Clinging to survival: Critically Endangered Chapman's pygmy chameleon Rhampholeon chapmanorum persists in shrinking forest patches
K RYSTAL A. TOLLEY,COLIN R. TILBURY,JESSICA M. DA S ILVA,GARY B ROWN Y ANKHO C HAPETA and C HRISTOPHER V. ANDERSON
Abstract The Critically Endangered Chapman’spygmycha- Introduction meleon Rhampholeon chapmanorum is endemic to the low verall extinction risk for chameleons is substantially elevation rainforest of the Malawi Hills in southern Malawi. higher than the global average of c. % for squamate Much of this forest has been converted to agriculture and O reptiles (Böhm et al., ), with % of species categorized it was uncertain whether chameleon populations have per- as threatened and % as Near Threatened (IUCN, ). sisted. We used current and historical satellite imagery to Most threatened species are forest specialists, with habitat identify remaining forest patches and assess deforestation. loss and degradation being the primary threat (Tolley We then surveyed forest patches for the presence of this et al., ). Given that forest chameleons are intolerant of chameleon, and assessed its genetic diversity and structure. transformed habitats, as forest is lost chameleon popula- We estimated that % of the forest has been destroyed tions decline and become locally extinct. There are five since , although we found extant populations of the Critically Endangered forest-living chameleon species in chameleon in each of the patches surveyed. Differentiation mainland Africa, all threatened by forest loss (IUCN, of genetic structure was strong between populations, suggest- ). They inhabit small rainforest patches on mountain ing that gene flow has been impaired. Genetic diversity was inselbergs that tend to have high endemicity (Menegon not low, but this could be the result of a temporal lag as et al., ; Conradie et al., ; Gereau et al., ; well as lack of sensitivity in the mitochondrial marker used. Lyakurwa et al., ). Thus, the loss of forest patches results Overall, the impact of forest loss is assumed to have led to in the extinction of endemic species (e.g. Betts et al., ), a large demographic decline, with forest fragmentation including chameleons. preventing gene flow. The Critically Endangered Rhampholeon chapmanorum Keywords Africa, Chapman’s pygmy chameleon, extinc- is a small (.–. cm body length), terrestrial chameleon en- tion, habitat loss, Malawi, reptile, Rhampholeon chapmano- demic to the low elevation rainforest of the Malawi Hills rum, threatened species near Nsanje in southern Malawi (Fig. , Plate ). These mountains rise to c. m and were once covered by dense rainforest, primarily on the eastern slopes between – m (Dowsett-Lemaire et al., ). At the time of the description of R. champanorum there were indications that substantial areas of forest were being lost (Tilbury, ). To safeguard the species, individuals ( females and males) from the type locality in the Malawi Hills were released in in a forest patch at Mikundi, KRYSTAL A. TOLLEY* (Corresponding author, orcid.org/0000-0002-7778-1963) † and JESSICA M. DA SILVA ( orcid.org/0000-0001-8385-1166) South African Malawi, c. km to the north (Dowsett-Lemaire et al., National Biodiversity Institute, Kirstenbosch Research Centre, Private Bag X7 ; Tilbury, ). Follow-up surveys at the release site Claremont 7735, Cape Town, South Africa. E-mail [email protected] in and confirmed the population was present COLIN R. TILBURY Department of Botany & Zoology, University of Stellenbosch, (C.R. Tilbury, unpubl. data). However, given the apparent Stellenbosch, South Africa loss of forest in the species’ natural range, it was uncertain GARY BROWN Les Terrasses, Sainte Innocence, France whether the remaining forest fragments still contained vi- ‡ YANKHO CHAPETA Museums of Malawi, Chichiri, Blantyre, Malawi able populations (Tolley et al., ). CHRISTOPHER V. ANDERSON ( orcid.org/0000-0001-7411-3102) Department of Rhampholeon chapmanorum is unlikely to inhabit trans- Biology, University of South Dakota, Vermillion, USA formed areas, and therefore forest degradation, fragmen- *Also at: School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, Johannesburg, South Africa tation and loss would be expected to cause an overall †Also at: Department of Zoology, Centre for Ecological Genomics & Wildlife demographic decline (e.g. Andrén, ), leading to local Conservation, University of Johannesburg, Johannesburg, South Africa extinctions. Furthermore, the loss of connectivity between ‡Also at: Biological Sciences Department, Mzuzu University, Luwinga, Mzuzu, Malawi populations affects ecosystem functioning, disrupts meta- Received February . Revision requested May . populations, reduces gene flow and increases genetic popu- Accepted September . lation structure even over small spatial scales. Thus, the
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, Downloadeddistribution, from https://www.cambridge.org/core and reproduction in any medium,. IP address: provided 170.106.33.22 the original work, onis 30 properly Sep 2021 cited. at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/termsOryx, Page 1 of 6 © The Author(s),. https://doi.org/10.1017/S0030605320000952 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 2 K. A. Tolley et al.
