Comparative avian phylogeography of ... - Wiley Online Library

MEC1032.fm Page 1505 Thursday, September 7, 2000 3:56 PM

Molecular Ecology (2000) 9, 1505 –1516

Comparative avian phylogeography of Cameroon and Equatorial Guinea mountains: implications for conservation

Blackwell Science, Ltd

T. B . S M I T H , * ‡ K . H O L D E R , * D . G I R M A N , * † K . O ’ K E E F E , * B . L A R I S O N * and Y. C H A N * *Center for Tropical Research and Department of Biology, San Francisco State University, San Francisco CA 94132, USA, †Department of Biology, Sonoma State University, Sonoma CA 94928, USA, ‡Center for Population Biology, University of California at Davis, Davis CA 95616, USA

Abstract We illustrate the use of Faith’s ‘Phylogenetic Diversity’ measure to compare the phylogeographic structure of two bird species with patterns of avian endemism across six mountains in Cameroon and Equatorial Guinea. The Mountain Greenbul and Cameroon Blue-headed Sunbird showed phylogeographic patterns that together defined three biogeographic regions: Bioko, Mt. Cameroon, and the northern mountains of Cameroon. In contrast, the distributions of endemic species were largely a function of geographical distance, with close mountains sharing more endemic species than distant mountains. Moreover, for both species, populations on Mt. Cameroon were distinctive with respect to the ecologically relevant character bill size. Our results, while preliminary, illustrate the utility of a comparative approach for identifying geographical regions that harbour evolutionarily distinct populations and caution against using only the distributional patterns of endemics to prioritize regions for conservation. Results show that patterns of endemism may not be concordant with patterns of phylogenetic diversity nor morphological variation in a character important in fitness. While incorporation of additional species from unrelated taxa will be necessary to draw definitive conclusions about evolutionarily distinct regions, our preliminary results suggest a conservation approach for the Afromontane region of the Gulf of Guinea that would: (i) emphasize protection of both Bioko and Mt. Cameroon, thereby maximizing preservation of within–species phylogenetic and morphologic diversity; (ii) emphasize protection within the northern mountains to further conserve intraspecific phylogenetic diversity and maximize protection of endemic species. Keywords: Afromontane, endemism, Gulf of Guinea, phylogeography, phylogenetic diversity, rainforests, Received 21 December 1999; revision received 26 April 2000; accepted 3 May 2000

Introduction Species richness, endemism, demography, genetic distinctiveness, and magnitudes of anthropogenic threats are all important sources of information for making informed conservation decisions (Mace & Lande 1991; Pressey et al. 1993; Balmford et al. 1998). Prioritization of areas for conservation ideally requires that data from these various components be integrated. For conservation planning to be Correspondence: Thomas B. Smith, Department of Biology, 1600 Holloway Ave., San Francisco State University, San Francisco, CA 94132, USA Fax: +415 405 – 0421; E-mail: [email protected] © 2000 Blackwell Science Ltd

successful, biologists should endeavour to preserve not only the pattern of biodiversity but also the evolutionary processes that generate and maintain it (Erwin 1991; Smith et al. 1993). In this context, the use of genetic markers to infer evolutionary histories at the intra- or interspecific level, and to prioritize areas for conservation, is widely recognized (Avise 1994; Humphries et al. 1995; Smith et al. 1997; Moritz & Faith 1998). One useful approach is to compare intraspecific phylogeographies of codistributed species to identify important geographical areas in which populations have undergone independent evolution (Avise 1992; Faith 1992; Crozier & Kusmierski 1994; Moritz & Faith 1998). If phylogenetic patterns are concordant across


1506 T. B . S M I T H E T A L . diverse taxa, extrapolating to multispecies communities may be possible. Thus, the application of comparative phylogeography (Moritz 1994; Vogler & DeSalle 1994), when coupled with data on endemism and ecologically relevant characters, has the potential to identify important geographical areas for conservation efforts (Balmford et al. 1998; Crandall et al. 2000). Moritz & Faith (1998) defined three requirements for the use of comparative phylogeography for conservation purposes: (i) an appropriate geographical area must be defined, based on biogeographic data, geomorphology or other criteria; (ii) a history of vicariance across regions to be compared must be shared among taxa, as indicated by congruence of phylogenetic division and biogeographic barriers; and (iii) combinations of regions that maximize diversity must be identifiable. Quantitative methods for testing congruence of phylogenies have recently been developed (Page 1994). One such approach, summarizing genetic data across codistributed taxa by incorporating branch length data, is the ‘Phylogenetic Diversity’ (PD) measure (Faith 1992; Faith 1994; Moritz & Faith 1998). The PD method is useful in characterizing the spatial pattern of genetic diversity irrespective of the taxonomic unit (Faith & Walker 1996; Moritz & Faith 1998). In this application, PD estimates the underlying pattern of phylogenetic diversity by calculating the total amount of the branching pattern that is spanned by individuals of one distinct geographical region, and the sum of diversity encompassed by two regions together (Moritz & Faith 1998). When combined with other relevant evolutionary and ecological data, this approach can help to focus conservation efforts. In this paper, we illustrate the use of the expanded PD measure approach of Moritz & Faith (1998) to compare phylogenetic data from two codistributed bird species with information on the distributions of montane endemic bird species in Cameroon and Equatorial Guinea (hereafter the Afromontane region of the Gulf of Guinea). In addition, we examine variation in ecologically relevant morphologic characters. These approaches allow us to make a preliminary assessment of the conservation value of mountains to diversity in this Afromontane region. Montane forests of western Cameroon and islands of the Gulf of Guinea are well known for their biogeographic significance, particularly with respect to the large numbers of endemic and threatened bird species (Louette 1981; Collar & Stuart 1985; Stuart 1986). Currently, considerable attention focuses on the conservation of these high elevational forests (Stuart 1986; Smith & McNiven 1993; Perez del Val et al. 1994; Larison et al. 2000). Conservation efforts are currently underway, or are being developed for many of the mountains, including Pico Basile and the southern highlands of Bioko in Equatorial Guinea, Mt. Cameroon, Mt. Kupe, Mt. Oku and Tchabal Mbabo in Cameroon. Despite intense conservation interest, comparative

phylogenetic and morphological data for montane species inhabiting these mountains are lacking. The objectives of this study were to illustrate the use of a quantitative framework to assess avian biodiversity and make a preliminary assessment of evolutionary units across the Afromontane region of the Gulf of Guinea. Specifically, we: (i) compare the intraspecific mitochondrial DNA (mtDNA) phylogenies of two montane bird species to identify which mountains or mountain groups support populations with independent evolutionary histories; (ii) use quantitative methods to describe avian endemism across mountains; (iii) contrast avian endemism with patterns of genetic diversity within species, and (iv) assess patterns of morphological variation across mountains. Finally, we discuss the utility of an integrated approach for setting priorities for conservation efforts for the region.

