ee - 
OF THE MATTO GROSSO EXPEDITION, 1891-92. 293 
REMARKS ON THE PROBABLE ORIGIN OF THE FOREGOING FLORAS. 
It will scarcely be disputed that the whole of Eastern and Central Brazil has, in times 
past, been a great area of evolution and of dispersal. Indeed, the large number of 
peculiar genera endemic there—genera, many of them, monotypic or oligotypic—the wide 
distribution of many of the species, and the high geological antiquity of this part of 
South America are all evidence in favour of the proposition. To the student of phyto- 
geography the question of interest here raised is, by what means was dispersal effected ? 
Doubtless the movements of quadrupeds and of birds, wind, dust-storms, &e. have played 
their part in South America as elsewhere, but I cannot help thinking another and even 
more effectual means to have been brought into operation. A glance at the map of 
Brazil will show that the country has two large water-sheds—an eastern, dividing the 
waters of the Sao Francisco and Parahyba on the east and north-east from those of the 
Tocantins on the west; and a western, running more or less east and west, interposed 
between the numerous feeders of the Araguaya, Xingu, Tapajos, and Guaporé flowing 
northward and westward, and those of the Paraguay with a southward flow. When these 
rivers are in flood many fruits and seeds must certainly be carried down to the lower 
country; consequently in the disposition of these water-sheds we can see means for 
ensuring, firstly, wide disposal of species through Eastern Brazil, and secondly dispersal 
of Central Brazilian types northward to Amazonia, and southward to the country watered 
by the Paraguay and its feeders, such as the Cuyabá, Sipotuba, Cabagal, and Jauru rivers. 
Moreover, from the Sáo Francisco-Tocantins water-shed may well have been derived 
species whose distribution is Northern Goyaz and Eastern Brazil; and this may be offered 
as a plausible explanation of the considerable resemblance there is between the floras of 
Minas Geraes and Northern Goyaz, although the two districts fall within the bounds of 
different phytogeographical provinces. Further, the large percentage of diffused Tropical 
American types in the floras already dealt with receives a simple elucidation from the 
fluvial geography of Brazil. The density of the vegetation upon the river-banks and its 
comparative sparseness elsewhere have already been mentioned ; this fact must obviously 
enhance the importance of the rivers as distributing agents. One may here refer to the 
occurrence at Corumbá, as far south as the nineteenth parallel, of several Amazonian 
plants not found in any immediate station, and, in spite of the strongly South- 
Brazilian character of the Corumba flora, quite unknown in the eastern part of the 
country. What hypothesis sounder than that here advocated can possibly be put 
forward to account for this ? 
But, besides this north and south distribution, there are clear indications as to the range 
of many species of the Matto Grosso flora being dependent upon latitude. This fact will 
be made clear to the reader on cursory examination of the distribution of species 
given in the following pages. He will see that a considerable percentage of East- 
Brazilian types pass through Matto Grosso to Eastern Bolivia, but do not occur to 
the northward. Similarity in the matter of rainfall is probably the reason of this, there 
being a marked difference in that respect between South and North Brazil, and marked 
agreement between the eastern provinces and Matto Grosso. 