PLATE 1 Male Rhampholeon chapmanorum in Mabunga forest.
historical satellite imagery as a proxy to gauge demographic decline, and examined the genetic structure and diversity of the populations. Because there are no historical baseline estimates of population size or genetic structure/diversity, we assume that demographic decline is proportional to for- est loss. We predicted that fragmentation has disrupted gene flow, leading to strong differentiation of genetic structure and low genetic diversity.
Methods
We estimated the extent of forest loss using historical (/ ) and recent () satellite imagery of Malawi Hills FIG. 1 (a) Overview map of sites searched for Rhampholeon chapmanorum. Broken circle shows the locality for Mikundi from Google Earth (Google, Mountain View, USA). His- (exact locality obscured) and the square shows the Malawi Hills torical satellite images were of poor quality, and therefore locality, and (b) forest patches in the Malawi Hills superimposed images from both and were used. Polygons were on a topographic map with contour lines. The present extent of created around each of the forest patches for these time per- forest patches is shaded, with the estimated extent in / iods, exported as kmz files, mapped in QGIS . (QGIS shown by the dotted lines. Type locality is indicated by the black Development Team, ) and their areas quantified. The dot. Forest patches are numbered as in Table , except patches m contour was used as a guide to assess where low and , which were not surveyed. elevation rainforest might have originally occurred (e.g. Dowsett-Lemaire et al., ). effect of habitat loss on the population is both direct (via Logistical constraints did not allow us to survey all the population decline) and indirect because genetic diversity forest patches, but we were able to access two forest patches and gene flow are expected to diminish over time, reducing in the Malawi Hills and one at Mikundi, during – March the adaptive potential of the population (Haddad et al., ; (Fig. ). We walked forest trails at night to record cha- Leigh et al., ). The latter effects can be amplified in small meleons, using torchlight. Transects were defined by pres- populations because genetic drift is greater, leading to strong ence of trails and were therefore neither randomly chosen differentiation in genetic structure between populations, nor of a set distance or time. The location of each chameleon and inbreeding depression through fixation of deleterious encountered was recorded (± m) and a tissue sample was alleles and elevated homozygosity, reducing resilience (Hanski, taken from a subset of adult chameleons (c. mm tail clip) ). Information on population size, spatial genetic struc- for genetic analysis. All individuals were returned to the ture and diversity is therefore required to understand fully same perch on which they were found. the extinction risk for R. chapmanorum. To assess the genetic diversity within and between popu- To investigate these issues, we surveyed forest patches lations from each forest patch, an bp fragment of the in both Malawi Hills and Mikundi, to assess whether ND mitochondrial gene was sequenced for individ- R. chapmanorum populations are extant, quantified the ex- uals. Genomic DNA was extracted using salt extraction tent of forest loss in the Malawi Hills using recent and (MacManes, ). ND was amplified in a PCR reaction
Oryx, Page 2 of 6 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 30 Sep 2021 at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000952 Chapman’s pygmy chameleon 3
TABLE 1 Forest patch areas in the Malawi Hills in / and , with estimated extent lost. Patch numbers (in parentheses) correspond to the numbers in Fig. b.