Methods Study sites The geological history of the region is complex, with volcanic activity beginning in the Upper Cretaceous. The chain of highlands and mountains (Fig. 1), referred to as the Cameroon line, extends north-east from the islands in the Gulf of Guinea through Mt. Cameroon to Tchabal Mbabo. Most of the volcanic activity likely occurred during the Pliocene (Wright et al. 1985), with signs of recent volcanic activity in Mt. Cameroon persisting through the Quaternary to present. The island of Bioko is Pliocene in age (Wright et al. 1985). Unlike the mountains of volcanic origin, Mt. Kupe is a massive horst of granite and syenite, formed by blockfaulting (Stuart 1986). For this paper, we sampled five mountains in Cameroon and one in Equatorial Guinea (Fig. 1). Descriptions of the montane forests on each mountain follow. Mt. Cameroon: Located on the coast of south-western Cameroon, this isolated, large (50 × 30 km) volcano is the highest mountain in West Africa at 4085 m. Its approximately 800 km2 of forest range from lowland forest at sea level to montane forest between 1400 and 2500 m. Because Mt. Cameroon is an active volcano, mature forest is interspersed with forest in various stages of regeneration due to periodic lava flows. The south and south-east portions of the mountain are well populated and vulnerable to deforestation from agriculture and logging. The only protected area is the 300 km2 Bambuku Forest Reserve on the northwest side of the mountain (Collar & Stuart 1985; Stuart 1986). Sampling took place in montane forest at an elevation of approximately 1500 m, at a site (N4°9′ E9°14′) above the town of Buea on the southern slope (31 May–2 June 1993, 25 May– 20 July 1995), and above the village of Mapanja at 1200 m (N4°5′ E9°09′; 30 March–3 April 1997 and 10 –12 July 1999). © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516


P H Y L O G E O G R A P H Y O F A F R O M O N TA N E AV I FA U N A 1507 Fig. 1 Afromontane region of Cameroon and Equatorial Guinea. Mountains on which birds were sampled are shown with abbreviations in bold.

Mt. Kupe: The first major peak inland from Mt. Cameroon reaches 2064 m. Montane rainforest begins at approximately 1600 m, and is estimated to cover only 0.21 km2 (Collar & Stuart 1985; Stuart 1986). Sampling took place in montane forest at an elevation of 1600 m (N4°47′ E9°43′; 26 – 29 May 1993, 26 February–2 March 1997). Bakossi Mountains: These mountains to the north-west of Mt. Kupe exceed 1500 m elevation in many areas, and support healthy montane forests. Sampling took place in montane forest at an elevation of 1600 m near the village of Kodmin (N4°57′ E9°42′; 5 – 6 March 1997). Mt. Oku: Part of the Bamenda Highlands, Mt. Oku is the second highest mountain in Cameroon, at an elevation of 3011 m. Due to pressures from high human population densities, agriculture and grazing, little forest exists on Mt. Oku below 2000 m (Stuart 1986). The remaining forest is montane, and extends from 2000 m to the summit, but is threatened by grazing, wood harvesting, burning and over-exploitation of small mammals and Pygaem bark. Sampling took place in montane forest at an elevation of 2070 m (N6°14′ E10°31′; 23 July 1995). Tchabal Mbabo: This plateau forms a 25–30 km crescentshaped ridge. Montane forest extends from over 2000 m down to approximately 1700 m, narrowing below this elevation into gallery forest and thickly wooded savanna. The north and west facing slopes harbor approximately 2500 hectares of montane forest and montane scrub, while the large montane gallery forests on the plateau are still in need of quantification (Thomas & Thomas 1996). The steep north slopes where most of the montane forest © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505–1516

occurs do not permit easy access, and are, therefore, relatively undisturbed. In contrast, the plateau is under heavy human pressure, including livestock grazing, crops, and wood gathering (Larison et al. 2000). We sampled the plateau on the western escarpment in a gallery at 2000 m from 1–5 December 1990 (Smith & McNiven 1993), and on the north slope about 5 km from Mayo Kélélé (N07°16′ E12°09) at an elevation of 1930–45 m from 11– 20 July 1995 (Larison et al. 2000). Southern highlands of Bioko: The region is thickly forested, with montane forest beginning as low as 800 m (Guinea 1951). Primary forest on the relatively uninhabited south slope of the island runs in a continuous gradient from lowland to montane forest (Jones 1994), representing one of the few intact gradients in West Africa. We sampled between 1600 and 1650 m on the north slope of the Caldera de Luba (N3°23′ E8°33′) above the village of Ruiché (23–26 June 1996; Larison et al. 2000).

