The center of the center of marine shore fish biodiversity: The Philippine Islands

Abstract

Multiple datasets show global maxima of marine biodiversity in the Indo–Malay–Philippines archipelago (IMPA). Analysis of distribution data for 2983 species reveals a pattern of richness on a finer scale and identifies a peak of marine biodiversity in the central Philippine Islands and a secondary peak between peninsular Malaysia and Sumatra. This pattern is repeated in diverse habitat and higher taxa classes, most rigorously for marine shore fishes, supporting geohistorical hypotheses as the most general unifying explanations. Specific predictions based on area of overlap, area of accumulation, and area of refuge hypotheses suggest that present day eastern Indonesia, or Wallacea, should be the center of marine biodiversity. Processes suggested by these three hypotheses contribute to the diversity in this region and are also a likely explanation for the secondary center of diversity. Our study indicates, however, that there is a higher concentration of species per unit area in the Philippines than anywhere in Indonesia, including Wallacea. The Philippine center of diversity is consistent with hypotheses that this area experienced numerous vicariant and island integration events and these hypotheses warrant further testing. Special attention to marine conservation efforts in the Philippines is justified because of the identification of it as an epicenter of biodiversity and evolution.

Full text

The center of the center of marine shore fish biodiversity: the Philippine Islands
Kent E. Carpenter
a
& Victor G. Springer
b
a
Department of Biological Sciences, Old Dominion University, Norfolk, VA 23529, U.S.A.
(e-mail: kcarpent@odu.edu)
b
Division of Fishes, Department of Zoology MRC-159, National Museum of Natural History, Smithsonian
Institution, P.O. Box 37012, Washington, DC 20013-7012, U.S.A.
Received 21 April 2004 Accepted 12 August 2004
Key words: Indo-Pacific biogeography, vicariance, island integration, conservation
Synopsis
Multiple datasets show global maxima of marine biodiversity in the Indo–Malay–Philippines archipelago
(IMPA). Analysis of distribution data for 2983 species reveals a pattern of richness on a finer scale and
identifies a peak of marine biodiversity in the central Philippine Islands and a secondary peak between
peninsular Malaysia and Sumatra. This pattern is repeated in diverse habitat and higher taxa classes, most
rigorously for marine shore fishes, supporting geohistorical hypotheses as the most general unifying
explanations. Specific predictions based on area of overlap, area of accumulation, and area of refuge
hypotheses suggest that present day eastern Indonesia, or Wallacea, should be the center of marine bio-
diversity. Processes suggested by these three hypotheses contribute to the diversity in this region and are
also a likely explanation for the secondary center of diversity. Our study indicates, however, that there is a
higher concentration of species per unit area in the Philippines than anywhere in Indonesia, including
Wallacea. The Philippine center of diversity is consistent with hypotheses that this area experienced
numerous vicariant and island integration events and these hypotheses warrant further testing. Special
attention to marine conservation efforts in the Philippines is justified because of the identification of it as an
epicenter of biodiversity and evolution.
Introduction
The Indo-Malay-Philippines Archipelago (IMPA)
has long been considered the area of highest
marine biodiversity, with decreasing latitudinal
and longitudinal gradients in species richness
radiating from this center (Bellwood &
Wainwright 2002, Mora et al. 2003). There is a
wide variety of hypotheses that have been pro-
posed to explain the remarkable diversity found in
the IMPA (Rosen 1988). These hypotheses can be
classified into four main categories: (1) area of
overlap, (2) area of accumulation, (3) area of ref-
uge, and (4) center of origin. There is little agree-
ment as to which of these hypotheses is the most
important in shaping diversity in the IMPA
(Bellwood & Wainwright 2002). Reconciling the
relative importance of these opposing hypotheses
has led some to propose what could be considered
a fifth major hypothesis: that a combination of all
processes evoked in these hypotheses is responsible
for the extreme biodiversity of the IMPA (Palumbi
1996, Wallace 1997, Carpenter 1998, Randall
1998, Wilson & Rosen 1998, Allen & Adrim 2003).
Predictions about concentrations of diversity
within the IMPA can be made based on the vari-
ous hypotheses and the geography and geology of
the IMPA. One assumption in these predictions is
that large shelf areas of the IMPA were exposed
and therefore experienced a series of local marine
extinctions during Pleistocene sea-level lows
(Springer & Williams 1990, Voris 2000, Figure 1)
Environmental Biology of Fishes (2005) 72: 467–480 Springer 2005

and potentially limited diversity in these areas.
With these Pleistocene exposed areas excluded, the
two major marine habitat areas remaining in the
IMPA are the Philippine Islands and the area re-
ferred to primarily by terrestrial biologists as
Wallacea. Wallacea is a term not frequently used
by marine biogeographers (Wallace 1997) and is
typically defined as the area between the famed
Wallace’s (1860) line and Lydekker’s line (Simpson
1977, Figure 1). We use it here as a convenient
term to refer to the area which encompasses the
island groups of Sulawesi, Moluccas, Halmahera,
and the Lesser Sunda Islands. The area of overlap,
area of accumulation, and area of refuge hypoth-
eses suggest Wallacea as an area of likely peak
marine diversity.
Area of overlap hypotheses contend the IMPA
is an area where different faunas from different
oceans or lithospheric plates aggregate. Overlap
between Pacific and Indian Ocean faunas
(Woodland 1983, Donaldson 1986) would con-
centrate diversity in the Wallacean corridor be-
cause it is centrally located and has ample habitat
to support species. Throughout the later half of the
Cenozoic, elements of Wallacea were positioned
across the central dispersal route between the Pa-
cific and Indian oceans. Up until the mid-Miocene,
the south equatorial current spanned the Pacific
and Indian oceans (Kennett et al. 1985, Figure 2).
During the Oligocene and much of the Miocene,
elements of Wallacea had mostly formed and were
in the path of the Indonesian Seaway that con-
nected the Pacific and Indian oceans (Hall 1998).
The Indian Ocean south equatorial current would
also have been in contact with elements of Wall-
acea (Figure 2).
Australian and Eurasian continental shelves
collided and deflected the southern equatorial
currents somewhere between 16 and 8 mya
(Kennett et al. 1985) and formed an equatorial
barrier between the Pacific and Indian oceans.
This transformed the southern equatorial currents
into boundary currents and the surface current
patterns seen today began to take shape
(Figure 3). Throughout the late Miocene and Pli-
ocene the major island groups of Wallacea were in
the primary pathway of currents flowing eastward
from the Indian Ocean and westward from the
Pacific Ocean (Hall 1998). During the Holocene
and times of Pleistocene sea-level highs, Indian
and Pacific Ocean exchange occurred primarily
Figure 1. The Indo-Malay-Philippine Archipelago with the
area shallower than the 120 m depth contour delineated in light
gray. This area estimates the sea bottom that would have been
exposed during Pleistocene sea-level lows. Major land masses
include the Palawan–Busuanga–Mindoro archipelago (PBM),
northern Luzon (NL), Visayas–Mindanao (VM), Halmahera
(HA), Moluccas (MU), and Sulawesi (S). Reproduced and
modified with permission from Voris (2000), copyright Field
Museum of Natural History, Chicago.
Figure 2. Position of land (diagonal lines), shallow seas (gray),
and deep ocean (no color) in Southeast Asia 30 million years
ago. Relative position of present-day land masses are drawn
with respect to these ancient features, including the Palawan–
Busuanga–Mindoro archipelago (PBM), northern Luzon (NL),
Visayas–Mindanao (VM), Halmahera (HA), Moluccas (MU),
and four components of Sulawesi (S1–S4). Modifed from Hall
(2002).
468

