Regarding the basal position of Ateles marginatus in our inferred
topology and the separation of A. paniscus from other species of
Ateles by the lower Amazon, it is worth noting that, in both of
the biogeographic scenarios we propose, A. paniscus likely diverged
from other Ateles somewhere 3.5–4.5 Ma, again before the full
development of the current Amazon system. At that time, still –
and even as early as 6.7 Ma when we estimate the A. marginatus
lineage originated – the Amazon drainage to the Atlantic probably
would have constituted a significant barrier to northward gene
flow from the Brazilian to the Guiana Shield in the eastern part
of the continent. Thus, if A. paniscus were indeed the next lineage
to diverge after A. marginatus (Fig. 5B) then a south-to-north vicariance
event across a long-established lower Amazon drainage
would have been required. In the other possible scenario we sug-gest (Fig. 5A), gene flow from the south-central Amazon to the
west and north (into what is now the northwestern Amazon of
Colombia, Ecuador, and Peru) as well as the origins of Ateles belzebuth,
chamek, and hybridus would have occurred somewhere
around 4.5–4.0 mya, in the Pliocene, and would have coincided
with a period of drying of the western Amazonian Pebas lake formation.
As Aleixo and Rossetti (2007) note, a number of avian species
also expanded from the Brazilian Shield into the western
Amazon lowlands at this time.