Field methods Sampling for genetic and morphological analyses concentrated on two species, the Mountain Greenbul (Andropadus tephrolaemus) and the Cameroon Blue-headed Sunbird (Nectarinia oritis). The Mountain Greenbul, which is restricted to montane forests above 1500 m, has a wide geographical distribution in Africa, occurring throughout many mountains of East Africa (Keith et al. 1992; Roy 1997). Two subspecies are recognized in the Afromontane region of the Gulf of Guinea: A. tephrolaemus tephrolaemus, occurring on Mt. Cameroon and Bioko, and A. t. bamendae, a darker race found from Mt. Kupe to Tchabal Mbabo


1508 T. B . S M I T H E T A L . (Louette 1981). In contrast, the Cameroon Blue-headed Sunbird is endemic to Cameroon and Bioko, Equatorial Guinea. It is a shy, sedentary species found in the lower canopy of high montane forest (Mackworth-Praed & Grant 1973). There are three recognized subspecies, N. oritis poensis on Bioko, N. o. oritis on Mt. Cameroon and N. o. bansoensis on mountains in northern Cameroon. At each site, we erected 10–33 mist nets (12 m, 30 × 30 mm mesh) in forest along cleared net lanes (Larison et al. 2000). Nets were opened at sunrise (0600 h) and closed between 1100 h and 1200 h. Individuals were banded with numbered aluminium bands, sexed, aged, measured (Smith 1990a; Smith 1990b; Smith et al. 1997), and a small (25–50 µL) sample of blood was taken via brachial venipuncture. Morphological measurements were taken by T. B. Smith using dial calipers, as follows: wing length, from carpal joint to the tip of the longest primary; tail length, from the uropygial gland to the end of the longest tail feather; tarsus length, from tibiotarsal joint to distal undivided scute; upper mandible length, chord length from point where culmen enters feathers of the head to tip; bill depth in vertical plane level at the anterior edge of nares. Mass was measured to the nearest 0.5 g using a 50-g Pesola spring scale. Age was determined by plumage characteristics (Mackworth-Praed & Grant 1973). Because both species show sexual dimorphism in size (Keith et al. 1992), it was important to distinguish between the sexes. N. oritis are sexually dimorphic in plumage, allowing adult males to be easily distinguished from females. However, juvenile males and adult females cannot be distinguished with confidence. Similarly, adult A. tephrolaemus are sexually monomorphic for plumage. We therefore distinguished males from females by screening for size differences in a sex–linked molecular marker (see below). Samples sizes for females of each species were small, so we examined morphological differences for adult males only.

DNA sequencing We collected blood samples from A. tephrolaemus individuals on five mountains, and from N. oritis on all six mountains. We stored blood in buffer (following Smith et al. 1997) at ambient temperature (short-term) or −20 °C. To isolate whole genomic DNA, we used protease digestion followed by phenol and chloroform/isoamyl alcohol extraction method (Kocher et al. 1989). For N. oritis, we amplified and sequenced a 578-bp portion of the cytochrome b gene of the mitochondrial genome with primers L14841 and H15547 (Kocher et al. 1989). For A. tephrolaemus, we amplified and sequenced a 635-bp fragment of the cytochrome b gene using primers MVZ–03′ (modified from L14841) and MI02 (Chikuni et al. 1995), and a 284-bp fragment of mitochondrial control region I using primers LGL2 and H417 (Tarr 1995). Double-

stranded polymerase chain reaction (PCR) products were generated in either 25 µL or 50 µL reactions consisting of buffer (100 mm Tris–HCl, pH 8.3, 50 mm KCl, 1% gelatin), 2.5 mm MgCl2, 0.2 mm dNTPs, 0.1 mm of each primer, 2.5 µg of bovine serum albumin, 1.25 units of AmpliTaq DNA polymerase, and 20–60 ng of genomic DNA. PCR products were purified using a QIA Quick® Kit following manufacturer’s instructions (Qiagen Inc., Valencia, CA). We used double-stranded cycle sequencing with either dye-primer or dye-terminator fluorescent labelling, and electrophoresed sequenced products through a 5% LongRanger gel in an ABI Prism 377® automated sequencer (Applied Biosystems Inc., Foster City, CA). Sequences were aligned using the program Sequencher version 3.0 (Gene-Codes-Corporation 1995). To identify the sex of A. tephrolaemus and N. oritis individuals, we used a PCR based approach which amplifies the sex-linked gene for chromo-helicase-DNA-binding protein (CHD). Copies of CHD occur on both the W and Z chromosomes (CHD1W and CHD1Z, respectively), but a constant size difference between introns in the two copies allows males to be distinguished from females (Ellegren 1996; Fridolfsson & Ellegren 1999). Sexing protocols followed Fridolfsson & Ellegren (1999).

Endemic species To compare patterns of endemic species distributions with patterns of genetic variation, we compiled data on the geographical distributions of species endemic to the Afromontane region of the Gulf of Guinea from several published sources (Mackworth-Praed & Grant 1973; Louette 1981; Brown et al. 1982; Stuart 1986; Urban et al. 1986; Fry et al. 1988; Keith et al. 1992; Smith & McNiven 1993; Perez del Val et al. 1994; Larison et al. 2000). We consider these distributions to be approximate and not definitive. Because the avifauna of some mountains, particularly Mt. Cameroon and Bioko, have received greater attention by taxonomists (Perez del Val et al. 1994; Stuart 1986) and may consequently have more described subspecies, we limit our discussion to only those endemics recognized as full species. Further, some mountains are under-surveyed for birds in general compared to other mountains, and are likely to reveal more endemics with additional surveys (Urban et al. 1986; Smith & McNiven 1993; Larison et al. 2000). For this reason, we excluded the Bakossi Mountains which have been poorly surveyed. Two of the species considered by Stuart (1986) to be endemic to this region were not included in our analyses: Tullberg’s Woodpecker, Campethera tullbergi, which has recently been joined taxonomically with its eastern sister species C. taeniolaema (Fry et al. 1988), and Bamenda Apalis, Apalis bamenda, which is more widespread than previously thought (Urban et al. 1986; Larison et al. 2000). © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516


P H Y L O G E O G R A P H Y O F A F R O M O N TA N E AV I FA U N A 1509 Fig. 2 Maximum-likelihood trees for (A) cytochrome b haplotypes of Nectarinia oritis (B) cytochrome b haplotypes of Andropadus tephrolaemus, and (C) control region haplotypes of A. tephrolaemus. Branch lengths not significantly different from zero (α < 0.01) were collapsed.