through Wallacea (Flemminger 1986, Fine et al.
1994, Figure 3). During Pleistocene sea-level lows,
the only pathway between equatorial Pacific and
Indian oceans was through Wallacea (Fleminger
1986, Figure 3).
The IMPA is also considered an area of over-
lap with different endemic shore faunas ‘rafting’
into the area through the movement of
lithospheric plates (Pandolfi 1992, Wilson & Ro-
sen 1998, Santini & Winterbottom 2002). Colli-
sion zones between plates would concentrate
different faunas. A primary overlap area would
be the triple juncture zone, located in eastern
Wallacea (Figure 4).
The area of accumulation hypothesis (also
called the vortex hypothesis) suggests most speci-
ation occurred on remote Pacific islands and pre-
vailing currents concentrated species in the IMPA
(Ladd 1960, Jokiel & Martinelli 1992, Connolly
et al. 2003). A prediction from this hypothesis is
that Wallacea would be a center of diversity
within the IMPA based on prevailing currents.
The middle to late Cenozoic Pacific equatorial
currents primarily flow towards elements of
Wallacea (Figures 2 and 3, Kennett et al. 1985,
Hall 1998).
Related to the area of accumulation hypothesis
is that the IMPA also serves as an area of refuge
(Bellwood & Hughes 2001) because it encompasses
Figure 3. Present-day surface current patterns in the IMPA. (a)
Winter, (b) Summer. Light gray indicates areas shallower than
the 120 m depth contour that would have been exposed during
extreme Pleistocene sea-level lows [current patterns modified
from Morgan & Valencia (1983), base map modified from Voris
(2000)].
Figure 4. Present-day location of major lithospheric plates of
the IMPA as delineated by major subduction zones (black
lines). The dashed line indicates the approximate location of the
rift zone that separated the Indian and Australian plates in the
early Cenozoic.
469

the most extensive and diverse tropical shallow
water marine habitat on earth (Carpenter 1998).
The significance of extensive tropical refugia
relates to the ecological maxim that larger areas
generally support more species than smaller areas
and that tropical low latitude habitats have greater
available energy to support biomass than cooler
latitudes (Gaston 2000). The area of refuge
hypothesis would predict that the Indonesian–
Malaysian part of the IMPA will have the highest
species richness. This area has the greatest extent
on earth of tropical shallow water habitat
(Carpenter 1998), including the greatest extent of
coral reefs (Spalding et al. 2001) and diversity of
habitat types (Woodland 1990, Randall 1998).
The center of origin hypothesis predicts that
areas of highest diversity would be areas that
hosted numerous allopatric speciation events
(McManus 1985, Briggs 1999a, Mora et al.
2003). The center of origin hypothesis has also
been proposed in terms of numerous competitive
or sympatric speciation events (Briggs 1999b) but
this mechanism has been dismissed as an
important means of speciation in the IMPA
(Springer & Williams 1990, Santini &
Winterbottom 2002). Both Wallacea and the
Philippines are potential areas of concentrated
allopatric speciation. Two major vicariant pro-
cesses occurred within the IMPA during the
Cenozoic when most of the shore fish fauna
evolved. These were: (1) potential isolation of
seas during Pleistocene sea-level lows (McManus
1985) and (2) geological origin through complex
tectonic movements (Hall 2002). A number of
seas within Wallacea and the Philippines were
potentially isolated from one another to varying
degrees during Pleistocene fluctuations in sea le-
vel. These included the South China, Sulu, Phil-
ippine, Celebes, Molucca, and Banda seas
(Figure 1). The origins of the Philippine and
Wallacean island arcs are particularly compli-
cated because of integration of varied Philippine
plate elements (Hall 2002).
The purpose of this paper is to examine results
of a geographical information system analysis of
marine species found in the western central Pacific
Ocean with respect to hypotheses proposed to
explain the biodiversity found in the IMPA. We
also comment on the importance of these findings
with respect to conservation.
Methods and results
The geographical information system database and
analyses
We used a geographical information system (GIS)
overlay of 2983 generalized distributions of marine
species (Carpenter & Niem 1998–2001) to examine
the pattern of diversity in the IMPA. Species
included seaweeds (62), corals (27), bivalves (189),
gastropods (249), cephalopods (87), stomatopods
(13), shrimps (110), lobsters (47), crabs (73), ho-
lothurians (18), sharks (150), batoid fishes (116),
chimaeras (6), bony fishes (1775), estuarine croc-
odile (1), sea turtles (6), sea snakes (21), and
marine mammals (33) produced by 84 specialists in
their respective taxonomic groups (Carpenter &
Niem 1998–2001 and authors cited therein). Taxa
were from continental shelf and epipelagic envi-
ronments. A subset of this database contains the
2047 shore fish species that included Elasmo-
branchii (sharks and batoid fishes), Holocephali
(chimaeras), Sarcopterygii (Sulawesi coelacanth),
and representatives from 24 orders of Teleostei.
The mix of species in this study includes mostly
common fisheries species whose distributions are
well known throughout the IMPA. It largely
excludes species such as smaller gobioid or blen-
nioid fishes for which distributions are mainly
derived from sampling in highly restricted areas.
Indonesia has a rich history of research on fishes
beginning in the early 19th century, including very
diverse collections made by the Dutch natural
historian, Pieter Bleeker (Weber & deBeaufort
1913). Extensive collections around Indonesia
continued through the present including notable
surveys of trawled (Gloerfelt-Tarp & Kailola 1984)
and coral reef fishes (Allen & Adrim 2003). There
were limited studies of Philippine fishes in the 19th
century (Herre 1953) but extensive collections be-
gan in the early 20th century (Smith & Williams
1999). Because of the extensive collections in both
Indonesia and the Philippines, distribution records
for common fisheries species would be thoroughly
covered in both countries. And, since the distri-
bution records are based primarily on museum
collections and not fisheries statistics (taxonomists
consider these mostly unreliable unless tied to a
museum specimen), uneven fisheries pressure also
would not influence these distributions. We do not,
470

