Geographic patterns, ecological gradients, and the maintenance of tropical fruit bat diversity: the Philippine model

Bat Biology and Conservation Edited, by Thomas H. Kunz and Paul A. Racey SMITHSONIAN INSTITUTION PRESS Washington and London 24 Geographic Patterns, Ecological Gradients, and the Maintenance of Tropical Fruit Bat Diversity The Philippine Model RUTH C. B. UTZURRUM Like most tropical areas worldwide, Philippine rainforests have been lost to exploitation (resource extraction and conversion), monotypic reforestation, and replacement of native species with exotics. The archipelago's forests, estimated to cover 80% of its total land area in the 1800s, have been reduced to 12% with most of the decline occurring within the last six decades (Myers 1988; Kummer 1990). Given the archipelagic nature of the Philippines, anthropogenic degradation and fragmentation of the forest habitat could very well exacerbate effects of isolation on populations of plants and animals. By the late 1980s, 2 of the 26 species of Philippine fruit bats (Family Pteropodidae) were reported to be extinct (Acerodon lucifer and Dobsonia chapmani; Heaney and Heideman 1987). A recent rediscovery of a population of A. lucifer on Boracay Island (off Panay) awaits confirmation (E. E. Maro, personal communication; but see Heaney et al. [n.d.] questioning the species status of Acerodon lucifer). It is recognized that at least 5 of the 24 extant species (including Acerodon jubatus, A. leucotis, Eonycteris robusta, Nyctimene rabori, and Pteropus leucopterus) are seriously threatened by habitat destruction and hunting (Heaney and Utzurrum 1991; Mickleburgh et al. 1992; Utzurrum 1992). In this chapter, I summarize results of elevational transect inventories conducted in recent years and use these data to assess the effects of habitat fragmentation on the maintenance of Philippine fruit bat diversity. Three questions are examined: (1) What are the patterns of species diversity in unffagmented (i.e., local gradients) and naturally fragmented areas (i.e., biogeographic) landscapes? (2) What impact will habitat fragmentation have on the maintenance of diversity? and (3) Do macroand microgeographic patterns of community associations provide practical insights into conservation? The conservation implications of these questions are (1) to examine properties of populations and community assemblages that could define their response to effects of habitat fragmentation; (2) to identify measures that may mitigate effects of fragmentation given limited information; and (3) to assess research needs for the development of sound conservation strategies. Primary emphasis is given to the patterns of local gradient (rather than biogeographic patterns) because these bear the most relevance to the discussion of habitat fragmentation and its impact on the maintenance of species diversity. 342 Maintenance of Fruit Bat Diversity 343 Data Sources and Methods of Analysis Study Sites The Philippine Archipelago, lying between 4°40' N to 2i°50' N latitude and 116°50' E to 136°35' E longitude, is tropical throughout its range. Habitat diversity is more marked along altitudinal than across latitudinal or longitudinal gradients, although variability in the local flora occurs in association with climatic subregions (Alcala 1976). Local gradients encompass three primary vegetation zones: lowland dipterocarp forest, montane forest, and mossy forest (dipterocarp, lower montane, and upper montane rain forest, in Whitmore 1984). The specific elevations at which these forest types occur vary among mountain sites, largely as a result of differences in maximum elevation and the amount and distribution of local rainfall (Whitmore 1984). Mountains that are at least 1,500 m in elevation support well-developed primary vegetational types over wide altitudinal ranges, and transition zones between major forest types exist as distinct bands (e.g., Mt. Guitinguiting, Sibuyan [Goodman and Ingle 1993]; Mt. Guisayawan, Negros [Heaney et al. 1989]; and Mt. Isarog, Luzon [Rickart et al. 1991]). Small mountains exhibit compression of vegetational zones, which thus occur at relatively lower elevations in what is known as the "Massenerhebung effect (e.g., Mt. Pangasugan, Leyte [Heaney et al. 1989; Rickart et al. 1993], and Mt. Konduko, Biliran [Rickart et al. 1993]) (Grubb and Whitmore 1966; Frahm and Gradstein 1991). Mt. Talinis, centered approximately at 9°16' N, 123 12 E on Negros Island, extends upward to 1850 m and exhibits well-defined lowland, montane, transitional montane-mossy, and mossy forests (see following section). The degree of anthropogenic disturbance of natural habitats varies greatly among these sites. In most locations, primary lowland forest is absent below 500 m, although limited natural or replanted secondary forest may exist below this elevation. Mosaic patches of disturbance within primary forests resulting from small-scale timber extraction or shifting agriculture are, likewise, a common feature of the forests, even in areas designated as national parks. Field Methods and Data Sources The principal information on biogeographic patterns of distribution is that reported by Heaney (1986, 1991a) and Heaney and Rickart (1990). Additional details are derived from Heaney and Rabor (1982), Heaney et al. (1991), Utzurrum (1992), Goodman and Ingle (1993), and Vincguerra and Mtiller (1993). The summary presented here does not take into account the proposed change in the species status of Acerodon lucifer (Heaney et al. n.d.). Patterns of elevational gradients are summarized from a survey I conducted on Mt. Talinis (Negros Island) and from published studies on Biliran (Mt. Konduko: Rickart et al. 1993), Leyte (Mt. Pangasugan: Heaney et al. 1989; Rickart et al. 1993), and Negros (Mt. Guinsayawan: Heaney et al. 1981, 1989; Heideman and Heaney 1989) (Figure 24.1). Where available and relevant, unpublished data from recent studies are provided. Data from the 1990 elevational transect study on Mt. Talinis, Negros Island, provided the focal point of analysis on elevational trends. Although previous studies on a neighboring mountain system, Mt. Guisayawan (Heaney et al. 1989; Heideman and Heaney 1989), were more