436 
THE WILSON JOURNAL OF ORNITHOLOGY . Vol 123, No. 3, September 2011 
underparts, but this breast patch is less extensive 
or absent in subgularis (Baptista et al. 1997: pers. 
obs.). The undertail coverts—dark chestnut in the 
two small-island taxa—are reported to be paler in 
epia (Baptista et al. 1997). but this is not 
conspicuous in the field (pers. obs.. Frontispiece). 
Thus, plumage variability in this fruit dove 
complex does not parallel the vocal leapfrog 
pattern. 
There is also no concordance between acoustic 
similarities of terminal taxa and geographic 
variability in plumage patterns in Andean brush 
finches of the Arremon torquatus complex 
(Cadena and Cuervo 2010). The unlinked nature 
of character variability in these groups supports 
Remsen (1984b) in that leapfrog patterns may be 
due to stochastic phenotypic changes in geo¬ 
graphically intermediate populations that have 
not affected terminal populations. However, it 
may also indicate differential selective pressures 
on various phenotypic traits within the same 
taxon. 
Speciation and the Evolutionary Origin of the 
Leapfrog Pattern .—ft is difficult to identify the 
cause of the vocal leapfrog pattern in P. 
subgularis based only on vocalizations. Phyloge¬ 
netic data would be useful in establishing 
relationships among the three taxa and inferring 
a speciation scenario. Morphological information 
suggests the Maroon-chinned Fin it Dove is most 
closely related to PtiUnopus species from Sula¬ 
wesi and the Philippines (P. marchei, P. merrilli, 
P. occipitalis , P. Jischeri ) and especially to the 
Black-chinned Fruit Dove ( P . leclancheri ) from 
the Philippines with which it forms a superspecies 
(Baptista et al. 1997). Its origin may lie in 
Sulawesi, and it could have colonized Peleng, 
Banggai, and the Sula Islands from there. Hall 
(2002) showed that Peleng, along with the other 
Banggai Islands and the Sula Archipelago, closely 
approached Sulawesi only within the last 4 
million years. 
and Chappell 2001. Siddall et al. 2003. Bintanjaet 
al. 2005, Thompson and Goldstein 2005, Caputo 
2007). The shallowest sea connection between the 
Banggai and Sula archipelagos is only -100 m 
deep (Becker et al. 2009). although the two island 
groups are —80 km apart. Narrow land connec¬ 
tions between Banggai and Sula must have existed 
on -20 occasions for 10.000-50.000 years each 
within the last 3 million years (Lambeck and 
Chappell 2001. Siddall et al. 2003. Bintanja cl al, 
2005, Thompson and Goldstein 2005. Caputo 
2007). 
P. v. subgularis and mangoliensis evolved strong 
vocal and plumage differences despite the repeated 
presence of glacial land connections. Some of these 
land connections may have been too low-lying and 
unstable to provide breeding opportunities for 
Maroon-chinned Fruit Doves. However, both taxa 
occur close to sea level and adjacent to mangrove 
forests (pers. obs.), and we assume the two taxa 
have repeatedly met during past glacial periods. 
Vocal and morphological characters of the two 
species have not amalgamated despite likely 
hybridization opportunities. A form of character 
displacement may have occurred during glacial 
encounters where mangoliensis evolved different 
underpart colors from subgularis with the latter 
presumably exhibiting the ancestral pattern shared 
with the aJJopatric epia. Vocally, the geographi¬ 
cally intermediate taxon subgularis probably 
evolved a much faster song type that radically 
differs from the presumed ancestral song type 
uttered by the two terminal taxa. 
Biological Species Status for All Three Taxa.- 
The three taxa epia, subgularis, and mangoliensis 
were each originally described as separate species 
(Quoy and Gaimard 1830. Meyer and Wigles- 
worth 1896. Rothschild 1898. Oberholser 1918). 
Their current treatment as one species Ptilinopus 
subgularis dates to Peters (1937) at a time when 
trinomial nomenclature came into widespread use 
and ornithologists started to lump similar allopai- 
ric terms into polytypic species. This taxonomic 
Colonization, in whichever direction it oc¬ 
curred, must have included overwater dispersal 
at least between Sulawesi and Peleng. as these 
two islands have not been connected by land 
bridges. The straits between Peleng and Sulawesi 
are only 14 km wide at their narrowest point 
( ig. 1). but constitute a deep sea trench of 400- 
700 m-depth (Becker et al. 2009). Consequently 
sea level reductions of up to 130 m during the 
glacial periods of the past 3 million years were 
"sufficient to connect the two islands (Lambeck 
approach has olten been attributed to have 
happened under the umbrella of Mayrs (1942. 
1969) Biological Species Concept (BSC). How¬ 
ever. many ol these mergers have not been re¬ 
examined under a modem BSC premise, which is 
more stringent concerning diagnosability and 
monophyly (Johnson et al. 1999). 
The question of biological species status for 
mangoliensis and subgularis is straightforward 
because their ranges cannot be considered alio- 