Data analysis To infer the phylogenies of haplotypes, we employed the maximum-likelihood (ML) method of tree reconstruction using the dnaml program in phylip version 3.5c; (Felsenstein 1993). For A. tephrolaemus, we used sequences from the congeneric species A. latirostris to root the phylogenies for cytochrome b and control region haplotypes. For N. oritis, we used a cytochrome b sequence from the congeneric species N. olivacea to root the phylogeny. We report results based on the empirical transition/transversion (ti/tv) ratio for each analysis, but use of ti/tv ratios from 2.0 to 17.0 did not change the results. To estimate the underlying diversity within and among montane populations of each species, we used the (PD) measure (Faith 1992; Faith 1994; Faith & Walker 1996; © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505–1516

Moritz & Faith 1998). For within-region diversity, this approach sums the branch lengths in a phylogeny along the minimum path connecting all haplotype unique to the region. For diversity spanned by combinations of regions, this approach sums the branch lengths both within and among regions and extending to the root of the tree. To obtain PD values for individual mountains and pairwise combinations of mountains, we used the ML branch lengths among cytochrome b haplotypes for N. oritis (Fig. 2). We combined the sequence data from both loci for A. tephrolaemus, used the ML algorithm for estimating the phylogeny of haplotypes (tree not shown) and repeated the PD analysis. We first calculated PD values within each mountain and combination of montane regions for each species separately. We then summed values from both species into total PD values for all montane regions.


1510 T. B . S M I T H E T A L . using Mantel’s test. This method of statistical analysis corrects for the presence of autocorrelation (Legendre & Fortin 1989). For comparison with phylogenetic data, endemic species were partitioned among mountains and mountain groups, and represented using a Venn diagram and area cladogram following the methods of Moritz & Faith (1998).

Following Moritz & Faith (1998), we partitioned total PD among individual montane regions and combinations of regions using a Venn diagram. This approach illustrates how much of the total diversity is shared between regions, and how much is unique to each region. We then inferred an area cladogram from the partitioning of PD in the Venn diagram to summarize the hierarchical relationships among the areas (Faith 1992; Moritz & Faith 1998). To describe differences in endemic species distributions among mountains, we calculated Bray–Curtis distances between mountains based on the presence or absence of each species (Ludwig & Reynolds 1988), and used the resulting data to cluster mountains in a dendogram using the upgma algorithm in ntsys-pc version 1.80 (Rohlf 1993). To test whether dissimilarity of mountains based on endemic species composition increased as distance increased, we correlated Bray– Curtis distance with geographical distance

Results Phylogeographic patterns For Andropadus tephrolaemus, we identified 13 cytochrome b haplotypes from 31 individuals, and 10 control region haplotypes from 19 individuals, across five mountains (Table 1); we were unable to obtain scoreable, unambiguous results for the control region from the remaining 12

Table 1 Distribution of mitochondrial DNA haplotypes among montane populations of Andropadus tephrolaemus (cytochrome b and control region haplotypes) and Nectarinia oritis (cytochrome b haplotypes) Andropadus tephrolaemus cytochrome b (AtCb) haplotypes

Mt. Cameroon Bioko Mt. Kupe Mt. Oku Tchabal Mbabo

1

2

3

4

5

3

1

1

1

1

6

7

1

1

8

2

9

5 2

10

1

11

1

12

1

13

14

15

16

17

Total 7 2 11 4 8 — 32

1 1 2

1 4

2

Andropadus tephrolaemus control region (AtCR) haplotypes 1 Mt. Cameroon Bioko Mt. Kupe Mt. Oku Tchabal Mbabo

2

3

2

2

4

5

6

7

8

9

10

Total 2 2 7 1 5 — 17

2 3

2

1

1

1 1

1

3

Nectarinia oritis cytochrome b (No Cb) haplotypes

Mt. Cameroon Bioko Mt. Kupe Bakossi Mtns Mt. Oku Tchabal Mbabo

1

2

3

4

5

6

5

3

2

1

2

1

7

8

9

10

2

1

1

1

11

12

13

14

15

16

17

18

19

20

21

22

Total 14

3

1

1

1

1

1 5

1

1 2

2

1 1

5 8 7 5 1 — 40

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516


P H Y L O G E O G R A P H Y O F A F R O M O N TA N E AV I FA U N A 1511 individuals. Three strongly supported lineages were identified in both gene trees: the northern mountains (comprising Mt. Kupe, Mt. Oku, and Tchabal Mbabo), Mt. Cameroon, and Bioko (Fig. 2). For Nectarinia oritis, we identified 22 cytochrome b haplotypes from 40 individuals across all mountains (Table 1). As in A. tephrolaemus, there were strongly supported lineages for both Bioko and Mt. Cameroon; moreover, the northern mountains (comprising Mt. Kupe, Mt. Oku, Bakossi Mountains and Tchabal Mbabo) supported two clades that fell within an unresolved polytomy with Mt. Cameroon haplotypes (Fig. 2). Sequences for A. tephrolaemus and N. oritis have been deposited in GenBank (Accession nos AF282775 to AF282823).

Table 2 PD values calculated from branch lengths of maximumlikelihood trees based on mtDNA sequences for Nectarinia oritis (cytochrome b only) and Andropadus tephrolaemus (cytochrome b and control region combined). Values are calculated for comparisons within and among three biogeographic units, Bioko (BI), Mt. Cameroon (MC) and the Northern Mountains (NOR; see Fig. 1) Region