Forest patch (no.) 1984/1985 (ha) 2019 (ha) % lost Type locality/Chief’s ancestors (1) 88.5 1.2 98.6 Patch 2 35.0 6.0 82.9 Mwanambweli (3 & 4) 44.7 16.5 (patch 3), 5.25 (patch 4) 51.3 Mabunga (5) 28.3 16.6 41.3 Total 196.5 45.55 76.8
volume of μl with . μlof mM dNTPs, . μlof mM We surveyed two of the five rainforest forest patches that MgCl, . μlof pmol forward (vMet) and reverse were furthest apart. The first patch (. ha) is c. . km (vTrp) primer (Cunningham & Cherry, ), . μlofMg+ south-west of the type locality (Fig. ; patch ). Local com- free buffer solution, . μl Taq polymerase, and – μlof munity leaders indicated that the Chief’s ancestors are bur- ng/μl DNA template. Thermal cycling included initial de- ied in the forest, and it is considered sacred ground. We were naturation for min at °C followed by cycles denatura- therefore only permitted to survey along a footpath that tion for sat °C, annealing for sat °C, extension skirted just inside the forest, where we recorded seven for sat °C, and final extension for min at °C. PCR adult chameleons (six females, one male) during one even- products were run on a % agarose gel and visualized under ing. Although we were not permitted further inside this a UV light to verify amplification. Amplicons were sequenced forest our observations of chameleons along the footpath using the forward primer at Macrogen Inc. (Amsterdam, The suggest there is a population within the sacred forest, despite Netherlands). Sequences were edited and aligned in Geneious the small size of this patch. The second, Mabunga forest R (Geneious, ) and deposited in GenBank (accession (Fig. ; patch ; . ha), is c. km to the south-west of the numbers: MT–MT). In addition, sequence from first. We did not observe chameleons on the forest edge, an individual originally sampled in from the type locality but we recorded individuals inside the forest along a was available on GenBank (AY) for comparison. footpath (seven females, three males) during one evening. Haplotype (h) and nucleotide (π) diversity for each locality Midway between the Chief’s ancestors and Mabunga for- was estimated (Tamura–Nei model of substitution, α = .) ests there are additional patches (Fig. ; patches , & ; using Arlequin . (Excoffier & Lischer, ). A haplotype the latter two at Mwanambweli forest) that we were not network was constructed using Network ... (Bandelt able to access, but given their location and extent, we assume et al., ; Polzin & Daneschmand, ) to assess shared there are chameleon populations present. Finally, we sur- haplotypes between localities, including the two GenBank se- veyed the forest patch at Mikundi (Fig. ) where chameleons quences from the type locality. To examine genetic structure had been released in . We recorded adult chameleons, between forest patches, an analysis of molecular variance plus juveniles and hatchlings (body size .–. cm) dur- Φ (AMOVA) was run, estimating pairwise ST in Arlequin. ing one evening, but we did not observe chameleons on the forest edge or outside the forest. Estimates of genetic diversity for all populations were Results within the range considered normal compared to other cha- The satellite imagery showed that the forest extent in the meleon and small-bodied reptile species (Table ; cf. Hague & Malawi Hills may have been c. ha in , in three Routman, ; Main et al., ; Rutherford et al., ). patches (Table ). As of , the patches were much re- There were no shared haplotypes between forest patches, in- duced, with two of the three patches fragmented into smal- cluding those previously sequenced from the original at the ler patches totaling c. ha, and the forest at the type locality type locality (Fig. ). Similarly, the AMOVA showed a signifi- Φ completely cleared (Fig. ). Currently, Mabunga forest is the cant difference between each of the forest patches ( ST = . , , largest intact patch (. ha), having been reduced from P . ) suggesting strong differentiation of genetic struc- ha since . Overall, we estimate c. % forest loss ture between populations and reduced gene flow across the during –. Given that the original extent (pre-) relatively small distances between the patches surveyed in is not known, the magnitude of loss could have been greater the natural range (Table ). as the original forest extent may have reached the lower slopes (e.g. Dowsett-Lemaire et al. ). There are also two ad- Discussion ditional low forest patches (at – m) on the western slopes of the Malawi Hills ( and . ha; patches and , Although populations of R. chapmanorum persist in the respectively, in Fig. ), but it is unknown if they provide Malawi Hills, the forest extent has been significantly re- suitable habitat for chameleons. duced to a total of c. ha. We can confirm there are extant
Oryx, Page 3 of 6 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 30 Sep 2021 at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000952 4 K. A. Tolley et al.
TABLE 2 Genetic diversity estimates (± SD) for Rhampholeon chapmanorum from three forest patches, showing haplotype diversity h, nucleotide diversity π and number of haplotypes H. Patch numbers (in parentheses) correspond to the numbers in Fig. b.