therefore, suspect sampling bias in our dataset
within the IMPA.
One potential source of bias in the current
dataset is taxonomic bias in that fishes and verte-
brates are covered in greater proportion to overall
diversity than invertebrates. Far more species of
invertebrates make up the diversity of the IMPA
than the fishes but these are not as important
proportionally in fisheries. This is particularly the
case when it comes to corals that make up a con-
spicuous part of the macrofaunal invertebrate
fauna of the IMPA. Taxonomic bias has been
criticized in other biogeographic studies (Baird
et al. 2002, Roberts et al. 2002). We examine all
the maps in combined analyses and also examine a
partition including only the shore fishes since this
later group was covered more rigorously.
All 2983 generalized maps were digitized in PC
ArcInfo and analyzed using ArcView (Environ-
mental Systems Research Institute, Inc.) with a cell
size set at 10 km ·10 km. Species were chosen
based on their likelihood to enter fisheries in the
western central Pacific FAO fishing area 71 and
the South Pacific Commission Area (Carpenter
1998). This region corresponds to most of the
tropical and subtropical western Pacific from
about 98E to 122W longitude excluding the
Hawaiian Islands and Johnston Island.
The distribution of each species was categorized
according to concordance with a major litho-
spheric plate (Figure 4). If the major portion of a
species distribution was found on one lithospheric
plate but it extended marginally onto another
plate it was considered endemic only to the plate
that it covered most completely. These plates
were the Eurasian plate (including the Philippine
and Moluccan marginal arc systems), the Pacific
plate, the Indian plate, and the Australian plate.
The Indian and Australian plates are currently
considered a single plate. However, these plates
were separated during the early Cenozoic (Hall
2002, Figure 4) and the shore faunas on these
plates were widely separated and presumably
evolved separately. In addition, the Indian por-
tion of the Indo-Australian plate was not directly
encompassed in this study. The Philippine Sea
plate is also a major lithospheric plate but is
mostly open sea with few identifiable endemics
(Myers 1989) when its marginal arc island systems
are considered part of Eurasian or Australian
shore faunas. Distributions covering more than
incidentally one lithospheric plate were catego-
rized as: Indian–Eurasian, Eurasian–Australian,
Western Pacific (distributions that encompassed
inclusively the Eurasian, Pacific, Philippine, and
Australian plates), Indo-westmost (species found
inclusively on the Indian, Australian, and Eur-
asian plates but not extending onto the Pacific
plate), and Indo-West Pacific (widespread on all
Indian and Pacific ocean plates). Habitat type was
categorized as to either primary coral reef, pri-
mary soft bottom, rocky, estuarine, generalized
demersal, generalized neritic, and epipelagic hab-
itat types if this was indicated in the habitat
section of each species account (Carpenter &
Niem 1998–2001).
Results from preliminary GIS analyses
Analysis of the 2983 combined ranges reveals the
central Philippines as the area of highest diversity
and endemism (Figure 5a, b). A secondary area of
high diversity is located between the tip of
Malaysia and Sumatra and extends along north-
eastern Sumatra and northern Java (Figure 5a).
Both diversity centers are repeated in subsets of
data based on distribution, habitat, all inverte-
brate taxa, and shore fishes (Figure 5c–f, Table 1).
Families and genera with predominantly small-
bodied species, which tend to have a higher pro-
portion of restricted-range endemics than taxa with
large-bodied species, were not included in our
study. Nevertheless, 120 restricted-range endemics
were included because they enter area fisheries. The
greatest concentration of these restricted-range
endemics is in the Philippines (Figure 5b) which
has 38. Indonesia/Malaysia has 19 such endemics,
Australia 18, New Guinea/Bismark/Louisade 18,
Coral Sea/New Caledonia/Vanuatu 17, and seven
other localities had either one or two endemics.
Analysis of distribution and habitat types in the
primary and secondary center of diversity
To examine the components of biodiversity we
analyzed the frequencies of general distribution
and habitat types in the primary and secondary
centers of diversity. We used the assemblage of
species in the single 10 km ·10 km pixel with the
most species (1736 or about 58% of all species in
471

the study), which is located in the Verde Island
Passage (VIP) between Mindoro and Luzon. The
cell with the greatest number of species in the
secondary center of diversity near Pulau Bintan
(PB) in Indonesia off the southern tip of peninsular
Malaysia has 1670, or about 56% of all species in
the study. Both the VIP and PB cells corresponded
very closely with the primary and secondary peaks
in shore fishes and were also used in a separate
analysis of shore fish distributions. Expected fre-
quencies were calculated from the combined
database or all shore fishes separately. For distri-
bution types this included all species that could
potentially be found on the Eurasian plate (i.e.
excluding Pacific and Australian plate endemics).
For habitat types this encompassed the entire
dataset. Expected frequencies of distribution cat-
egories and habitat types from the VIP and PB
Figure 5. (a) Pattern of species richness in the IMPA from an overlay of the 2983 ranges of species. Each change in color or shade
represents an increase or decrease in 43 species (40 classes total or a 2.5% change per class). The top 10% of species richness is in
shades of red and yellow and the remaining decreasing increments of species richness are indicated by lighter shades of blue. The
greatest diversity is red (1693 to 1736 species), followed by pink (1650 to 1692 species), yellow (1606 to 1649 species), and light yellow
(1563 to 1605 species). (b) Overlap of restricted-range endemic species in the IMPA from the 120 of these endemics covered in this
study. Each darker shade of blue indicates an increase in one restricted-range endemic overlapping in the area; red indicates the highest
area of overlap with 19–20 endemics; pink indicates overlap of 17–18 endemics; yellow indicates an overlap of 15–16 endemics; light
yellow indicates an overlap of 13–14 endemics. (c–f) Each change in color or shade represents an increase or decrease in 2.5% of the
maximum number of species found in any one spot; the top 10% of species richness is in shades of red and yellow and remaining
decreasing increments of species richness are indicated by lighter shades of blue. The greatest diversity is red (highest 2.5%), followed
by pink (highest 2.5–5%), yellow (highest 5–7.5%), and light yellow (highest 7.5–10%). (c) All fish species (maximum overlap, 1010
species). (d) All invertebrate species (maximum overlap, 623 species). (e) Eurasian–Australian plate distributions (maximum overlap,
179 species). (f) Indo-West Pacific distributions (maximum overlap, 750 species).
472