N. oritis

A. tephrolaemus

Total

Mt. Cameroon (MC) Bioko (BI) Northern Mountains (NOR) MC–BI MC–NOR BI–NOR

28 29 26 66 63 64

15 15 29 33 47 47

43 44 55 99 110 111

Phylogenetic diversity analysis PD values derived from ML trees of mtDNA haplotypes (Fig. 2) were compared within and among the mountains of Bioko, Mt. Cameroon and the northern mountains for both N. oritis and A. tephrolaemus (Table 2). Because PD values between pairs of northern mountains (Mt. Kupe, Mt. Oku, Bakossi Mountains and Tchabal Mbabo) were low, we combined the northern mountains into a single biogeographic unit for subsequent analyses. The distribution of phylogenetic diversity across the three montane regions (Table 2) can be easily visualized using a Venn diagram (Fig. 3A). First, the unique PD units for Mt Cameroon, Bioko and the northern mountains (43, 44 and 55, respectively) were assigned to the nonoverlapping portions of the Venn diagram. Second, the PD units shared among regions were partitioned among the overlapping

portions of the Venn diagram. For example, the combined PD value for Bioko and northern mountains is 111 (Table 2) which exceeds the sum of the regions’ unique PD units. This is because the remaining 12 units correspond to the PD shared among the regions (based on shared branches in the phylogenies), and are thus partioned among the overlapping portions of the Venn diagram. The northern mountains are clearly shown as supporting the highest unique PD value, followed by relatively equal contributions from both Bioko and Mt. Cameroon. The low values of overlapping diversity among the regions indicate that the phylogenies are strongly structured and informative, with little evidence of homoplasy among regions. Because Bioko and the northern mountains shared the most PD

Fig. 3 Venn diagrams showing (A) components of mtDNA phylogenetic diversity (PD) for two species across mountain regions, (C) counts of endemic species across mountain regions, and the corresponding area cladograms for (B) mtDNA data and (D) endemic species. See Moritz & Faith (1998) for detailed methods.

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505–1516


1512 T. B . S M I T H E T A L . units (i.e. nine) these regions grouped together in an area cladogram, suggesting a weak sister relationship.

Geographic patterns of endemism Based on our survey of the published literature 19 bird species are endemic to the Afromontane region of the Gulf of Guinea. These species include: Cameroon Mountain Francolin Francolinus camerunensis, Bannerman’s Turaco Tauraco bannermani, Cameroon Mountain Roughwing Psalidoprocne fuliginosa, Cameroon Mountain Greenbul Andropadus montanus, Cameroon Olive Greenbul Phyllastrephus poensis, Grey-headed Greenbul Phyllastrephus poliocephalus, Yellowbreasted Boubou Laniarius atroflavus, Green-breasted Bush-shrike Malaconotus gladiator, Mount Kupe Bushshrike Malaconotus kupeensis, Mountain Robin-chat Cossypha isabellae, White-throated Mountain Babbler Lioptilus gilberti, Black-capped Woodland-warbler Phylloscopus herbeti, Green Longtail Urolais epiclora, White-tailed Warbler Poliolais lopesi, Banded Wattle-eye Platysteira laticincta, Cameroon Blue-headed Sunbird Nectarinia oritis, Ursula’s Mouse-coloured Sunbird Nectarinia ursulae, Fernando Po White-eye Speirops brunneus, Bannerman’s Weaver Ploceus bannermani. Interestingly, the diversity and distributions of endemic species (Figs 3C,D) were not congruent with the distribution of phylogenetic diversity within two species (Figs 3A,B). The distributions of montane endemics analysed using upgma analysis resulted in three main groupings: (1) Bioko (2) Mt. Cameroon and Mt. Kupe and (3) Mt. Oku and Tchabal Mbabo (Fig. 4A). Thus, in contrast to relationships suggested by phylogenetic data, Mt. Kupe groups with nearby Mt. Cameroon rather than with the other northern mountains, Mt. Oku and Tchabal Mbabo. Moreover, there is a positive relationship between dissimilarity of endemic montane communities (Bray–Curtis distance) and geographical distance, such that a given mountain shares more endemic species with nearer mountains than with more distant mountains (Fig. 4B). At the species level, Bioko and Mt. Cameroon support only one unique endemic species each and, therefore, add little to the species diversity of the Afromontane region of the Gulf of Guinea as a whole.

Patterns of morphological divergence Because ecological differences among mountains may give rise to unique adaptive variation, we examined variation in fitness related morphological characters. Adult males of A. tephrolaemus and had significantly longer upper mandibles on Mt. Cameroon than on Tchabal Mbabo (Fig. 5; Mann–Whitney U-test, U = 17, P < 0.001). Upper mandible length of N. oritis adult males differed significantly across all mountains (Kruskal–Wallis test, H = 22.86, P 0.001). When the population on Mt. Cameroon was excluded

Fig. 4 (A) Relationships among mountains of Cameroon and Equatorial Guinea inferred from the distributions of endemic species by upgma clustering of Bray–Curtis distances, (B) Relationship between the dissimilarity of endemic communities measured by Bray–Curtus distances and geographical distance (y = 0.001x + 0.239; Mantel’s test, r = 0.7, P < 0.01).

from the analysis, there were no significant differences among the remaining populations (H = 1.79, P = 0.41), suggesting that, like A. tephrolaemus, N. oritis on Mt. Cameroon had significantly larger mandibles than those on other mountains. Additionally, A. tephrolaemus from Mt. Cameroon were significantly larger in mass, wing, tail, tarsus and lower mandible length than those on Tchabal Mbabo (Mann–Whitney U-tests: U = 172, P = 0.009; U = 178, P = 0.004; U = 172, P < 0.001; U = 58, P = 0.043; U = 17, P < 0.001, respectively). In contrast, N. oritis, while differing in upper mandible length, showed no significant trends in body size as indexed by mass, wing or tarsus length across mountains (Mann–Whitney U-tests, all P > 0.1). © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516


P H Y L O G E O G R A P H Y O F A F R O M O N TA N E AV I FA U N A 1513

Fig. 5 Box plots illustrating upper mandible length of (A) Andropadus tephrolaemus adult males (Bioko, n = 1; Mt. Cameroon, n = 33; Mt. Kupe, n = 2; Mt Oku, n = 3; Tchabal Mbabo, n = 24) and (B) Nectarinia oritis adult males (Bioko, n = 5; Mt. Cameroon, n = 6; Mt. Kupe, n = 7; Mt Oku, n = 8). Single lines indicate medians for samples sizes of fewer than five individuals. See Fig. 1 for mountain abbreviations.