Forest patch (no.) n h ± SD π ± SD H Chief’s ancestors forest (1) 6 0.800 ± 0.172 0.003 ± 0.0021 4 Mabunga (5) 9 0.722 ± 0.159 0.002 ± 0.0014 5 Mikundi (released population) 4 0.833 ± 0.222 0.014 ± 0.0098 3
(Dowsett-Lemaire & Dowsett, ). This is approximately double our estimate of ha for , but the date of the aerial photographs used was not indicated. If the low ele- vation rainforest covered the mountain down to m (Dowsett-Lemaire et al., ) the forest extent could have been , ha historically. Although speculative, this could mean that % of the original forest has been lost. Our es- timate of a minimum of % forest loss since the mid s is likely to have resulted in a similar decline of R. chapma- norum, or at least a dramatic population decline over at least – FIG. 2 Haplotype network for R. chapmanorum with haplotypes years. Despite this, genetic diversity is not low, but this shaded according to locality where chameleons were sampled. can be explained by the lack of a linear relationship between Branch lengths are proportional to the number of mutations genetic diversity and population declines that result from and circle size is proportional to the number of individuals habitat loss (Pflüger et al., ). Erosion of genetic diversity with that haplotype. at detectable levels has a long time lag that is influenced by the effective population size, the number of generations and the degree of habitat loss and fragmentation (e.g. Tilman TABLE 3 Pairwise analysis of molecular variance for R. chapmano- et al., ; Krauss et al., ; Hoban et al., ;Wu Φ rum from three forest patches, with ST values (bottom matrix) and et al., ). Conversely, genetic structure responds more P values (top matrix). Patch numbers (in parentheses) correspond rapidly to habitat fragmentation (Keyghobadi et al., ; to the numbers in Fig. b. Pflüger et al., ) because gene flow is impeded by isola- Chief’s tion of local patches (Balkenhol et al., ). Populations of ancestors (1) Mabunga (5) Mikundi R. chapmanorum do not have reduced genetic diversity but Chief’s ancestors (1) , 0.010 , 0.001 they do have strong differentiation of genetic structure, with Mabunga (5) 0.088 , 0.001 no shared haplotypes despite the short distance between Mikundi 0.032 0.083 them. This indicates that gene flow has been disrupted over a scale of just a few kilometres and that the metapopu- lation is not intact, increasing extinction risk. populations in at least two of the five forest patches at Individuals from the released population at Mikundi are higher elevations, and we assume the other higher elevation in the same haplotype group as those from the type locality patches also hold populations. There are two patches at and the Chief’s ancestors sacred forest, although the haplo- lower elevation that we did not survey, and we cannot as- types occur at different frequencies. Given that the original sume that R. chapmanorum populations are present there, released population was small, the observed difference is given these are below m and low elevation rainforest probably the result of a founder effect in which the released was not considered to have occurred there (Dowsett-Lemaire individuals were not a random sample of the range of haplo- et al., ). Chameleons were not observed outside forest types in the original population. This can result in haplotype although observations at the sacred forest were made at slightly frequency differences between the new and source popula- degraded forest edge. Rhampholeon species are forest special- tions, and the effect can be magnified over time as genetic ists (Branch et al., ) and cannot tolerate transformed ha- drift changes the proportions of haplotypes. In this case, as- bitats, so where forest has been destroyed, the populations will suming the generation time of Rhampholeon is –. years, have gone locally extinct (as in the type locality). Conversely, there would have been – generations since the release, chameleons appear to be abundant at the Mikundi release which is sufficient time for genetic drift to have differentiated site, with chameleons of all age classes observed. the populations given the small founding population. From aerial photography, the extent of the Malawi Overall, our findings indicate there has been a severe Hills forest was previously estimated to be c. ha demographic decline with disruption of gene flow as a result
Oryx, Page 4 of 6 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 30 Sep 2021 at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000952 Chapman’s pygmy chameleon 5