centers of diversity differed significantly (X
2
¼
p< 0.001; Table 2) from frequencies expected for
all distributions and for shore fishes separately.
The VIP and PB centers had significantly greater
frequencies (p< 0.001) of widespread Indo-Pacific
distributions, widespread Indo-westmost Pacific
distributions, and coral reef species than overall
and shore fish expected frequencies. Both centers
had significantly fewer (p< 0.001) Eurasian plate
endemics and generalized demersal species com-
pared to expected frequencies. The VIP center had
significantly fewer (p< 0.001) Indian–Eurasian
plate distributions than expected frequencies.
Statistical comparison of observed versus predicted
primary centers of diversity
According to predictions from area of overlap,
area of accumulation, and area of refuge hypoth-
eses, Wallacea should have the highest species
richness. To test if the observed center of diversity
had a significantly higher species richness than that
observed in the predicted area, a random sub-
sample of 50 cells were chosen from continental
shelf areas within each of the Philippine area and
the Sulawesi–Moluccan area of Indonesia. After a
test of normality, a t-test was performed (more
sophisticated Monte Carlo simulation methods
were considered but the strength of the t-test and
readily apparent distribution of species concen-
trations make it obvious that such a test is
unnecessary). The Philippine center of diversity
was found to have significantly higher species
richness (p< 0.001) than Wallacea for all distri-
butions combined and also when shore fish dis-
tributions are treated separately. A similar
comparison with the Pulau Bintan area of sec-
ondary diversity also indicates that the central
Philippines has a significantly higher species rich-
ness (p< 0.001).
Discussion
A Philippine center of marine diversity is both
supported and refuted by previous studies. The
Philippines is shown with highest diversity at dif-
ferent taxonomic levels in some biogeographical
studies (e.g., Springer 1982, Woodland 1983).
Randall (1998), however, used expected rates of
discovery of new species to predict that the
greatest diversity of shore fishes would eventually
be recorded from Indonesia, which has far greater
shelf area and over twice the reef area of the
Philippines (Spalding et al. 2001). Allen & Adrim
(2003) indicate that Indonesia has the highest coral
reef fish diversity. Because of its greater area,
Indonesia may eventually be shown to have a
greater overall marine biodiversity than the Phil-
ippines. However, there is a higher concentration
of species per unit area in the Philippines than
anywhere in Indonesia, including Wallacea,
according to our study.
Identification of the center of marine biodiver-
sity provides tests of the various hypotheses that
have been proposed to explain the remarkable
diversity found in the IMPA. The central Philip-
pine center of diversity falsifies predictions from at
least three of four major categories of these
Table 1. Areas with highest diversity according to different
categories of distribution type, habitat type, and major
taxonomic groups.
Distribution/
Habitat/Taxa
Highest diversity Secondary
diversity
All maps Central
Philippines
NE
Sumatra–N Java
Australian plate SE Queensland none
Eurasian plate W Taiwan SW Luzon
Indo-West Pacific Central
Philippines
Sulawesi, N Java
Indian-Eurasian plate Straits of Malacca none
Western Pacific Central
Philippines
none
Indo-westmost
Pacific
NE
Sumatra–N Java
Central
Philippines
Eurasian
+ Australian plates
Central
Philippines
none
Coral Reefs Central
Philippines
Sulawesi,
N Sunda Ids.
All soft bottoms NE Sumatra Central
Philippines
Estuarine NE Sumatra S Kalimantan
All invertebrates Central
Philippines
NE Sumatra,
N Sabah
All fish Central
Philippines
N Java
Highest diversity was defined as cells with the upper 5% of the
maximum diversity (the red and pink areas of Figure 5a).
Secondary diversity was defined as those cells with 5–10% of
maximum diversity.
473

hypotheses. Different interpretations of the area of
overlap hypotheses indicate that Wallacea is the
likely center of overlap. The central Philippines are
removed northwesterly from the potential central
zone of overlap between Pacific and Indian Ocean
biotas. The Philippines are also removed from the
triple juncture zone of Pacific, Australian, and
Eurasian plates and further removed from the
broad collision zone between Indian–Australian
and Eurasian plates (Figure 4). The distribution of
species found only on the Eurasian and Australian
plates shows a Philippine center of diversity (Fig-
ure 5e, Table 1). Assuming that species with this
distribution type originated on one of the plates
and subsequently dispersed to the other plate, the
expected zone of overlap would be the boundary
zone. The Philippine center of diversity for this
distribution type does not support the idea that
species originating on these plates should overlap
at their boundary, unless there is uneven extinction
occurring in the IMPA. There is evidence that
extinction within the IMPA influenced distribu-
tions, presumably because of land loss (Voris 2000,
Figure 1), cooling (Flemminger 1986), or pre-
sumed increased turbidity during Pleistocene gla-
cial maxima (Springer & Williams 1990). However,
both the Philippines and Wallacea are in the
warmest region of the oceans, the western Pacific
warm pool, which would have undergone
approximately 3C decrease during the last glacial
maximum (Lea et al. 2000). Both areas also have
ample deep water as refuge from lowered sea level
(Figure 1). Although Pleistocene glacial events
undoubtedly excluded shallow water species from
large areas and presented barriers to dispersal
(Flemminger 1986), there is no apparent reason to
predict that extinction would have occurred more
frequently in either the Philippines or Wallacea.
Therefore, the Philippine center of diversity of
Eurasian–Australian endemics suggests predomi-
nate origin on the Eurasian plate. This is consis-
tent with routes of gene exchange away from the
Philippines toward Australian and western Pacific
areas observed in populations of the giant clam
(Benzie & Williams 1997).
The direction of prevailing equatorial currents
implicates Wallacea as the primary area of
accumulation of species that originated on the
Pacific plate. A center of diversity in the Philip-
pines does not coincide with predictions from this
Table 2. Observed and expected frequencies of distribution and habitat types in the Philippine Verde Island Passage (VIP) area of
highest diversity and secondary area of high diversity near Pulau Bintan (PB) in Indonesia for all distributions combined and
separately for shorefishes.
Distribution & Habitat type All taxa Fish
VIP PB Expected VIP PB Expected
Circumtropical + Circumglobal 7.09 6.59 7.92 7.33 6.42 8.58
Eurasian plate 6.39
*
6.41
*
12.81
*
5.75
*
5.80
*
13.79
*
Indo–West Pacific 42.91
*
41.68
*
32.92
*
40.63
*
38.41
*
29.11
*
Indian–Eurasian plate 3.80
*
8.68 7.13
*
3.07
*
9.94 7.22
*
Western Pacific 9.45 6.11 8.86 10.60 6.21
*
9.35
*
Indo-westmost Pacific 20.45 22.46
*
18.93
*
21.41 23.91
*
19.35
*
Eurasian + Australian plates 9.91 8.08 11.44 11.20 9.32 12.60
All Soft bottoms 35.37 37.72 36.74 25.97
*
30.33 30.92
*
Coral Reef 28.57
*
25.45
*
21.99
*
37.07
*
32.09
*
25.16
*
Mostly rocky 14.63 13.35 12.34 7.73 6.63 8.11
Estuarine/fresh water 3.17 3.59 4.46 2.58 2.59 4.49
Demersal hard & soft bottoms 2.13
*
2.28
*
5.80
*
3.27
*
3.21
*
7.43
*
Neritic demersal and pelagic 6.97 7.60 7.81 9.71 10.46 9.53
Epipelagic & deep pelagic 9.16 10.00 10.86 13.68 14.70 14.36
An asterisk (*) indicates a significant (p< 0.001) difference in observed versus expected frequencies. All numbers are percentages
based on: 1736 total taxa at VIP, 1670 total taxa at PB, 2552 total taxa expected distributions possible on the Eurasian plate, 2983 total
taxa possible for expected habitat types, 1009 shorefishes at VIP, 966 shorefishes at PB, 1690 shorefishes expected distributions possible
on Eurasian plate, and 2047 total shorefishes for expected habitat types.
474