extensive in scope and effort, a greater proportion of the netting was done at sites representing a single habitat type, and the full range of elevational sampling was not undertaken within the same year. Thus, results of the Mt. Guisayawan studies may be less comparable with data from more standardized sampling used in recent surveys conducted elsewhere in the Philippines. Standardized surveys of the fruit bat fauna along elevational transects typically involved running a series of understory mist nets, during three to five nights, within a 50to 100-m elevational band at sites corresponding to each major habitat type (Rickart 1993). On Mt. Talinis, I ran an elevational transect between May and July 1990 using six netting sites. The sites were (1) an area of low-intensity agriculture and secondary growth (500 m), characterized by stands of coconuts, coffee, and plots of cultivated vegetables and flowers, interspersed with patches of shrubs (Melastoma spp.), sawgrass (Imperata sp.), and scattered wild figs (Ficus spp.); (2) a naturally regenerated secondary forest (500 m) with remnant anthropogenic plants such as banana and abaca (Musa spp.), avocado (Persea americana), and jackfruit (Artocarpus hetephyllus); (3) primary lowland forest (750 m) punctuated by small (<0.5-ha) disturbed patches in various stages of regeneration; (4) primary montane forest (1,100 m) relatively free of and distant from disturbance, (5) transitional montane-mossy forest (1,250 m) with elements of previous anthropogenic disturbance (e.g., Musa spp. and Bambusa sp.); and (6) primary mossy forest (1,625 m) at the summit of the ridge system on which the five other sampling sites were located. At each sampling location, ten nylon mist nets (12 m long X 2.6 m high) were run from 1800 to 0600 for five consecutive nights. External measurements (size measurements and body mass) were recorded for each fruit bat captured. All samples were identified to species following Heaney et al. (1987) and Ingle and Heaney (1992), and assessed for age and 344 R. C. B. UTZURRUM Pleistocene islands as defined by current 1 20 m bathymetric line 0 200 1 1 1 1 1 II Kilometers •18° Luzon Faunal Region 1 6:8:1 Greater Mindoro 1 1:6:1 Palawan Faunal Region 7:1:1 NegrosPanay Faunal Region Figure 24.1. Extent of Pleistocene land connections (shaded areas, which correspond to the current 120-m bathymetric line) in comparison to the current topography (solid lines) of the Philippines. The first of the three numbers following each Pleistocene region name corresponds to the total number of fruit bat species present; the second number, the subset of this total that are endemic to the Philippines; and the third, the number of endemic species that are unique to the faunal region. Letters indicate the principal locations of study sites mentioned in the text: C, mountain on Catanduanes; G, Mt. Guisayawan, Negros; Gt, Mt. Guitinguiting, Sibuyan; I, Mt. Isarog, southeastern Luzon; Ki, Mt. Kitanglad, Mindanao; K, Mt. Konduko, Biliran Island; P, Mt. Pangasugan, Leyte; T, Mt. Talinis, Negros. reproductive status by methods modified from Heideman (1987). Voucher specimens were deposited in the Field Museum of Natural History (Chicago), Philippine National Museum (Manila), Silliman University Museum of Natural History (Dumaguete City), and the teaching collection of the Department of Biology, Boston University. Data Analysis Analyses of data sets other than from Mt. Talinis included calculations of species richness (S, the total number of different species captured at each site), total abundance (expressed as numbers of bats captured per net-night), and relative abundance (number of bats of a given species captured per net-night). Similar treatments were applied to the Mt. Talinis data set. The Shannon-Wiener index of diversity (H') was also calculated for the Mt. Talinis (this chapter) and Mt. Pangasugan (Heaney et al. 1989; Rickart et al. 1993) elevational transect data. Within-mountain and betweenmountain diversity indices were compared using a t-test (Magurran 1988). Large flying-foxes (Acerodon and Pteropus species) are customarily excluded from data analysis beMaintenance of Fruit Bat Diversity 345 Table 24.1 Distributions of the 26 Species of Philippine Fruit Bats (Pteropodidae) across Pleistocene Faunal Regions (sensu Heaney 1986) as Updated from Heaney (1991) Distribution and endemism Spec' Species widespread in Indo-Australia (6/26; 23%) Species shared with nearby archipelagos; of limited distribution in the Philippines (4/26; 15%) Endemic species widespread in oceanic Philippines (6/26; 23%) Endemic species on two or more Pleistocene islands (2/26; 8%) Endemic species on only one Pleistocene island (8/26; 31%) Cynopterus brackyotis Eonycteris spelaea Macroglossus minimus Pteropus hypomelanus Pteropus vampyrus Rousettus amplexicaudatus Dyacopterus spadiceus Megaerops wetmorei Pteropus dasymallus3 Pteropus speciosus Acerodon jubatus Eonycteris robusta Haplonycteris fischeri Harpyionycteris whiteheadi Ptenochirus jagori Pteropus pumilus Pteropus leucopterusuz Nyctimene rabori4-6 Acerodon leucotis Acerodon lucifer Alionycteris paucidentata Dobsonia chapmani Haplonycteris sp.4 Otopteropus cartilagonodus Ptenochirus minor Pteropus sp.5 "Superscript numbers indicate sources of updates: (1) Heaney and Rabor 1982; (2) Heaney et al. 1991; (3) Utzurrum 1992; (4) Goodman and Ingle 1993; (5) Heaney 1993, and unpublished; and (6) Vinciguerra and Muller 1993. cause they are not reliably sampled in understory net sets (Heaney et al. 1989; Heideman and Heaney 1989; Rickart et al. 1993). However, I included the smallest of the flyingfoxes, Pteropus pumilus, in the analysis of the Mt. Talinis data. The nets for the Mt. Talinis transect were set on narrow ridges and were effective in capturing flying-foxes that commute over ridgetops. Results Pteropodids presently known from the Philippines range in size from 16 g (e.g., Alionycteris paucidentata; Macroglossus minimus) to more than 1 kg (e.g., Acerodon jubatus and Pteropus vampyrus). Of the 26 species present, 16 (62%) are restricted to the Philippines, including 6 species in the endemic genera: Alionycteris (1), Haplonycteris (2), Otopteropus (1), and Ptenochirus (2) (Heaney et al. 1987; Heaney 1991a; Utzurrum 1992). Biogeographic Distribution Patterns Biogeographic analysis of species richness among Philippine fruit bats shows distributions concordant with land masses formed during lowering of sea levels in the Pleistocene (Heaney 1991a). Six (23%) of the nonendemic species are widespread within the Philippines and in the Indo-Australian region; the other four (15%) are restricted within the Philippines and are shared with nearby islands (Table 24.1; see also Figure 24.1) (Heaney et al. 1987; Heaney 1991a). Among the endemic species, three basic patterns of geographic distribution emerge: (1) species that are widespread in oceanic Philippines (six); (2) species that are shared at least between two Pleistocene islands (two); and (3) species that are confined to only one Pleistocene island (eight) (Table 24.1; Figure 24.1) (Heaney and Rabor 1982; Heaney 1991a 1993; Utzurrum 1992; Goodman and Ingle 1993). 346 R. C. B. UTZURRUM Data from well-inventoried present-day islands reveal a significant positive relationship between species numbers and island size (Heaney 1991a). However, this relationship does not hold true for endemic species. Luzon Island (108,171 km2) is the largest island to hold an endemic (i.e., Otopteropus cartilagonodus). The islands of Mindanao (99,078 km2; Alionycterispaucidentata), Negros (13,670 km2; Dobsonia chapmani), Panay (12,300 km2; Acerodon lucifer), Palawan (11,785 km2; Acerodon leucotis), and Mindoro (9,735 km2; undescribed Pteropus sp.) also have one endemic species each (Heaney 1991a, 1993; Heaney et al. n.d.). Sibuyan Island (463 km2; undescribed Haplonycteris sp.), an oceanic island with no historical connection to any of the Pleistocene islands, is the smallest Philippine island currently known to have an endemic bat species (Utzurrum 1992; Goodman and Ingle 1993). The high degree of species overlap within and among Pleistocene islands suggests the importance of overwater colonization in shaping diversity and distribution (Heaney and Rickart 1990; Heaney 1991a). Measures of gene flow confirm these patterns (Peterson and Heaney 1993). Pleistocene land-bridge islands intermediate in size between Negros and Sibuyan that lack endemic species (e.g., Dinagat and Leyte) further support the importance of Pleistocene land connections, or conversely the lack thereof, in shaping speciation events within the archipelago. Species with disjunct distributions pose an interesting puzzle in our understanding of these faunal patterns and the mechanisms that shaped them. Pleistocene land connections among islands do not account for the disjunct distribution of Nyctimene rabori (in Cebu, Negros, and Sibuyan islands only; Heaney et al. 1987; Goodman and Ingle 1993, Vinciguerra and Miiller 1993), or of Pteropus leucopterus (in Catanduanes, Dinagat, and northern Luzon; Heaney and Rabor 1982; Heaney et al. 1987, 1991). These cases suggest processes that have influenced extinction events in the past—or may simply reflect information gaps that need to be filled from more thorough inventories. Elevational Gradients in Species Diversity and Abundance: General Trends General patterns have emerged from elevational transect surveys: (1) species richness reaches its maximum in primary lowland forest and declines with elevation; (2) total abundance is highest in disturbed areas and decreases with elevation in forest habitats; and (3) species assemblages in disturbed areas tend to be characterized by the presence of geographically widespread species, whereas endemic species tend to be associated with forest habitats (Heaney et al. 1981, 1989; Heideman and Heaney 1989; Rickart et al. 1993). Endemic species are sometimes found in low numbers in moderately disturbed habitats provided that these are adjacent (within 1 km) to primary habitats (Heideman and Heaney 1989; Rickart et al. 1993). Despite the abundance of fruit bats in urban orchards and agricultural areas far removed from forested sites, endemic species have not been recorded at these locations (Heideman 1987; Heaney et al. 1989; Heideman and Heaney 1989; Rickart 1993), with the exception of Pteropus pumilus (R. C. B. Utzurrum, unpublished records of captures in Dumaguete City [1986] and Siaton [1992] on Negros). In most cases, species at high elevations represent a subset of the lowland community, indicating a lack of highelevation specialists. In 1992 and 1993, Alionycterispaucidentata was captured in considerable numbers on Mt. Kitanglad, Mindanao, but only at elevations above 1,500 m in montane and mossy forest (L. R. Heaney, personal communication). Elevational Gradients on Mt. Talinis, Negros Island Of the 15 species of fruit bats occurring in Negros Island, 9 were netted in this study (Table 24.2). The capture of Cynopterus brachyotis at 1,250 m is a new elevational record for the species; all other species have been recorded at elevations similar to or greater than in the present study (Heaney and Heideman, unpublished data). Missing from the samples were 3 species of flying-foxes (Acerdon jubatus, Pteropus hypomelanus, and P. vampyrus) not expected to be captured in understory nets, an uncommon endemic species, Eonycteris robusta (Utzurrum 1992), and the reportedly extinct Dobsonia chapmani (Heaney and Heideman 1987). The total numbers of fruit bats captured were highest at the agricultural site (3.24 bats/net-night) and decreased with elevation in forest (from 1.9 bats/net-night in lowland forest to 0.22 bats/net-night in mossy forest; Table 24.2). This overall trend was true for all the nonendemic species as well as the endemic Ptenochirus jagori (Table 24.2). All other endemic species were uncommon or absent at the agricultural site. Instead, they were found in higher numbers in lower elevation forest (lowland or montane), although Nyctimene rabori was relatively uncommon even in forest habitat. These patterns of abundance and distribution are consistent with a previously observed ecological dichotomy between endemic and nonendemic species (Heaney et al. 1989). As predicted by Heaney et al. (1989), species richness was highest in lowland primary forest (S = 9) and lowest in mossy forest (S = 3) (Table 24.2). Unexpectedly