Although the broad general pattern of reef diversity is known detailed knowledge is far from complete for any region. Considering the huge gaps in taxonomic information for many groups, particularly invertebrates, it is convenient to use relatively well known "flagship" groups such as fishes as indicators of overall biodiversity. Fishes are finely adapted to a 

Methods
The present study involves analysis of reef fish distribution in the Indo-Pacific region for the purpose of identifying hotspots applicable to marine conservation priority. The analysis is restricted to shallow reef fishes, which are defined here as species that are wholly or mainly confined to coral reefs and intermingled habitats (sand and rubble patches, adjacent seagrass beds, etc.) less than 60 m deep. The 60-metre depth is near the lower limit of compressed air scuba diving, and also approximates the limit of most reef-building corals. 

More than 100 families of fishes are represented on Indo-Pacific coral reefs, but a relatively small portion of these constitute the majority of species. The bulk of the fauna is composed of the 29 families listed in Table 1.

This is especially true in the species-rich Indo-Australian Archipelago, where these families routinely constitute 85-90 percent of the total reef fish fauna. Of particular importance are the Gobiidae, Labridae, Pomacentridae, Apogonidae, Blennidae, Serranidae, Chaetodontidae, and Acanthuridae, which collectively make more than 50 percent of the fishes on any given reef.

Results
  Species Diversity
A review of relatively recent published faunal checklists of colleagues was undertaken and combined with the author's personal survey data for Australia, Indonesia, Malaysia, Singapore, Philippines, and Papua New Guinea. In order to maintain consistency, only coral-reef fishes (see definition above) were considered. The information from this analysis was used to construct a generalized "species contour" map for the Indo-Pacific region (Fig.3). 

  Endemism
Results of the analysis reveal both regional and local patterns of endemism. The largest definable endemic 

Table 3. Top 10 areas for endemic reef fishes based on absolute numbers.

The reader is reminded that the values in this table were determined from selected indicator groups only. The total number of endemics for all families from each area is estimated to be approximately twice this total judging from two areas (Hawaii and Marquesas) where accurate total figures are available (Randall, pers. comm.).

Another way to evaluate the endemism data involves ranking various areas on the basis of their percentage of endemic species in relation to the total reef-fish fauna (Table 4). The highest ranking areas using this

Table 4. Top 10 coral-reef hotspots based on percentage of endemic reef fishes for the total fish fauna.

The values shown in Table 4 are based on the overall fauna of each location, rather than being restricted to the indicator groups.

* The ranking is determined by dividing the total endemic species (E) by the area (km2)

Species Diversity
Results for the analysis of species diversity are shown in Figure 3. The pattern presented here for reef fishes agrees well with other wide-ranging groups such as corals and molluscs (Veron 1995; Briggs 1999), and no doubt reflects diversity trends for other Indo-Pacific marine organisms. Basically, the center of diversity is the area occupied by eastern Indonesia and the southern Philippines. There is an attenuation of the fauna in all directions away from the center, although this trend is less pronounced in a westward direction. Numerous families of reef fishes exhibit similar patterns, for example the highly speciose Damselfishes (Allen 1975, 1991).

Most conservation actions are initiated at the country level or regional level, involving several adjacent countries. Therefore, it is useful to rank individual countries on the basis of overall species diversity (Table 6). The total for each country may at first appear very conservative, compared to various published works involving these areas. However, these figures apply to coral-reef fishes as defined above and do not reflect the total inshore fish fauna.

The majority of tropical marine families of fishes and invertebrates demonstrate their greatest concentration of species within the area encompassing Australia, Indonesia, Philippines, and Papua New Guinea (Briggs 1999). This region is sometimes referred to by zoogeographers as the Indo-Australian Archipelago or East Indian Triangle (Briggs 1999). The area (Fig. 2) has also been tagged
the "Coral Triangle" by conservation biologists (Werner and Allen 1998). Although its extraordinary biological rrichness is widely acknowledged, there is still a lack of basic taxonomic information for most locations within this huge region. However, unpublished field research on reef fishes by the author and various colleagues indicates the "epicenter" of the Coral Triangle, may well be situated in eastern Indonesia.

Although there are no authenticated cases of recent extinctions involving reef fishes, coral reefs and their myriad inhabitants are increasingly at risk as a result of human activities. Greenhouse emissions and resultant sea temperature increases, destructive fishing practices, and haphazard logging /agricultural methods that lead to erosion and consequent sedimentation of reefs are among the most obvious factors. There is an urgent need to establish an effective network of marine protected areas throughout the Indo-Pacific and other tropical seas. Admittedly, there is still much to learn concerning "source-sink" relationships of marine larva, but if we wait for researchers to provide this information it may already be too late for many species. It would be far better to implement a strategic network of MPAs as soon as possible. If a conservation "umbrella" was based on the 35 Indo-Pacific hotspots recognized in Figure 2 it is conservatively estimated that at least 90 percent of all fish species in the region would receive at least some measure of protection. Most importantly, the rarest elements of the fauna, the highly restricted endemics, would receive maximum attention following this approach.

Acknowledgments
I am grateful for the assistance of my Conservation International colleagues R. Mittermeier and T. Werner, who encouraged me to undertake this analysis and offered many useful suggestions. I also acknowledge the numerous contributions of J. Randall and V. Springer to our knowledge of Indo-Pacific fish zoogeography. Their works were crucial to this study. Thanks are also due C. Roberts, F. Wells, and J. Veron for their discussions and input concerning Indo-Pacific zoogeography.