Discussion Interpreting phylogeographic patterns If geological history and ecological factors are important determinants of genetic differentiation through their effects on patterns of drift and gene flow, then one should find congruent phylogeographic patterns among sympatric taxa that are influenced by the same processes (Rosen 1978; Wiley 1988; Moritz & Faith 1998). Combining phylogenies of mtDNA haplotypes thus provides an integrated picture of general patterns of genetic differentiation resulting from vicariant isolation of populations. While the molecular genetic results presented here on two taxa are of limited scope, and data on additional species are necessary to fully evaluate the evolutionary history of the Afromontane region of the Gulf of Guinea, we nevertheless offer a cautious interpretation of our results. © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505–1516

The Mountain Greenbul and the Cameroon Blue-headed Sunbird show relatively concordant phylogeographic patterns that together define three relevant biogeographic regions: Bioko, Mt. Cameroon, and the northern mountains (comprising Mt. Kupe, Mt. Oku, Bakossi Mountains and Tchabal Mbabo). Patterns of variation within each species reveal monophyletic mtDNA lineages associated with both Mt. Cameroon and Bioko. The pattern is less clear for the northern mountains, where populations of greenbuls comprise one monophyletic lineage but populations of sunbirds support members of two lineages, both of which are polytomous with the Mt. Cameroon lineage. The complex geological history of the area makes inferring any simple pattern of isolation and divergence difficult. Palaeoclimatic data suggests that many Afromontane vegetation types extended into the lowland 20 000 years before present (Maley 1996), providing a possible means of dispersal within suitable ecological habitats. The northern mountains remain connected by high ridges that may provide contemporary corridors for dispersal and gene flow even among widely separated mountains. In contrast, despite the geographical proximity of Mt. Cameroon to Mt. Kupe, there are no high elevation forest corridors connecting Mt. Cameroon with this or any other northern mountain. Further, because the island of Bioko has been isolated from the mainland repeatedly during the Pleistocene (Wright et al. 1985), the phylogegeography of mountain greenbuls and sunbirds suggests that avian populations on Bioko and Mt. Cameroon diverged in allopatry rather than by a simple model of isolation by distance.

Phylogeography and endemism Although the pattern revealed by PD analysis of genetic variation is limited to only two avian species, results contrast with the pattern of regional diversity based on endemism. The distribution of phylogenetic diversity within the greenbul and sunbird suggests that the Afromontane region of the Gulf of Guinea supports highly divergent genetic lineages with little overlap of diversity and a weak hierarchical relationship among three regions. In contrast, the distributions of endemic species are largely a function of geographical distance, with the closest mountains showing greater overlap of endemics than more distant mountains. Why are the two measures different? Given that we examined the phylogeography of only two species, a different pattern could emerge as additional species are examined. However, if the phylogeographic pattern revealed in this paper is indeed a general pattern representative of other species with similar ecological requirements and vagility, it suggests that the mechanisms determining patterns of neutral genetic variation differ from those determining patterns


1514 T. B . S M I T H E T A L . and levels of endemism. Regional associations based on independent phylogenies are consistent with the relative degree of geographical isolation rather than simply geographical distance. Bioko and Mt. Cameroon are both more isolated by topography than are any of the northern mountains. In contrast however, ridges and intervening highland could provide dispersal routes and avenues for gene flow among populations of the northern mountains. It is plausible that the same geographical connectedness among northern mountains leading to genetic exchange among populations may not necessarily lead to a similar association of endemics. The probability of a dispersal event leading to the successful establishment of a new population is much lower than the probability of a single individual dispersing and breeding, or hybridizing, with an already successfully established population (Grant & Grant 1989; Grant & Grant 1992). In other words, establishing an entirely new population of a species on a different mountain would likely require a large propagule of individuals. Such ‘colonizations’ would, therefore, tend to involve mountains in close proximity, giving rise to the pattern of increasing dissimilarity with increasing distance across the Afromontane region (Fig. 4B). In contrast, movement of single individuals through the system of ridges and highlands among the northern mountains is more probable, and would be sufficient to prevent significant population differentiation, among even widely separated montane populations, if demographic exchange exceeded one migrant per generation (Wright 1931). While incomplete sampling may be responsible for our failure to detect haplotypes shared among regions, the lack (historical and contemporary) of suitable high elevation forest habitat between the northern mountains and either Mt. Cameroon or Bioko may also have prevented movement of individuals and, therefore, may explain the absence of shared haplotypes among the three montane regions.

Ecological considerations Populations experiencing similar ecological conditions that shape adaptive traits are considered to exhibit ecological exchangeability (Templeton 1994). In other words, if the selective regimes experienced by different populations are similar enough, individuals of each population may be exchanged without a reduction in fitness (Crandall et al. 2000). Avian bill characters are particularly useful in assessing ecological exchangeability because they typically show high heritabilities and are closely correlated with feeding efficiency and fitness (Boag & van Noordwijk 1987; Smith 1993). Our results suggest that Mt. Cameroon populations of Andropadus tephrolaemus and Nectarinia oritis are morphologically distinct, having significantly longer bills than conspecifics on other mountains. Exactly what feature(s) of the environment may be associated

with a longer bill is unclear, but our results suggest that the ecology of Mt. Cameroon may be different from that of other mountains, thereby driving change in morphological characters. The possibility that the bill lengths of A. tephrolaemus and N. oritis on Mt. Cameroon have changed by such magnitude as a result of genetic drift rather than selection is less likely because feeding performance, and fitness, are associated with avian mandible length (Smith 1990a; Temeles & Roberts 1993; Grant & Grant 2000).