high levels of species richness and abundance were recorded at the montane-mossy transitional zone. Maintenance of Fruit Bat Diversity 347 Table 24.2 Summary of Captures of Fruit Bats (Pteropodidae) on Mt. Talinis, Negros Oriental, May-July 1990 Site (and elevation) A B c D E F Species (500 m) (500 m) (725 m) (1,100 m) (1,250 m) (1,625 m) Cynopterus brachyotis 0.74 0.68 0.54 0.08 0.10 0.02 1.04 0.10 0 0 0.16 0.04 0.02 0 0 0 0 Eonycteris spelaea 0.38 0.14 0.02 0 0.10 0.02 1.04 0.10 0 0 0.16 0.04 0.02 0 0 0 0 * Haplonycteris fischeri 0.04 0.10 0.26 0.42 0.10 0.02 1.04 0.10 0 0 0.16 0.04 0.02 0 0 0 0 *Harpyionycteris whiteheadi 0 0 0.04 0 0.10 0.02 1.04 0.10 0 0 0.16 0.04 0.02 0 0 0 0 Macroglossus minimus 0.94 0.20 0.54 0.10 0.36 0 0 0.16 0.04 0.02 0 0 0 0 *Nyctimene rabori 0 0 0.06 0.06 0.02 0 0.04 0 0 0 0.16 0.04 0.02 0 0 0 0 'Ptenochirus jagori 0.36 0.04 0.16 0 0.02 0 0.04 0 0 0 0.16 0.04 0.02 0 0 0 0 *Pteropus pumilus 0.02 0 0.22 0 0.02 0 0.04 0 0 0 0.16 0.04 0.02 0 0 0 0 Rousettus amplexicaudatus 0.76 0.08 0.06 0 0.02 0 0.04 0 0 0 0.16 0.04 0.02 0 0 0 0 All nonendemics 2.82 1.10 1.16 0.18 0.48 0.02 All endemics (*) 0.42 0.14 0.74 0.48 1.20 0.20 AH species 3.24 1.24 1-90 0.66 1.68 0.22 (S = 7) (S = 6) (S = 9) (S = 4) (S = 7) (S - 3) Notes: Data are given as number of bats captured per net-night Each site had a total netting effort of 50 net-mgjus. Site designations: A, mixed agriculture/secondary growth; B, secondary lowland forest; C, primary lowland forestD montane forest; E, transitional montane-mossy; F, mossy forest. An asterisk denotes endemic species. Discussion Biogeographic Information and the Design of a System of Protected Areas Patterns of biogeographic distribution of species are relevant to fruit bat conservation at two levels. First, they provide a biological basis for the selection of important sites for protection. Second, island size and species diversity relationships revealed from biogeographic analysis provide estimates of areal requirements for the maintenance of species diversity based on estimated rates of colonization and extinction (Heaney 1986). Given the high degree of similarity in the composition of fruit bat assemblages between Pleistocene faunal regions and among islands of a region, a minimum of eight protected areas, to be located in the current islands of Luzon (one in the northern tip and a second in the southeastern peninsula), Mindanao, Negros, Panay, Palawan, Mindoro, and Sibuyan, may theoretically protect all 26 species (Table 24.3). Interestingly, the patterns of species richness and levels of endemism seen in fruit bats relative to Pleistocene faunal regions are concordant with those of other vertebrate groups, including nonvolant mammals (Heaney 1986, 1993; Heaney and Rickart 1990), birds (Dickerson 1928; Dickinson et al. 1991), and amphibians and reptiles (Brown and Alcala 1970; Hague et al. 1986). These zoogeographic patterns also overlap well with phytogeographic patterns of diversity and endemism (D. A. Madulid, personal communication; Heaney 1993). Thus, a parks system modeled on the biogeography of these more speciose vertebrates and plants would subsume protection of the fruit bats. In the recent Integrated Protected Areas Systems (IPAS) initiative toward the redevelopment of the Philippine parks system (IUCN 1991), recommended priority areas based on mammalian and floristic diversity patterns have led to the inclusion of sites in Luzon, Negros, and Mindanao (Heaney 1993). This plan will incorporate habitats for 85% (22 of 26) of all fruit bats, including 75% (12 of 16) of the endemic species (see Table 24.3). It should be emphasized that areal size requirements based on estimated rates of colonization and extinction should be treated as conservative guidelines, given that these and other evolutionary processes have occurred in a context of habitats (i.e., continuous expanse of forests) largely different from present-day conditions (i.e., discontinuous patches of forests). Thus, whenever possible, the largest continuous area of suitable habitat available at each priority site should be chosen. There are perceived political barriers against the designation of small islands for protection even when strong biological reasons exist (see Utzurrum 1991 for discussion). 348 R. C. B. UTZURRUM Table 24.3 Theoretical Percentages of Species That Would Be Protected by a Designated Park in the Philippine Islands All species Endemic species (total, 26) (total, 16) Island Cumulative Cumulative (and area) No. % No. % Mindanao (99,078 km2) 17 65% 8 50% Luzon (108,171 km2) +3 77% +2 62% Negros (13,670 km2) +2 85% +2 75% Panay (12,300 km2) + 1 88% + 1 81% Mindoro (9,735 km2) + 1 92% + 1 88% Palawan (11,785 km2) + 1 96% + 1 94% Sibuyan (463 km2) + 1 100% + 1 100% Notes: Mindanao is ranked first because it has the most species. Luzon and Negros follow, based on the number of species that these islands will add to the theoretical protected pool. The last four islands are ranked on the basis of conservation priority of the additional species unique to each theoretical reserve (see Heaney 1993; Heaney and Utzurrum 1991; Utzurrum 1992). Four of six forest parks designated by the Integrated Protected Areas System (IPAS) will be located in Mindanao (Mt. Kitanglad), Luzon (Palanan Wilderness Area and Subic Bay), and Negros (Mt. Kanlaon). Such may be the case for Sibuyan Island. At least five undescribed species of mammals were recently (i.e., since 1990) discovered on this island (one fruit bat, Haplonycteris sp., and four murids, Apomys (2), Chrototnys (1), and Tarsomys (1); Goodman and Ingle 1993). Its degree of isolation and relative state of "underdevelopment," however, elements inherently favoring species persistence, could work against its selection for protection under the national parks scheme. Ongoing political machinations exist that exploit the rich timber resources on Mt. Guitinguiting (N. Ingle, personal communication), but the island's relative isolation from trade and communication precludes the national visibility that could highlight a need to include it in a short list of priority parks. Instead, protection of areas such as Sibuyan, and other sites supporting single island endemics, may