hypothesis. Both northern and southern equato-
rial currents enter Wallacea directly (Figure 3).
The northern equatorial current does directly
enter the Philippines but much of this is deflected
as a boundary current and the majority of Pacific
islands are in the South Pacific. To explain a
Philippine center of diversity through accumula-
tion requires invocation of secondary dispersal
and special habitats in the Philippines that
allowed species to thrive there preferentially over
habitats across the Wallacean southern equatorial
current corridor. We have no evidence of such
special habitats in the Philippines.
The central Philippines center of diversity also
appears to falsify predictions from the area of
refuge hypothesis. The available shallow water
habitat and insolation of the Philippines are sub-
stantially less than those of the Indo-Malay part of
the IMPA.
The hypothesis that the IMPA is a center of
origin is not falsified by the central Philippine
epicenter of diversity. However, it is not readily
apparent why geological events that promoted
allopatric speciation may have been more pre-
valent in the Philippines than in Wallacea. Both
regions host a number of sea basins that may have
been isolated during Pleistocene sea level lows
(McManus 1985, Figure 1). The geological origin
of both the Philippines and Wallacea involved
highly varied and concentrated island integration
events (Rotondo et al. 1981, Hall 2002). Molecular
phylogeographic studies of extant marine species
show that population structuring within the IMPA
occurs over distances as short as 300 km (Barber
et al. 2000, Perrin & Borsa 2001). Allopatric spe-
ciation at the scale of Pleistocene isolated seas and
Miocene and Pliocene elements of the Philippines
and Wallacea is therefore plausible.
The Philippine epicenter of diversity suggests
that the geological events that lead to allopatric
speciation within the IMPA were more prevalent
in the Philippines than in Wallacea. The sea basins
around the Philippines do appear to have more
potential for isolation than those around Walla-
cea. During Pleistocene ice ages, the South China,
Sulu, Celebes, and Philippine seas were potentially
isolated from one another through land barriers
(Figure 1) and perhaps cooler surface sea tem-
peratures or restriction of currents. It is also pos-
sible that smaller seas within the central
Philippines became isolated. The potential for
isolation is not as readily apparent across the
Celebes, Philippine, Molucca, and Banda seas
(Figure 1).
Another aspect of the geological history that
could have contributed to a concentration of spe-
cies in the Philippines is the integration of islands
that created the archipelago. Springer (1982) drew
attention to Indonesia as a potential area of
highest diversity because of proximity to frequent
tectonic events associated with plate margins
(Figure 2). Allen & Adrim (2003) point to the
varied geological origins of Indonesia as a
probable cause for the high diversity of coral reef
fishes found there. The amalgamation of the di-
verse elements of Sulawesi is perhaps the most
complicated of all major islands in the IMPA (Hall
2002, Figures 1 and 2). However, Hall’s (2002)
reconstructions of southeast Asia and southwest
Pacific during Cenozoic show the Philippines as
also having a highly complicated geological his-
tory. The Philippines are integrated from at least
three major island systems that were widely sepa-
rated during much of the Cenozoic (Hall 2002,
Figure 2). Most of present-day Mindanao and
eastern Visayas was a shallow sea in the early
Cenozoic at a position equivalent to that currently
occupied by Papua New Guinea. Northern Luzon,
western Visayas, and other elements of Southeast
Asian islands originated at this time near present-
day eastern Borneo. Also in the early Cenozoic,
the Palawan–Calamianes–Mindoro archipelago
was associated with continental Eurasian litho-
spheric plate and were an established island arc
continuous with the present-day island of Taiwan.
Each of these three major elements were displaced
over 1000 km to reach their current locations, with
the Palawan–Calamianes–Mindoro archipelago
moving mostly southeast, the northern Luzon-
western Visayas mass moving mostly north and
the Mindanao-eastern Visayas arc moving mostly
west and north. During this time, the northern
Luzon-western Visayas and Mindanao-eastern
Visayas arcs were associated mostly with the
Philippine lithospheric plate but later coalesced on
what is currently recognized as the Eurasian
lithospheric plate, with the primary subduction
zone shifting to the eastern side of the Philippines.
The amalgamation process created barriers when
the larger islands took shape and potentially
475

separated populations and provided conditions for
allopatric speciation. For example, present-day
Negros collided with the southern tip of Luzon
around 22 million years ago (Hall 2002) and sep-
arated previously contiguous eastern and western
basins. In addition to potentially generating
vicariant events, the accretion of the archipelago
would also have concentrated diversity, assuming
that the different elements of the Philippines
developed their own endemic biotas. This island
integration bioconcentration is consistent with the
conclusions of Santini & Winterbottom (2002)
who consider the derived fauna of the IMPA a
consequence of different faunas rafting into the
region through association with geologically di-
verse terranes. Hall’s (2002) reconstructions of
Wallacea also show complex geological origins
with elements from Eurasian, Australian, and
Philippine plates integrating. However, strong
prevailing southern equatorial currents (Figure 2,
Kennett et al. 1985, Hall 1998) and associated
eddies may have prevented barriers and hence
endemic biotoas from forming, within the basins
of proto-Wallacea during the Miocene and Plio-
cene.
The Philippines center of diversity has a signif-
icant overabundance of both widespread Indo-
Pacific species (Figure 5f, Table 2) and species
found on coral reefs indicating that it could be a
source of widespread diversity. These two catego-
ries are expected to co-occur because coral reef
species tend to be more widespread in their dis-
tributions than other shore species (41% of wide-
spread Indo-Pacific species are coral reef species,
although species in this primary habitat type
comprise 22% of all distribution categories). This
preponderance of widespread species is consistent
with arguments by Springer & Williams (1990)
that extinction may have played an important role
in shaping Indo-Pacific diversity. Local extinction
events would have a higher probability of elimi-
nating limited-range endemics while widespread
species would have had a better chance to survive.
A Philippine center of origin is also consistent with
recent evidence that the IMPA was a center of
origin based on apparent dispersal to marginal
areas distant from an IMPA center of endemism
(Mora et al. 2003).
The potential for Miocene, Pliocene, and Pleis-
tocene vicariance provides a credible explanation
for the primary center of diversity observed in the
Philippines, but fails to explain the secondary con-
centration of diversity near Pulau Bintan of Indo-
nesia (Figure 5a). This area would have
experienced total extinction of local marine popu-
lations several times during the Pleistocene (Fig-
ure 1). Area of refuge, area of overlap, and area of
accumulation hypotheses offer best explanations
for this secondary area of diversity. This secondary
center is in the middle of the largest and longest
standing Cenozoic equatorial shelf area of the
world, the Sunda Shelf, with ample shallow water
habitats as refugia. Unlike the Philippine center,
Pulau Bintan does not have a significantly lower
frequency of Indian–Eurasian plate endemics (Ta-
ble 2) and, therefore, potentially serves as an area of
overlap between Indian and Pacific Ocean faunas.
Similar to the Philippine center of diversity, the
Pulau Bintan center has an overabundance of
widespread Indo-West Pacific species (Table 2). As
a potential refuge, this area of Indonesia could have
accumulated widespread species that originated
anywhere in the Indo-Pacific.
However, if the Philippine center of diversity is a
result of concentrated vicariance, proximity to this
center may have allowed the sunda shelf to flourish
through accumulation. This area of accumulation
hypothesis is different from the vortex hypothesis
proposed by Jokiel & Martinelli (1992) in that the
primary source for species origin is not the remote
Pacific Islands, but the Philippine center of origin.
And, the primary current dispersal vehicle is not
the southern equatorial current but the surface
circulation generated by monsoons. Holocene and
Pleistocene current patterns within the IMPA are
governed primarily by monsoons, although fluc-
tuations in relative strength of paleomonsoons
have varied (Huang et al. 1997). Monsoons would
have also been a major factor for currents in the
region during earlier Cenozoic, although the
IMPA would also have been strongly influenced
by the southern equatorial current prior to its
deflection in the Miocene. The Pulau Bintan region
is at a primary crossroads of monsoon currents
and could have received surface dispersal from the
Philippines during both seasonal monsoons (Fig-
ure 3). This area is bathed by southeasterly direc-
ted circulation from the South China Sea in winter
(Figure 3a), and northwesterly directed circula-
tions around Borneo from Wallacea and the
476

