Dispersal centers of the Amazonian acridids

The dispersal centers of the Amazonian acridid arboreal fauna are defined on the basis of the study of the subfamilies Bactrophorinae and Proctolabinae. Four dispersal centers are revealed which correspond to the main centers proposed by Miiller (1973) and Haffer (1979). The ecological characteristics of their fauna and the relations between the centers are tentatively approached through the study of some polyceniric dis­ tr ibution patterns (Ommatolampinae, Proctolabinae) as well as through the more restricted relations between Hit Amazonian centers and the Atlantic Forest.


INTRODUCTION
In order to define the principal distribution centers of the Amazonian acridids, in this study we will mainly use two of the most important arboreai groups that have recently been re vised : the Bactrophorinae (Descamps, 1978, s/d) and the Proctolabinae (Descamps. 1976(Descamps. , 1980)).
The Bactrophorinae account for 31 to 47%

METHODS
In order to separate and characterize ths different regional faunas within the Amazonian fauna, we used the following taxonomic cate gories : -monospecific and limited-species gene ra, endemic in parts of the Amazonian basin; -groups of closely related species until now only known by punctual or quasi-punctual signalizations; -species with a medium-range distri bution; -subspecies, used in a few cases.

DISPERSAL CENTERS
The four dispersal centers defined in the above manner are almost the same as those recognized by Muller (1973) for terrestrial vertebrates.We will, therefore keep his termi nology.However, Muller differentiates a fift n center (Para), located to the east of the Rio Tocantins, a region for which we have no information on the groups concerned.
The two western centers, Napo in the north and Ucayali in the south, are clearly differentiated for the Bactrophorinae (Fig. 1) .
However, with regard to the Proctolabinae (Fig. 2 These results also agree with those of 1) The region south of the Lower Amazon between the Rio Xingu and the Atlantic coast.
2) The Venezuelan and the Rio Negro regions.Thus, the Madeira center is characterized by some Proctolabinae, often thamnophilous and having a polycentric distribution (Fig. 3).

OVERLAPPING
The The sizes of the gaps between the different taxons present in the dispersal area show that the changes which gave rise to them took place at different periods and were probably related to several expansions of a suitable vegetation (Prance, 1978).
Although the arboreal fauna of the Ucayali center is insufficiently known, the contours of the center were able to be delimited through the study of the Pactrophorinae and the Proc tolabinae.
However, the occidental part of this center is also characterized by its diversity of species of Ommatolampae ( 4 ) (Fig. 4) and par ticularly by their abundant number.This latter group, comprised only of micropterous and   No longer situated at the genus or the species level but at the tribe level, the peculiar

RELATIONS WITH THE ATLANTIC FOREST
The arboreal fauna of the Atlantic Forest is very poorly known at the present time.
Moreover, as a consequence of intense agri culture and breeding, any hope of obtaining more knowledge about this region seems to be already lost.
One of the most striking features of the Amazonian forest fauna is its extraordinary richness, as is recognized by all the specialists who have undertaken its study.Although often considered as typical open-biotope insects, the acridids ( 2 ) also greatly contribute to this wealth, and such a diversity of acridid species cannot be observed anywhere else in the world to a more important species diversity of the New World compared to that of the Old World, and is particularly character istic of recent fractions of old American acridid groups or of insects newly implanted in South America.However, this species diversity is much more noticeable in forest groups, 74 to 84% of the species and 84 to 92% of the individuals.Each one of these subfamilies includes groups of genera that exhibit different and often opposite ecological vahnces.Their representation in the different sampels is related to their geogra phic location, to the forest types, and to the state of degradation ot the forest.
of the species studied (11 out cf 201 species of Bactrophorinae and Proctolabi nae ( 3 ) .Moreover, we observed 22 cases of overlapping at the higher level of groups of parent species (i.e., with allopatric speciation between the centers).This concerned primarily the Bactrophorinae-But, the present state of our knowledge does not allow us to say if the fluvial barrier is really a more effective factor of differentiation than the distance between stations within a given center, for as in a great number of cases, each station is characterize by its own form.These boundary overlappings appear on the maps to be limited to the main rivers, but as most of the samples that enabled us to detect them were located along these barriers, the(3) -7 species of this group are known only from Benjamin Constant and Tabatinga and, therefore, cannot be assigned to one of the centers.This explains wh/ center overlapping doss not appear on the map (Fig.2).true range oi these overlappings remains largely unknown.However, the sample from Jutai, located half-way between Madeira and Purus, sesms to show that the boundary overlappings may be wide and are probably greatly dependent on species or groups.Despite these marked overlappings, it seems that rivers do function somewhat as obstacles to dispersal.This is particularly true for the less vagile Bactrophorinae in which were found, among the 33 cases cited above, most of the differences arid conversely the smallest number of the similarities.The wide distri bution of most of the species of Proctolabinae compensates for the apparent lack of an Upper Amazonian overlapping.Even of the rivers constitute effective barriers to dispersal, we must note that along the Rio Madeira, the Rio Negro, and th° Lower Amazon, they are rein forced by heterogeneous bands in which the forest is interspersed with areas of less-dense vegetation (campos cerrados, campinas) .These zones, which subsist in fact as ecological barriers, must have been much more pro nounced during the Quaternary xeric periods and were obviously more effective obstacles to the dispersal of the fauna than were the rivers-Such a structure does not exist in the zone of contact between the Napo and the Ucayali centers where climate and vegetation are similar.As a result of these ecological conditions, it is coherent that the occidental faunas are more closely related than the oriental and the occidental faunas or those of the Madeira and Guianan centers.However, the barrier zones comprised of heterogeneous vegetation and major rivers are not absolute, for the insects cross them in all directions in the overlapping zones which are, in fact, zones of Interpenetration.