depend on the development of a network of secondary regionally managed conservation parks that complements the national parks system (Utzurrum 1991). There are many potentially species-rich and biogeographically interesting areas that remain relatively unknown in the Philippines. Recent efforts to systematically inventory local mammalian fauna have been concentrated mostly on a latitudinal band extending from southern Luzon (in the north) to northern Mindanao (in the south) (see Figure 24.1). Although Palawan has been very attractive to researchers of birds (Dickinson et al. 1991) and, in part, of mammals (Heaney 1986), its bat fauna is poorly studied (Heaney 1991a). The northernmost regions of Luzon (especially the northeastern border) still support extensive forests, yet most of the recent surveys in the area focused primarily on birds (Mallari and Jensen 1993). These research trends reflect in part the opportunistic nature of research work in the Philippines, in terms of funding, political stability, and the expertise available. Habitat Affinity and Its Implications for Conservation It is widely recognized that forests are essential for the conservation of biodiversity in tropical regions. More importantly, the design and management of protected areas should incorporate provisions for different affinities among species for gradients in vegetational structure and composition and abiotic conditions throughout the local range of a forest habitat. Thus, a basic understanding of local patterns of species distribution and species-habitat associations is of utmost importance. Lowland forests are essential for the maintenance of maximum local and, therefore, overall diversity of Philippine fruit bats (see Results). This requirement necessitates the inclusion of lowland forest into parks or reserves. Forests at low elevations may also be critical for the persistence of higher-elevation forests. These upper-elevation habitats, in turn, may be integral to the maintenance of lowland forest diversity. Montane forest habitats on Philippine mountains have experienced episodes of expansion and retraction associated with climate changes in the Pleistocene (Heaney 1991b). Although the relevance of these historical changes in vegetational cover to speciation events is more apparent for rodents (Heaney and Rickart 1990), they may have influenced speciation events or patterns of habitat specialization in fruit bats as well. The distributional pattern of Alionycteris paucidentata on Mt. Kitanglad indicates that other upper-elevation habitat specialists among the fruit bats are likely to occur in association with broadly distributed and well-developed montane-mossy forests. As on Mt. Kitanglad, these conditions may occur in mountains exceeding 1,500 m in elevation. However, few Philippine mountains bearing elevations close to 2,000 m or greater have been surveyed. Furthermore, we have a very limited understanding of the nature and dynamics of microgeographic preferences among and within fruit bat species. While surveys indicate lowland forests as the local centers of species richness, they do not necessarily reveal to what extent adjoining tracts of forests contribute to the maintenance of this diversity. For example, preliminary analysis of within-species differences in the elevational distribution of Otopteropus cartilagonodus on Mt. Isarog and Zambales revealed elevational segregaMaintenance of Fruit Bat Diversity 349 tion between males and females of the species, a mechanism that may reduce intersexual competition (Ruedas et al. 1994). The importance of montane and mossy forest habitats for the maintenance of biodiversity is even more crucial for nonvolant mammals. Philippine murid rodents exhibit peaks in species richness and endemism in montane and mossy forests (Heaney and Rickart 1990; Rickart et al. 1991). In many of the mountains surveyed, patterns of avian diversity mirror that of fruit bats, with lowland forest as the locus of maximum species richness. However, endemism itself may be centered in the upper-elevation forests, as was documented in the northern Sierra Madre (Mallari and Jensen 1993). Together, these findings indicate the need for a full elevational complement of forest habitats in any designated conservation site if maximum species diversity is to be protected. Impacts of Forest Fragmentation on Fruit Bat Diversity Conservation of biodiversity must address whether and how contemporary degradation and fragmentation of habitats within islands will influence ecological processes and, thus, local patterns of species assemblage and distribution. If strong preference for forest habitats inhibits fruit bat movements over fragmented landscapes, despite their inherent vagility, then mechanisms of species maintenance and geographic structuring may be affected. Indeed, genetic variation in Cynopterus brachyotis (a nonendemic) and Haplonycteris fischeri (an endemic) suggests that the reduced gene flow seen in the latter species relates, in part, to its greater affinity for specific habitats (Peterson and Heaney 1993). The occurrence of an endemic species of Haplonycteris on Sibuyan Island demonstrates evolutionary stability in small isolated populations (Peterson and Heaney 1993). Thus, it is difficult to predict whether increased fragmentation of once continuous populations (within islands) in the recent past will shift evolutionary processes toward the negative trajectories associated with increasingly smaller population sizes (see Lande 1988). Elevational gradients in fruit bat diversity and abundance on Negros and Leyte islands illustrate how habitat fragmentation or degradation affects local community structure. Negros and Leyte were parts of two different Pleistocene islands but exhibit a moderately high degree of similarity in their fruit bat fauna (Heaney 1991a). Both islands support 13 extant fruit bat species, with 12 species shared by both areas (Heaney et al. 1989). Whereas Negros Island has Nyctimene rabori, Leyte has Ptenochirus minor, both of which are endemic and share an affinity for forest habitats (Heaney et al. 1989; Rickart 1993). Thus, they may be considered ecological equivalents for the purpose of comparing