Celebes Sea in the summer (Figure 3b, Huang
et al. 1997).
The Philippine and northern Sumatra areas are
not the only diversity centers identified in our
study. Another pattern that emerges from our
analysis is the apparent marginal centers of
diversity on the Australian and Eurasian plates
(Table 1). Australian plate endemic diversity
reaches a peak in southeastern Queensland, just
south of the Great Barrier Reef. This peak in
diversity is presumably the zone of overlap
between tropical and temperate faunas. Likewise,
an apparent zone of tropical–temperate overlap
appears in Eurasian plate endemic diversity that
reaches a peak at the island of Taiwan. Because
the Philippine center of origin is substantially
north of Wallacea, there is also the suspicion that
it may be a zone of tropical–temperate overlap.
However, the Philippine center of diversity has
significantly fewer Eurasian plate endemics
(Table 2) where most of the northern temperate
biota in this study resides and, therefore, would
not be substantially enhanced by temperate biota.
The identification of the Australian and Eurasian
plate endemic diversity peaks is likely a conse-
quence of inclusion of many non-coral reef species.
Only 22% of the species in our study were con-
sidered primary coral reef species and much of the
temperate fauna that contribute to these marginal
diversity peaks are obviously not primary coral
reef species. A recent study utilizing a large data-
base of both corals and coral reef fishes by
Connolly et al. (2003) shows a peak in diversity
outside the tropical latitudes of the IMPA. This
observation may be a result of a temperate and
tropical overlap of biotas since many reef species
restricted to subtropical and temperate waters in
the western Pacific are also found at extreme lati-
tudinal limits of corals.
A test of the Philippine center of origin would be
a comparative molecular phylogeographic study
(Arbogast & Kenagy 2001, Zink 2002) of marine
populations across the basins proximal to the
Philippines and the major geological elements of
the archipelago. The most parsimonious explana-
tion for diverse taxonomic groups showing the
same distribution pattern of phylogenetically re-
lated taxa is that they have a shared biogeographic
history (Wiley 1988, Avise 2000). Historical
vicariance biogeography relies on well supported
phylogenetic hypotheses and accurate distribution
information to infer relationships between geo-
graphical areas. Santini & Winterbottom (2002)
amassed a dataset of available phylogenetic and
distribution organisms for coral reef organisms of
the Indo-Pacific and implied that the diversity of
the IMPA is a result of amalgamation of varied
geological elements. However, they concluded that
there is a need to develop new methods in order to
further test hypotheses relating to the complex
IMPA region. Phylogeography holds promise
in elucidating both the timing and detail of
biogeographic history through examination of
populations and closely related species and has
already contributed to our knowledge of IMPA
biogeography (Perrin & Borsa, 2001, Barber et al.
2000). Comparative phylogeography incorporates
individual phylogeographic studies of widely sep-
arated taxa in the same geographic area to help
reveal common geohistorical isolating events that
shape population structure and drive speciation
(Arbogast & Kenagy 2001, Zink 2002). Compar-
ative phylogeography as a form of vicariance
biogeography may be crucial to supporting or
falsifying the notion that a Philippine primary
center of allopatric origins constitutes the elusive
unifying hypothesis to explain IMPA biodiversity.
An additional test of the Philippine center of
origin would be expanding the present database to
include all shore fishes of the western Pacific. This
would require extensive taxonomic work and
careful review of distribution records, many of
which are available in museum databases but
would require identification verification by taxo-
nomic specialists before they can become reliably
used. Unfortunately, support for taxonomists has
dwindled perilously in recent years despite their
importance to biodiversity studies (Cotterill 1995,
Carpenter & Paxton 1999, Blaber 2002). Our study
is similar to all previous studies on IMPA bioge-
ography in that it relies on the work of taxono-
mists to examine specimens and distribution
records in order to determine the limits of the
species distributions. Analytical studies that uti-
lized location lists (Bellwood & Hughes 2001,
Mora et al. 2003), range data (Hughes et al. 2002,
Connolly et al. 2003), and general distribution
information (Roberts et al. 2002, Santini & Win-
terbottom 2002) ultimately are derived from the
same taxonomic work that forms the basis for the
477

present study (Carpenter & Niem 1998–2002).
Continued expansion of these databases and more
precise judgments on conservation strategies will
rely on continued rigorous taxonomic work.
Recent analytical studies examining the bioge-
ography or biodiversity of the IMPA have focused
on coral reef biota. These have included distribu-
tional analyses of corals (Wallace 1997), a select
group of coral reef fishes (Mora et al. 2003), a
select group of corals and fishes (Bellwood &
Hughes 2001, Hughes et al. 2002, Connolly et al.
2003), and a broad range of select coral
reef groups (Roberts et al. 2002, Santini &
Winterbottom 2002). An assumption of all these
studies and our study is that the subset of taxa
chosen is representative of overall biodiversity
patterns and that omission of any other taxa will
be compensated by the use of an extensive data-
base. Our study adds the dimension of marine
biota beyond the coral reef. Soft bottom, rocky
intertidal, rocky reef, and estuarine habitats also
harbor an extremely high biodiversity and are
under extreme anthropogenic stress within the
IMPA (Carpenter & Paxton 1999).
In terms of the different hypotheses of IMPA
diversity outlined here, recent analytical studies
have concluded nearly the full range: area accu-
mulation (Connolly et al. 2003), area of refuge
(Bellwood & Hughes 2001), center of origin (Mora
et al. 2003), and a combination of some or all of
the hypotheses (Wallace 1997, Santini & Winter-
bottom 2002). The overall conclusion from our
study is that the diversity found in the IMPA is
likely a combination of many processes, as evi-
denced by the secondary center of diversity we
observed. However the higher number of species
per unit area in the Philippines than elsewhere in
the IMPA indicates that concentrated allopatric
speciation and island integration across the Phil-
ippine archipelago appears to have played an
important role in shaping the diversity of the
IMPA. These hypotheses warrant further testing
through refined vicariance biogeography methods
and molecular phylogeographic approaches.
The identification of the Philippines as the major
center of marine biodiversity is troubling because
of the heightened level of threat to marine envi-
ronments there (Bryant et al. 1998). Roberts et al.
(2002) list the Philippines as being the most highly
threatened center of endemism. The concentration
of limited-range endemics in the Philippines poses
a danger of mass extinctions on a marine scale
similar to endangered Brazilian rainforests. The
concentration of species indicates that there are a
variety of unique biotic communities in the area.
These unique communities would also be under
threat due to habitat degradation. Knowledge of
the underlying processes that govern uneven dis-
tribution of biodiversity is crucial to understand-
ing ecology, and for effective conservation (Gaston
2000, Mora et al. 2003). Understanding factors
that control patterns of endemism and richness
should also help prioritize sites for conservation
even when data are sparse, as they often are in
marine environments. Solely as an example of
peak diversity and endemism, there is ample jus-
tification to prioritize the Philippines for conser-
vation. As a probable epicenter of allopatric
speciation and island integration bioconcentra-
tion, it is imperative to conserve the habitats and
diversity that can help us understand the processes
of evolution that govern biodiversity in the marine
realm. Clearly, marine conservation efforts in the
Philippines warrant special attention.
Acknowledegments
We thank the Food and Agriculture Organization
of the United Nations who made these data
available to us. Special thanks to M. Smith and R.
Waller of Conservation International for help with
geographical information system analyses and
useful comments on the manuscript. We are
grateful to the following for comments and help on
the manuscript: M. Butler, B. Collette, D. Halas,
J. Holsinger, A. Mahon, J. Martin, I, Nakamura,
L. Parenti, W. Resetarits, R. Rose, W. Smith-Va-
niz, and R. Winterbottom. This work is dedicated
to all the taxonomists who provide the distribution
information that forms the basis for this work and
for so much of our information on biodiversity, in
particular, the following authors who provided the
distribution maps in this study: G. Allen,
K. Amaoka, W. Anderson, T. Bagarinao, D.
Bellwood, J. Caruso, M. Carvalho, T. Chan, B.
Collette, L. Compagno, C. Conand, L. Crowley,
J. Dooley, M. Dunning, W. Eschmeyer, R. Feltes,
C. Ferraris, R. Fricke, R. Fritzsche, A. Gill,
O. Gon, E. Gomez, D. Greenfield, I. Harrison, P.
478

