Because of the lower quality of our Ateles marginatus sample, we
were unable to initially amplify our target fragment in a long range
PCR and, subsequently, to sequence the entire 3480 bp we
obtained for most other samples. When the analyses above are
repeated excluding sequence data from the NAD6 locus, the
inferred phylogenies are broadly similar to those noted above.
Again, A. marginatus is reconstructed as basal with 100% support
in both the maximum likelihood and Bayesian analyses, as are
clades comprising A. belzebuth + A. chamek and A. fusciceps + A.
geoffroyi. Details of the inferred phylogeny, however, are not
always identical. In the likelihood analysis, A. hybridus is joined
to A. belzebuth + A. chamek in a weakly supported clade, and the
relationships among A. paniscus, A. belzebuth + A. chamek + A. hybridus,
and A. fusciceps + A. geoffroyi are unresolved. In the Bayesian
analysis, the inferred branching order conforms to what was recovered
in the likelihood analysis using all three loci, but with higher
support for a clade of A. hybridus + A. paniscus + (A. fusciceps + A.
geoffroyi) (Fig. 3C and D).