general patterns between the two islands. On Mt. Talinis (Negros), primary forest was absent from below the 500-m elevation but extended over a greater distance to a summit of 1,850 m. Disturbed patches within the primary forest were not uncommon. The primary forest in Mt. Pangasugan (Leyte) was compressed over a narrower elevational range because the summit was lower (1,150 m). Disturbances within forests were minimal and confined within the lower 300-m elevation; lowland forest graded into agricultural areas below 200 m. Mt. Talinis has sharper topographical features consisting of deep, steep gulleys bisecting sharp narrow ridges. Thus, the interfaces between lowand high-elevation zones, and between forest and agricultural or secondary growth areas, were greater there than on Mt. Pangasugan. On Negros, endemic species, such as Haplonycteris fischeri and Ptenochirus jagori, were absent at elevations below 500 m, with the exception of Pteropuspumilus (Figure 24.2). This lack coincides with the absence of forest habitat below this elevation. Conversely, the persistence of endemic species near sea level on Leyte coincides with the lower extent of forest on Mt. Pangasugan. Both areas exhibited maximum species richness in primary lowland forest (Table 24.4). The transition from lowland to montane forest was the upper limit of a significant shift in species richness. Diversity indices of sites above lowland forest do not differ significantly, although values show a general downward trend with increasing elevation (Magurran 1988: t-test on H' values and variances of H',p > 0.05) (Table 24.4). Thus, lowland forest zones may define the upper elevational boundary of maximum fruit bat diversity. Species diversity levels (H') do not differ statistically (Magurran 1988: t-test on H' and variance of H', p > 0.05) between pairs of equivalent habitats even when these zones occurred at different elevations on each mountain (Table 24.4). Attenuation in species numbers occurred at the transition from montane to mossy forest on Mt. Pangasugan, but not on Mt. Talinis. On Mt. Talinis, species richness was higher in the montane-mossy transition forest than expected. I attribute this to the atypical occurrence of the nonendemic species Cynopterus brachyotis and Eonycteris spelaea at the site (see Table 24.2). On Mt. Guinsayawan (northeast of Talinis), neither of these species occurred above upper montane forest (Heaney et al. 1989), and unpublished data from Mt. Isarog (in Luzon; L. R. Heaney, personal communication) revealed the same pattern of elevational distribution for these species. I interpret this unexpectedly high level of species richness at the transition zone of montane and mossy forests on Mt. Talinis as an upward range extension of species typical of lowland 350 R. C. B. UTZURRUM (A) Negros c o •*— 03 > 0) LU 2000 (B) Leyte OJ c75 3 % £ CC >• O f0<W3 ^ J. OUJSO^fzatfll O UJ 2 £ Q. CL Figure 24.2. Comparative elevational distribution of fruit bats on (A) Negros Island (a composite from Mt. Guinsayawan and Mt. Talinis studies) and (B) Leyte (Mt. Pangasugan). Species are identified by the first two letters of the genus and species names (see Table 24.1 for list). The horizontal lines denote lower elevational limits of the various types of forests: solid line, primary forest; dashed line, montane forest (approximate); dotted line, transition into mossy forest; dashed-and-dotted line, well-developed mossy forest. Primary forest on Negros begins at an elevation of 500 m, whereas on Leyte it is still present at 50 m. Maximum elevations at the study sites were 1,800 m (Negros) and 1,150 m (Leyte). or disturbed habitat in response to habitat disturbance at this site and its increased proximity to cultivated fields on adjoining slopes. This response should be differentiated from increased diversity that may occur at zones where communities form ecotones (Ricklefs 1979). In this particular case, the latter phenomenon does not truly apply because the fruit bat species found in the adjoining montane and mossy sites were not distinct from each other. I draw two points of relevance from the preceding comparisons for the maintenance of species diversity. First, changes in the quality and quantity of forest habitats alter the nature of species assemblages within forest types by affecting both the relative numbers among species and the types of species present. As a corollary, these results suggest that (1) light to moderate levels of habitat disturbance, where the primary forest structure and composition and climatic conditions are retained, result in higher species richness than would be expected in undisturbed forest of Table 24.4 Total Abundance of Fruit Bats and Measures of Their Species Richness, Diversity, and Evenness along Two Mountains Site Elevation (m) Abundance (bats/net-night) No. of bat species Diversity (H'y Evennes Mt. Pangasugan, Leyte Island'' Agriculture and disturbed lowland forest 50 6.08 6 1.690* 0.943 Lowland forest, 2 sites: (1) Primary forest, disturbed 300 0.60 6 1.277* 0.712 (2) Primary forest 500 0.82 8 1.829* 0.880 Primary montane forest 700 0.97 7 1.065 0.547 Primary mossy forest 950 0.45 3 0.730 0.660 Mt. Talinis, Negros Island Agriculture and lowland secondary growth 500 3.24 7 1.618* 0.831 Lowland forest, 2 sites: (1) Secondary forest 500 1.24 6 1.360* 0.759 (2) Primary forest, disturbed 750 1.90 9 1.792* 0.816 Primary montane forest 1,100 0.66 4 1.047 0.755 Transitional montane-mossy forest 1,250 1.68 7 1.157 0.595 Primary mossy forest 1,625 0.22 3 0.760 0.691 "An H' value (the Shannon-Weiner index) marked by an asterisk (*) differs significantly from the H' value of the next higher elevation on the same mountain. 