Heemstra, D. Hensley, G. Hodgson, P. Hulley,
B. Hutchins, W. Ivantsoff, W. T. Iwamoto,
P. Kailola, C. Kinze, H. Kishimoto, L. Knapp, M.
Kottelat, H. Larson, P. Last, J. Leis, K. Matsuura,
R. Manning, R. McDowall, R. McKay, E. Miclat,
R. Mooi, M. Moteki, T. Munroe, I. Nakamura, P.
Ng, J. Nielsen, M. Nizinski, M. Norman, J. Olney,
N. Parin, J. Paxton, T. Paulus, S. Poss, J. Poutiers,
R. Pyle, J. Randall, A. Rasmussen, A. Reid,
W. Richards, P. Rosenzweig, B. Russell, K. Sakai,
K. Sasaki, H. Senou, D. Smith, W. Smith-Vaniz,
W. Starnes, K. Thiesfeld, G. Trono, R. Vari,
M. Westneat, T. Wongratana, and D. Woodland.
References
Allen, G.R. & M. Adrim. 2003. Coral reef fishes of Indonesia.
Zool. Stud. 42: 1–72.
Arbogast, B.S. & G.J. Kenagy. 2001. Comparative phyloge-
ography as an integrative approach to historical biogeogra-
phy. J. Biogeog. 28: 819–825.
Avise, J.C. 2000. Phylogeography. The History and Formation
of Species, Harvard College, Massachusetts. 447 pp.
Baird, A.H., D.R. Bellwood, J.H. Connell, H.V. Cornell, T.P.
Hughes, R.H. Karlson & B.R. Rosen. 2002. Coral reef bio-
diversity and conservation. Science 296: 1026–1027.
Barber, P.H., S.R. Palumbi, M.V. Erdmann & M. Kasim
Moosa. 2000. A marine Wallace’s line? Nature 406: 692–693.
Bellwood, D.R. & T.P. Hughes. 2001. Regional-scale assembly
rules and biodiversity of coral reefs. Science 292: 1532–1534.
Bellwood, D.R. & P.C. Wainwright. 2002. The history and
biogeography of fishes on coral reefs. pp. 5–32. In: P.F. Sale
(ed.), Coral Reef Fishes, Academic Press, New York.
Benzie, J.A.H. & S.T. Williams. 1997. Genetic structure of giant
clam (Tridacna maxima) populations in the West Pacific is
not consistent with dispersal by present-day currents. Evo-
lution 51: 768–783.
Blaber, S.J.M. 2002. ‘Fish in hot water’: the challenges facing
fish and fisheries research in tropical estuaries. J. Fish Biol.
61A: 1–20.
Briggs, J.C. 1999a. Coincident biogeographic patterns: Indo-
West Pacific Ocean. Evolution 53: 326–335.
Briggs, J.C. 1999b. Extinction and replacement in the Indo-
West Pacific Ocean. J. Biogeogr. 26: 777–783.
Bryant, D., L. Burke, J. McManus & M. Spalding. 1998. Reefs
at Risk: A Map-based Indicator of Potential Threats to the
World’s Coral Reef, World Resources Institute, Washington,
DC. 000 pp.
Carpenter, K.E. 1998. An introduction to the oceanography,
geology, biogeography, and fisheries of the tropical and
subtropical western and central Pacific. pp. 117. In: K.E.
Carpenter & V.H. Niem (eds), FAO Species Identification
Guide for Fishery Purposes. The Living Marine Resources of
the Western Central Pacific, FAO, Rome.
Carpenter, K.E. & V.H. Niem (eds), 1998–2001. FAO Species
Identification Guide for Fishery Purposes. The Living Mar-
ine Resources of the Western Central Pacific, FAO, Rome. 6
volumes. 4218 pp.
Carpenter, K.E. & J.R. Paxton. 1999. The future of systematic
ichthyological research in the tropical Indo-Pacific. pp. 683–
693. In: B. Seret & J.-Y. Sire (eds), Proceedings of the Fifth
Indo-Pacific Fish Conference.
Connolly, S.R., D.R. Bellwood & T.P. Hughes. 2003. Indo-
Pacific biodiversity of coral reefs: deviations from a mid-
domain model. Ecology 84: 2178–2190.
Cotterill, F.P.D. 1995. Systematics, biological knowledge and
environmental conservation. Biodiv. Conserv. 4: 183–205.
Donaldson, T.J. 1986. Distribution and species richness pat-
terns of the Indo-West Pacific Cirrhitidae: support for
Woodland’s hypothesis. pp. 623–628. In: T. Uyeno, R. Arai,
T. Taniuchi & K. Matsuura (eds), Indo-Pacific Fish Biology.
Proceedings of the Second International Conference on Indo-
Pacific Fishes.
Fine, R.A., R. Lukas, F.M. Bingham, W.J. Warner & R.H.
Gammon. 1994. The western equatorial Pacific: a water mass
crossroads. J. Geophys. Res. 99: 25063–25080.
Flemminger, A. 1986. The Pleistocene equatorial barrier be-
tween the Indian and Pacific oceans and a likely cause for
Wallace’s line. pp. 84–97. In: Pelagic Biogeography, UNE-
SCO technical papers in marine science.
Gaston, K.J. 2000. Global patterns in biodiversity. Nature 405:
220–227.
Gloerfelt-Tarp, T. & P.J. Kailola. 1984. Trawled Fishes of
Southern Indonesia and Northwestern Australia, Tien Wah
Press, Singapore. 406 pp.
Hall, R. 1998. The plate tectonics of Cenozoic SE Asia and the
distribution of land and sea. pp. 99–131. In: R. Hall & J.D.
Holloway (eds), Biogeography and Geological Evolution of
SE Asia, Backhuys Publishers, Leiden.
Hall, R. 2002. Cenozoic geological and plate tectonic evolu-
tion of SE Asia and the SW Pacific: computer-based
reconstructions, model and animations. J. Asian Earth Sci.
20: 353–431.
Herre, A.W. 1953. Check List of Philippine Fishes, U.S. Gov-
ernment Printing Office, Washington, DC. 977 pp.
Huang, C.-Y., P.-M. Liew, M. Zhao, T.-C. Chang, C.M. Kuo,
M.-T. Chen, C.-H. Wang & L.-F. Zheng. 1997. Deep sea
and lake records of the Southeast Asian paleomonsoons for
the last 25 thousand years. Earth Planet. Sci. Lett. 146: 59–
72.
Hughes, T.P., D.R. Bellwood & S.R Connolly. 2002. Biodi-
versity hotspots, centres of endemicity, and the conservation
of coral reefs. Ecol. Lett. 5: 775–784.
Jokiel, P. & F.J. Martinelli. 1992. The vortex model of coral
reef biogeography. J. Biogeogr. 19: 449–458.
Kennett, J.P., G. Keller & M.S. Srinivasan. 1985. Miocene
planktonic foraminiferal biogeography and apleogeographic
development of the Indo-Pacific region. In: J.P. Kennett
(ed.), The Miocene Ocean: Paleogeography and Biogeogra-
phy. Geological Society of America Memoir 163: 197–236.
Ladd, H.S. 1960. Origin of the Pacific Island molluscan fauna.
Am. J. Sci. 258A: 137–150.
479