Conservation implications Comparative phylogeographic variation is rarely examined simultaneously with patterns of morphologic divergence and endemism (but see Schneider et al. 1999; Schneider & Moritz 1999). However, nonconcordant patterns among such disparate measures of biodiversity should not be looked on despairingly by conservation planners. Rather than representing conflicting outcomes of similar ecological and evolutionary processes, nonconcordant data sets likely represent different measures of historical, adaptive and demographic process. All such sources of information are useful, and conservation practitioners should endeavour to maximize their use in conservation planning (Balmford et al. 1998; Crandall et al. 2000). In particular, greater efforts should be made to integrate morphologic with phylogenetic data. Substantial periods of isolation, even over many millions of years, may not lead to phenotypic evolution, whereas selection along gradients in which populations exist in parapatry may result in significant changes in phenotype (Endler 1977). For example, two recent studies (Schneider & Moritz 1999; Schneider et al. 1999) found substantial intraspecific genetic divergence within several lizard species distributed among historically isolated regions, but little or no detectable phenotypic evolution despite long-term isolation and major reductions in population size. In contrast, adjacent populations found in ecologically distinct habitats showed large phenotypic differences suggesting that selection was acting differentially in the two habitats (Schneider et al. 1999). While data from additional, unrelated taxa will be required for a comprehensive survey of intraspecific phylogeographic variation in the Afromontane region of the Gulf of Guinea, our results illustrate the utility of combining morphologic and phylogenetic data with information on endemism to prioritize areas for conservation and management. While information on many factors not discussed in this paper are important in conservation planning for the region (see Larison et al. 2000), a pragmatic conservation approach for the region based on our preliminary data, would be to: (i) emphasize conservation efforts on Bioko and Mt. Cameroon because each of these sites supports highly genetically differentiated populations and more phylogenetic diversity than © 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516


P H Y L O G E O G R A P H Y O F A F R O M O N TA N E AV I FA U N A 1515 any single northern mountain, and together they support more phylogenetic diversity than all northern mountains combined. In addition, Mt. Cameroon harbours morphologically distinct populations of both species; And (ii) emphasize protection within the northern mountains so as to conserve intraspecific phylogenetic diversity and maximize protection of endemic species.

Acknowledgements We thank the Cameroon government for permission to conduct the research. We thank R. Wayne, L. Bernatchez and two anonymous reviewers for discussion and comments on the manuscript, A. French, M. Kimura, R. Fotso, D. McNiven for assistance in the field, and C. Saux for assistance in the lab. The work was supported through grants and support from the USAID-CARPE program, National Geographic Society, Wildlife Conservation Society, WWF– Cameroon, a GAANN fellowship to KO and grants NSFDEB9726425, NSF-IRCEB9977072 and NIH Office of Research on Minority Health #5 P20 RR11805 grant to TBS.

References Avise JC (1992) Molecular population structure and the biogeographic history of a regional fauna: a case history with lessons for conservation biology. Oikos, 63, 62–76. Avise JC (1994) Molecular Markers, Natural History and Evolution. Chapman & Hall, New York. Balmford A, Mace GM, Ginsberg JR (1998) The challenges to conservation in a changing world: putting processes on the map. In: Conservation in a Changing World (eds Mace GM, Balmford A, Ginsberg JR), pp. 1– 28. Cambridge University Press, Cambridge. Boag PT, van Noordwijk A (1987) Quantitative Genetics. In: Avian Genetics: a Population and Ecological Approach (eds Cooke F, Buckley PA), pp. 45 – 78. Academic Press, New York. Brown LH, Urban EK, Newman K, eds (1982) Birds of Africa (Vol. 1). Academic Press, London. Chikuni K, Minaka N, Ikenaga H (1995) Molecular phylogeny of some Passeriformes, based on cytochrome b. Journal of Yamashina Institute for Ornithology, 28, 1–8. Collar NJ, Stuart SN (1985) Threatened Birds of Africa and Related Islands. International Council for Bird Preservation, Cambridge. Crandall KA, Bininda-Emonds ORP, Mace GM, Wayne RK (2000) Considering evolutionary processes in conservation biology: an alternative to ‘Evolutionary significant units’. Trends in Ecology and Evolution, 15, 290–295. Crozier RH, Kusmierski RM (1994) Genetic distances and the setting of conservation priorities. In: Conservation Genetics (eds Loeschcke V, Tomiuk J, Jain SK), pp. 227– 237. Birkhäuser Verlag, Basel. Ellegren H (1996) First gene on the avian W chromosome (CHD) provides a tag for universal sexing of non-ratite birds. Proceedings of the Royal Society of London B, 263, 1635–1641. Endler JA (1977) Geographic Variation, Speciation, and Clines. Princeton University Press, Princeton, New Jersey. Erwin TL (1991) An evolutionary basis for conservation strategies. Science, 253, 750 – 752. Faith DP (1992) Conservation evaluation and phylogenetic diversity. Biological Conservation, 61, 1–10.