'Mt. Pangasugan data are from Heaney et al. (1989) and Rickart et al. (1993). Note that, unlike Mt. Talinis, Mt. Pangasugan had no transitional montane-mossy forest. Maintenance of Fruit Bat Diversity 351 similar type and elevation; (2) the changes in the community structure relating to moderate habitat disturbance result primarily from local range extensions of nonendemic species that are typically associated with disturbed habitats and are rare or absent in primary forest (especially above lowland forest); (3) large-scale habitat disturbances resulting in degradation of the principal forest structure, the alteration of associated climate conditions, and (or) marked fragmentation of formerly continuous tracts of forests may result in lower levels of species richness than would be expected; and (4) the decrease in species richness in heavily disturbed or fragmented forest habitats is associated primarily with the disappearance of endemic species or with their increasing rarity. Second, differences in the topographical features of mountain forests influence the responses of both plant and fruit bat community structure. Steepness, ruggedness, and irregularity in topography determine the degree of interface between habitat bands as well as the depth and expanse of a given band. Disturbance of similar scales may have different effects on two landscapes of dissimilar topographical features. Hence, efforts should be made toward the analysis of landscape features as they may influence local bio tic communities. In general, I predict that the upper montane zone will constitute the upper elevational limit of maximum fruit bat diversity. As forests at lower elevations disappear and the lower edges of forest progressively shrink upward in elevation, we may see shifts in the region of highest diversity from lower to higher forest regions. However, this "elevational retreat" may reach its limit when conditions (food, roosting, climatic) of the environment necessary for supporting viable populations in themselves become limiting, as may be the case in mossy forests. While this prediction results from studies associated with anthropogenic destruction of forests, habitat disturbances resulting from natural catastrophes (e.g., hurricanes) may generate similar results. One difference between these two types of disturbances is this: Although natural catastrophes do not typically generate sustained destructive stresses, anthropogenic processes typically do so. Future Conservation Research Needs Further inventories of continuous habitat gradients in the tropics are expected to demonstrate the close association between forest habitat and endemic species. In this context, comparisons of elevational distributions against a backdrop of changing habitat are useful as preliminary indices of the nature of diversity patterns and the ecological processes that may affect local community structure. Arguably, a more thorough analysis of relationships between habitat fragmentation and changes in species diversity requires quantitative measurements of habitat disturbance, such as relative areal coverage, spatial geometry, and extent of edge habitats, and the impact on local fruit bat assemblages. Additionally, there is a need to examine the actual processes and mechanisms that underpin community structuring and correlations among species diversity, population structure, and habitat quality. These include: (1) identification and quantification of those elements of the habitat of direct importance to fruit bats, specifically food resources and roosts; (2) determination of critical ranges of environmental conditions (temperature and humidity) that are physiologically compatible with the persistence of species at given habitats; and (3) overlaying spatial analyses on analyses at temporal scales to assess how annual or seasonal dynamics of fruit bat activities (e.g., reproduction) will influence their responses to habitat changes on a spatial scale. Conclusions The potential effects of habitat fragmentation on the maintenance of biodiversity are varied (see Lande 1988 and Terborgh 1992 for recent reviews). Species extinctions are possible end results of these effects. It is widely believed that such extinctions result from demographic stochasticity rather than accumulations of deleterious genetic changes (Lande 1988). In the final analysis, both demographic and genetic stochasticity often become relevant only after populations have been decimated to inviable numbers. Hence, the pivotal issue for the conservation of biodiversity is the prevention of declines in population sizes. The studies summarized in this chapter strongly indicate that changes in fruit bat communities can occur within the brief time scale in which forest habitats are being destroyed. In this regard, it is apparent that the prevention of further forest destruction is most crucial for the long-term maintenance of fruit bat diversity. Acknowledgments Financial support for the 1990 field research on Mt. Talinis was provided by the American Society of Mammalogists (grants-in-aid for research), Bat Conservation International, Inc., Chicago Zoological Society, Marshall Field and Ellen Thorne Smith Funds of the Field Museum of Natural History, the Lubee Foundation, Inc., and Boston University. I thank F. Catalbas, M. Furacan, E. Maro, and L. Tag-at for their valuable assistance in the field. Many other studies on Philippine fruit bats cited, in which I have had the opportunity to collaborate, were supported by the U.S. National 352 R. C. B. UTZURRUM Science Foundation (BSR-8514223), the John D. and Catherine T. MacArthur Foundation, and the Field Museum of Natural History (with L. R. Heaney as RI.). L. R. Heaney, T. H. Kunz, E. A. Rickart, and J. O. Seamon provided valuable comments on this manuscript. I am especially thankful to L. R. Heaney, P. D. Heideman, and E. A. Rickart for providing the intellectual atmosphere and research opportunities that have shaped my interest in Philippine bat research. Literature Cited Alcala, A. C. 1976. Philippine Land Vertebrates: Field Biology. New Day Publishers, Quezon City. Brown, W C., and A. C. Alcala. 1970. The zoogeography of the herpetofauna of the Philippine islands, a fringing archipelago. Proceedings of the California Academy of Science 38:105-130. Dickerson, R. E. 1928. Distribution of life in the Philippines. Bureau of Science Monograph (Manila) 21:1-322. Dickinson, E. C., R. S. Kennedy, and K. C. Parkes. 1991. The birds of the Philippines: An annotated checklist. British Ornithologists Union Checklist 12:1-507. Frahm, J.-R, and S. R. Gradstein. 1991. An altitudinal zonation of tropical forests using bryophytes. Journal of Biogeography 18: 669-678. Goodman, S. M., and N. R. Ingle. 