meal, it is unnecessary for mie to say. Resides these, an occa- 
sional Cara-Cara eagle or two went careering around the dried-up swampy 
ground, while small companies of //>/.s- alhicoHtn Vieill. enlivened the 
* According to Sanson's map of lH5('i, Parimc is the name given it by the Carib, but Ro- 
ponowitii by the laoyi (Eil.) 
t The word is derived from (tlie Spanish) Sabana, a cloth or large sheet (Ed.^ 
B 2. 