Lea, D.W., D.K. Pak & H.J. Spero. 2000. Climate impact of
late quaternary equatorial Pacific sea surface temperature
variations. Science 289: 1719–1724.
McManus, J.W. 1985. Marine speciation, tectonics, and sea-
level changes in south-east Asia. Proc. Fifth Int. Coral Reef
Symp. 1984 4: 133–138.
Mora, C., P.M. Chittaro, P.F. Sale, J.P. Kritzer & S.A. Ludsin.
2003. Patterns and processes in reef fish diversity. Nature
421: 933–936.
Morgan, J.R. & M.J. Valencia (eds). 1983. Atlas for Marine
Policy in Southeast Asian Seas, University of California
Press, Berkeley. 144 pp.
Myers, R.F. 1989. Micronesian Reef Fishes. Coral Graphics,
Guam. 298 pp.
Palumbi, S.R. 1996. What can molecular genetics contribute to
marine biogeography? An urchin’s tale. J. Exp. Mar. Biol.
Ecol. 203: 75–92.
Pandolfi, J.M. 1992. Successive isolation rather than evolu-
tionary centres for the origination of Indo-Pacific reef corals.
J. Biogeogr. 19: 593–609.
Perrin, C. & P. Borsa. 2001. Mitochondrial DNA analysis of
the geographic structure of Indian scad mackerel in the Indo–
Malay archipelago. J. Fish Biol. 59: 1421–1426.
Randall, J.E. 1998. Zoogeography of shore fishes of the Indo-
Pacific region. Zool. Stud. 37: 227–268.
Roberts, C.M., C.J. McClean, J.E.N. Veron, J.P. Hawkins,
G.R. Allen, D.E. McAllister, C.G. Mittermeier, F.W. Schu-
eler, M. Spalding, F. Wells, C. Vynne & T.B. Werner. 2002.
Marine biodiversity hotspots and conservation priorities for
tropical reefs. Science 295: 1280–1284.
Rosen, B.R. 1988. Progress, problems and patterns in the bio-
geography of reef corals and other tropical marine organ-
isms. Helgola
¨nder Meeresuntersuchungen. 42: 269–301.
Rotondo, G.M., V.G. Springer, G.A.J. Scott & S.O. Schlan-
ger. 1981. Plate movement and island integration – a pos-
sible mechanism in the formation of endemic biotas, with
special reference to the Hawaiian Islands. System. Zool. 30:
12–21.
Santini, F. & R. Winterbottom. 2002. Historical biogeography
of Indo-western Pacific coral reef biota: is the Indonesian
region a center of origin? J. Biogeogr. 29: 189–205.
Simpson, G.G. 1977. Too many lines; the limits of the Oriental
and Australian zoogeographic regions. Proc. Am. Philos.
Soc. 121: 107–120.
Smith, D.G. & J.T. Williams. 1999. The great Albatross Phil-
ippine expedition and its fishes. Mar. Fish. Rev. 61: 31–41.
Spalding, M.D., C. Ravilious & E.P. Green. 2001. World Atlas
of Coral Reefs, University California Press, Berkeley. 424 pp.
Springer, V.G. 1982. Pacific plate biogeography, with special
reference to shorefishes. Smithsonian Contribut. Zool. 367:
1–182.
Springer, V.G. & J.T. Williams. 1990. Widely distributed Pacific
plate endemics and lowered sea-level. Bull. Mar. Sci. 47: 631–
640.
Voris, H.K. 2000. Maps of Pleistocene sea levels in Southeast
Asia: shorelines, river systems and time durations. J. Bioge-
ogr. 27: 1153–1167.
Wallace, A.R. 1860. On the zoological geography of the Malay
Archipelago. J. Linn. Soc. Lond. 4: 172–184
Wallace, C.C. 1997. The Indo-Pacific center of coral diversity
re-examined at species level. pp. 365–370. In: H.A. Lessios &
I.G. Macintyre (eds), Proceedings of the Eighth International
Coral Reef Symposium, Smithsonian Tropical Research
Institute, Panama.
Weber, M. & L.F. deBeaufort. 1913. The Fishes of the Indo–
Australian Archipelago. II, E. J. Brill, Leiden. 404 pp.
Wiley, E.O. 1988. Vicariance biogeography. Annu. Rev. Ecol.
System. 19: 513–542.
Wilson, M.E.J. & B.R. Rosen. 1998. Implications of paucity of
corals in the Paleogene of SE Asia: plate tectonics or center
of origin. pp. 165–195. In: R. Hall & J.D. Holloway (eds.),
Biogeography and Geological Evolution of SE Asia, Back-
huys Publishers, Leiden.
Woodland, D.J. 1983. Zoogeography of the Siganidae (Pisces):
an interpretation of distribution and richness patterns. Bull.
Mar. Sci. 33: 713–717.
Woodland, D.J. 1990. Revision of the fish family Siganidae
with descriptions of two new species and comments on dis-
tribution and ecology. Indo-Pacific Fish. 19: 1–136.
Zink, R.M. 2002. Methods in comparative phylogeography,
and their application to studying evolution in the North
American Aridlands. Integr. Comp. Biol. 42: 953–959.
480