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505–1516

Faith D (1994) Phylogenetic pattern and the quantification of organismal biodiversity. Philosophical Transactions of the Royal Society of London B, 345, 45–58. Faith D, Walker P (1996) How do indicator groups provide informaton about the relative biodiversity of different sets of areas?: on hotspots, complementarity and pattern-based approaches. Biodiversity Letters, 3, 18– 25. Felsenstein J (1993) PHYLIP (Phylogeny Information Package), Version 3.5c. University of Washington, Seattle. Fridolfsson AK, Ellegren H (1999) A simple and universal method for molecular sexing of non-ratite birds. Journal of Avian Biology, 30, 116–121. Fry CH, Keith S, Urban EK, eds (1988) The Birds of Africa (Vol. III). Academic Press, New York. Gene-Codes-Corporation (1995) Sequencher 3.0 for the Apple Macintosh. Ann Arbor. Grant BR, Grant PR (1989) Evolutionary Dynamics of a Natural Population: the Large Cactus Finch of the Galapagos. University of Chicago Press, Chicago. Grant PR, Grant BR (1992) Hybridization of bird species. Science, 256, 193–197. Grant PR, Grant BR (2000) Quantitative genetic variation in populations of Darwin’s Finches. In: Adaptive Genetic Variation in the Wild (eds Mousseau TA, Sinervo B, Endler JA), pp. 1– 40. Oxford University Press, Oxford. Guinea E (1951) Esquema de la vegetation de Fernado Poo. Conferencia Internacional Deafricanistas Occidentales Circular, 4, 41–45. Humphries CJ, Williams PH, Vane-Wright RI (1995) Measuring biodiversity value for conservation. Annual Review of Ecology and Systematics, 26, 93–111. Jones PJ (1994) Biodiversity in the Gulf of Guinea: an overview. Biodiversity and Conservation, 3, 772–784. Keith S, Urban EK, Fry CH (1992) The Birds of Africa (Vol. 4). Academic Press, New York. Kocher TD, Thomas WK, Meyer A et al. (1989) Dynamics of mitochondrial DNA evolution in animals: amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences of the USA, 86, 6196 – 6200. Larison B, Smith TB, Fotso R, McNiven D (2000) Comparative avian biodiversity of five mountains in northern Cameroon and Bioko. Ostrich, 71, 269–276. Legendre P, Fortin M-J (1989) Spatial pattern and ecological analysis. Vegetatio, 80, 107–138. Louette M (1981) The birds of Cameroon. An annotated checklist. Verhandelingen van de Koniklijke Academie voor Wetenschappen, Letteren Schone Kunsten van Belgie. Klasse de Wetenschappen, 43, 1–295. Ludwig JA, Reynolds JF (1988) Statistical Ecology. John Wiley & Sons, New York. Mace G, Lande R (1991) Assessing extinction threats: toward a reevaluation of IUCN threatened species categories. Conservation Biology, 5, 148–157. Mackworth-Praed CW, Grant CHB (1973) Birds of West and WestCentral Africa (Vol. 2). Longman, London. Maley J (1996) The African rain forest — main characteristics of changes in vegatation and climate from the Upper Cretaceous to the Quaternary. Proceedings of the Royal Society of Edinburgh, 104B, 31–73. Moritz C (1994) Defining evolutionarily significant units for conservation. Trends in Ecology and Evolution, 9, 373 – 375.


1516 T. B . S M I T H E T A L . Moritz C, Faith DP (1998) Comparative phylogeography and the identification of genetically divergent areas for conservation. Molecular Ecology, 7, 419 – 429. Page R (1994) Maps between trees and cladistic analysis of historical associations among genes, organisms, and areas. Systematic Biology, 43, 58 –77. Perez del Val J, Fa JE, Catroviejo FJ, Purroy FJ (1994) Species richness and endemism of birds on Bioko. Biodiversity and Conservation, 3, 868 – 892. Pressey RL, Humphries CR, Margules CR, Vane-Wright RI, Williams PH (1993) Beyond opportunism: key principles for systematic reserve selection. Trends in Ecology and Evolution, 8, 124 –128. Rohlf RJ (1993) NTSYS-pc, Numerical Taxonomy and Multivariate Analysis System. Exeter Software, Setauket. Rosen DE (1978) Vicariant patterns and historical explanation in biogeography. Systematic Zoology, 27, 159–188. Roy M (1997) Recent diversification in African greenbuls (Pycnonotidae: Andropadus) supports a montane speciation model. Proceedings of the Royal Society of London B, 264, 1337– 1344. Schneider C, Smith TB, Larison B, Moritz C (1999) A test of alternative models of diversification in tropical rainforests: Ecological gradients vs. rainforest refugia. Proceedings of the National Academy of Sciences of the USA, 94, 13869–13873. Schneider CJ, Moritz C (1999) Refugial isolation and evolution in Australia’s wet tropics rainforest. Proceedings of the Royal Society of London B, 226, 191–196. Smith TB (1990a) Natural selection on bill characters in the two bill morphs of the African finch Pyrenestes ostrinus. Evolution, 44, 832 – 842. Smith TB (1990b) Resource use by bill morphs of an African finch: Evidence for intraspecific competition. Ecology, 71, 1246– 1257. Smith TB (1993) Disruptive selection and the genetic basis of bill size polymorphism in the African finch. Pyrenestes Nature, 363, 618 – 620. Smith TB, McNiven D (1993) Preliminary survey of the avifauna of Mt Tchabal Mbabo, west-central Cameroon. Bird Conservation International, 3, 13 –19. Smith TB, Bruford MW, Wayne RK (1993) The preservation of

process: the missing element of conservation programs. Biodiversity Letters, 1, 164–167. Smith TB, Wayne RK, Girman DJ, Bruford MW (1997) A role for ecotones in generating rainforest biodiversity. Science, 276, 1855–1857. Stuart SN (1986) Conservation of Cameroon Montane Forests. International Council for Bird Preservation, Cambridge. Tarr CL (1995) Primers for amplification and determination of mitochondrial control region sequences in oscine passerines. Molecular Ecology, 4, 527–529. Temeles EJ, Roberts WM (1993) Effect of sexual dimorphism in bill length on foraging behavior: an experimental analysis of hummingbirds. Oecologia, 94, 87–92. Templeton AR (1994) The role of molecular genetics in speciation studies. In: Molecular Ecology and Evolution: Approaches and Applications (eds Schierwater B, Streit B, Wagner GP, DeSalle R), pp. 455–477. Birkhäuser Verlag, Basel. Thomas DW, Thomas J (1996) Tchabal Mbabo botanical field survey. WWF, Cameroon. Urban EK, Fry CH, Keith S, eds (1986) The Birds of Africa (Vol. 2) Academic Press, London. Vogler AP, DeSalle R (1994) Diagnosing units of conservation management. Conservation Biology, 8, 354 – 363. Wiley EO (1988) Vicariance biogeography. Annual Review of Ecology and Systematics, 19, 513–542. Wright JB, Hastings DA, Jones WB, Williams HR (1985) Geology and Mineral Resources of West Africa. George Allen & Unwin Publishers, London. Wright S (1931) Evolution in a Mendelian population. Genetics, 16, 97–159.

This study is a component of a long-term project by TBS to test alternative speciation hypotheses in rainforests. KH is a NSERC postdoctoral fellow at the Center for Tropical Research and works on the phylogeography of African birds. DG is currently working on projects on the phylogeography of birds, mammals and insects and is now a Assistant Professor of Biology at Sonoma State University. BL, KO and YC are graduate students with interests in the molecular ecology of birds.

© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1505 –1516