1993. Sibuyan Island in the Philippines: Threatened and in need of conservation. Oryx 27: 174-180. Grubb, P. J., and T. Whitmore. 1966. A comparison of montane and lowland rain forest in Ecuador. II. The climate and its effects on the distribution and physiognomy of the forest. Journal of Ecology 54:303-333. Hague, P., J. Terborgh, P. Winter, and J. Parkinson. 1986. Conservation priorities in the Philippine archipelago. Forktail 2:83-91. Heaney, L. R. 1986. Biogeography of mammals in SE Asia: Estimates of rates of colonization, extinction, and speciation. In Island Biogeography of Mammals, L. R. Heaney and B. D. Patterson, eds., pp. 127-165. Academic Press, London. Heaney, L. R. 1991a. An analysis of patterns of distribution and species richness among Philippine fruit bats (Pteropodidae). Bulletin of the American Museum of Natural History 206: 145-167. Heaney, L. R. 1991b. A synopsis of climatic and vegetational change in Southeast Asia. Climatic Change 19:53-61. Heaney, L. R. 1993. Biodiversity patterns and conservation of mammals in the Philippines. Asia Life Sciences 2:261-270. Heaney, L. R., and P. D. Heideman. 1987. Philippine fruit bats: Endangered and extinct. Bats 5(1): 3-5. Heaney, L. R., and D. S. Rabor. 1982. Mammals of Dinagat and Siargao islands, Philippines. Occasional Papers of the Museum of Zoology of the University of Michigan 699:1-30. Heaney, L. R., and E. A. Rickart. 1990. Correlations of clades and clines: Geographic, elevational, and phylogenetic distribution patterns among Philippine mammals. In Vertebrates in the Tropics, G. Peters and R. Hutterer, eds., pp. 321-332. Museum Alexander Koenig, Bonn. Heaney, L. R., and R. C. B. Utzurrum. 1991. A review of the conservation status of Philippine land mammals. Association of Systematic Biologists of the Philippines Communications 3:1-13. Heaney, L. R., P. C. Gonzales, and A. C. Alcala. 1987. An annotated checklist of the taxonomic and conservation status of land mammals in the Philippines. Silliman Journal 34:32-66. Heaney, L. R., P. D. Heideman, and K. M. Mudar. 1981. Ecological notes on mammals in the Lake Balinsasayao region, Negros Oriental, Philippines. Silliman Journal 28:122-131. Heaney, L. R., P. C. Gonzales, R. C. B. Utzurrum, and E. A. Rickart. 1991. The mammals of Catanduanes Island, Philippines. Proceedings of the Biological Society of Washington 104:399-415. Heaney, L. R., P. D. Heideman, E. A. Rickart, R. B. Utzurrum, and J. S. H. Klompen. 1989. Elevational zonation of mammals in the central Philippines. Journal of Tropical Ecology 5:259-280. Heaney, L. R., D. S. Balete, L. Dolar, A. C. Alcala, A. Dans, P. C. Gonzales, N. Ingle, M. Lepiten, W Oliver, E. A. Rickart, B. R. Tabaranza, Jr., and R. C. B. Utzurrum. 1998. A synopsis of the mammalian fauna of the Philippine islands. Fieldiana Zoology (new series) 88:1-61. Heideman, P. D. 1987. The reproductive ecology of a community of Philippine fruit bats (Pteropodidae, Megachiroptera). Ph.D. dissertation, University of Michigan, Ann Arbor. Heideman, P. D., and L. R. Heaney. 1989. Population biology and estimates of abundance of fruit bats (Pteropodidae) in Philippine submontane forest. Journal of Zoology (London) 218:565-586. Ingle, N. R., and L. R. Heaney. 1992. A key to the bats of the Philippine islands. Fieldiana Zoology (new series) 69:1-44. IUCN (International Union for Conservation of Nature and Natural Resources). 1991. Protected Areas of the World: A Review of National Systems. Vol. 1: Indomalaya, Oceania, Australia, and Antarctica. Island Press, Washington, D.C. Kummer, D. M. 1990. Deforestation in the post-war Philippines. Ph.D. dissertation, Boston University, Boston. Lande, R. 1988. Genetics and demography in biological conservation. Science 241:1455-1460. Magurran, A. E. 1988. Ecological Diversity and Its Measurement. Princeton University Press, Princeton. Mallari, N. A. D., and A. Jensen. 1993. Biological diversity in northern Sierra Madre, Philippines: Implication for conservation and management. Asia Life Sciences 2:101-102. Mickleburgh, S. P., P. A. Racey, and A. M. Hutson, eds. 1992. Old World Fruit Bats: An Action Plan for the Family Pteropodidae. IUCN Press, Gland, Switzerland. Myers, N. 1988. Environmental degradation and some economic consequences in the Philippines. Environmental Conservation 15:205-214. Peterson, A. T., and L. R. Heaney. 1993. Genetic differentiation in Philippine bats of the genera Cynopterus and Haplonycteris. Biological Journal of the Linnean Society 49:203-218. Rickart, E. R. 1993. Diversity patterns of mammals along elevational and diversity gradients in the Philippines: Implications for conservation. Asia Life Sciences 2:251-260. Rickart, E. A., L. R. Heaney, and R. C. B. Utzurrum. 1991. Distribution and ecology of small mammals along an elevational Maintenance of Fruit Bat Diversity 353 transect in southeastern Luzon, Philippines. Journal of Mammalogy 72:458-469. Rickart, E. A., L. R. Heaney, P. D. Heideman, and R. C. B. Utzurrum. 1993. The distribution and ecology of mammals on Leyte, Biliran, and Maripipi islands, Philippines. Fieldiana Zoology (new series) 72:1-62. Ricklefs, R. E. 1979. Ecology, 2nd Ed. Chiron Press, New York. Ruedas, L. A., J. R. Demboski, and R. V Sison. 1994. Morphological and ecological variation in Otopteropus cartilagonodus Kock, 1969 (Mammalia: Chiroptera: Pteropodidae) from Luzon, Philippines. Proceedings of the Biological Society of Washington 107:1-16. Terborgh, J. 1992. Maintenance of diversity in tropical forests. Biotropica 24:283-292. Utzurrum, R. C. B. 1991. Philippine island biogeographic patterns: Practical applications for resource conservation and management. Association of Systematic Biologists of the Philippines Communications 3:19-32. Utzurrum, R. C. B. 1992. Conservation status of Philippine fruit bats (Pteropodidae). Silliman Journal 36:27-46. Vinciguerra, L. B„ and R. A. Miiller. 1993. Neue Erkenntnisse iiber die Verbreitung des Rohrennasen-Flughunds Nyctimene rabon Heaney & Peterson, 1984 auf den Philippinen (Mammalia: Pteropodidae). Jahrbuch des Naturhistorisches Museum der Stadt Bern 11:125-129. Whitmore, T. C. 1984. Tropical Rain Forests of the Far East, 2nd Ed. Oxford University Press, Oxford.