Rheindt et al. • LEAPFROG PATTERN IN PTILINOPUS FRUIT DOVES 
435 
01 
02 
3 
I Group Centroid 
FIG. 4. Canonical function plot of discriminant analysis; discriminant function 1 versus 2; 1: subgularis ; 2: epia; 
3; mungnliensis. 
range than the two other taxa, and our vocal 
material originates from two geographically 
distant sites on Sulawesi which displayed slight 
vocal differentiation, consistent with phylogeo- 
graphic patterns in amphibians and monkeys 
(Evans et al. 2003a, b; Evans el al. 2008) that 
exhibit differentiation across the same regional 
boundaries within Sulawesi. The over-water 
distance of the North Sulawesi population is 
much less to subgularis on Peleng than to the 
other epia population in Central Sulawesi. Nev¬ 
ertheless, the treatment of both epia populations 
as a single taxon is justified by morphological 
uniformity and the limited level of vocal differ¬ 
entiation from each other, which was significantly 
less than from subgularis. Wallacean forest birds 
such as fruit doves have been known to be poor 
over-water dispersers since Alfred Russel Wallace 
noted the scarcity of bird colonizations across 
narrow sea channels characterized by deep sea 
trenches (e.g.. Wallace [1869]: 187-190). 
The present characterization of the vocal 
similarity between epia and mungoliensis is the 
first thorough demonstration to our knowledge of 
a vocal leapfrog pattern in non-passerines, as well 
as the first vocal leapfrog pattern from outside the 
Neotropics. Remsen (1984b) hypothesized the 
Andes may be conducive to producing leapfrog 
patterns because of their geographic linearity, 
topographic relief, and high species richness. 
However, the present case suggests they can 
occur in an island setting in Wallacea where none 
of the three Andean factors is as great. Leapfrog 
patterns, especially vocal ones, may be much 
more widespread in birds than assutned, and may 
be frequently overlooked. 
Does Plumage Variability Parallel the Vocal 
Leapfrog Pattern? —Plumage differences among 
the three P. subgularis taxa pertain mainly to 
underpart coloration (Baptista et al. 1997, Coates 
and Bishop 1997, Gibbs et al. 2001). Plumage of 
the eastern terminal taxon tnangoliensis is most 
conspicuous: the gray color of the underparts and 
neck of epia and subgularis is replaced with lime- 
green (Gibbs el al. 2001: pers. obs.; the labels of 
the drawings of ntangoliensis and epia in Baptista 
et al. [1997] are accidentally reversed). Differ¬ 
ences between subgularis and epia are subtler: 
epia is characterized by a conspicuous yellowish- 
buff breast patch contrasting with the gray 