of extreme forest loss. The effects of this decline have yet to be References fully manifested in the genetic composition, as genetic diver- sity and differentiation can continue to be negatively affected ANDRÉN,H.( ) Effects of habitat fragmentation on birds and mammals in landscapes with different proportions of suitable for many generations after habitat loss is halted, even with – habitat: a review. Oikos, , . subsequent population size increases (Pflüger et al., ). BALKENHOL, N., PARDINI, R., CORNELIUS, C., FERNANDES,F.& The genetic differentiation shows that gene flow has already SOMMER,S.() Landscape–level comparison of genetic diversity been disrupted, and the forest loss requires immediate at- and differentiation in a small mammal inhabiting different tention to prevent further loss of genetic diversity. Urgent fragmented landscapes of the Brazilian Atlantic Forest. – conservation action is needed, including halting of forest Conservation Genetics, , . BANDELT, H.-J., FORSTER,P.&RÖHL,A.() Median–joining destruction and recovery of habitat to promote connectivity. networks for inferring intraspecific phylogenies. Molecular Biology Although part of the Malawi Hills falls within a Key & Evolution, , –. Biodiversity Area (Matandwe Forest Reserve) most of the BETTS, M.G., WOLF, C., RIPPLE, W.J., PHALAN, B., MILLERS, K.A., forest falls outside the reserve boundary (c.f. BirdLife DUARTE,A.etal.() Global forest loss disproportionately erodes International, ), and the effectiveness of the forest reserve biodiversity in intact landscapes. Nature, , . BIRDLIFE INTERNATIONAL () World Database of Key Biodiversity is questionable, given that most of the destruction has been Areas. keybiodiversityareas.org [accessed February ]. within its boundaries. Although extending the reserve to en- BÖHM, M., COLLEN, B., BAILLIE, J.E., BOWLES, P., CHANSON, J., COX, compass all the forest patches would be a first step, measures N. et al. () The conservation status of the world’s reptiles. are needed to avert the destruction of the remaining patches. Biological Conservation, , –. In addition to R. chapmanorum, other endemic species could BRANCH, W.R., BAYLISS,J.&TOLLEY, K.A. ( ) Pygmy chameleons of the Rhampholeon platyceps complex (Squamata: occur there, which would add weight to the need to conserve Chamaeleonidae): description of four new species from isolated the remaining patches. Unfortunately, few surveys have been ‘sky islands’ of northern Mozambique. Zootaxa, , –. carried out in the area, with only widespread species of plants, BRITISH HERPETOLOGICAL SOCIETY () British Herpetological birds and butterflies recorded (Dowsett-Lemaire et al., ). Society: ethical policy and guidelines. The Herpetological Bulletin – A species action plan that defines conservation activities , . CONRADIE, W., BITTENCOURT-SILVA, G.B., ENGELBRECHT, H.M., is urgently needed for R. chapmanorum. However, both the LOADER, S.P., MENEGON, M., NANVONAMUQUITXO, C. et al. () planning and the recommended actions require strong lead- Exploration of the hidden world of Mozambique’s sky island forests: ership, personnel, stakeholder engagement, including with new discoveries of reptiles and amphibians. Zoosystematics & government departments, and sufficient funding to ensure Evolution, , –. success. Many international organizations support the pro- CUNNINGHAM,M.&CHERRY, M.I. () Molecular systematics of duction and execution of species action plans (e.g. IUCN African -chromosome toads (Anura: Bufonidae). Molecular Phylogenetics & Evolution, , –. Species Survival Commission, WWF) and the involvement DOWSETT-LEMAIRE,F.&DOWSETT, R.J. () Threats to the of these or analogous organizations will be essential to pre- evergreen forests of southern Malawi. Oryx, , –. vent the extinction of this species in the wild. DOWSETT-LEMAIRE, F., DOWSETT, R.J. & DYER,M.() Malawi. In Important Bird Areas in Africa and Associated Islands: Priority Sites Acknowledgements We thank the Museums of Malawi for assist- for Conservation (eds L.D.C. Fishpool & M.I. Evans), pp. –. ing with logistics and facilitating permits for this work; the South Pisces Publications, Newbury, & BirdLife International, African National Biodiversity Institute for support; and for funding, Cambridge, UK. the Scion Natural Science Association and the numerous contributors EXCOFFIER,L.&LISCHER, H.E.L. () Arlequin suite version .: to crowd-source funding at Rockethub. A new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources, Author contributions Project conceptualization, sourcing of fund- , –. ing: KAT, CVA; field work, analysis: KAT; field work, planning: CRT, GENEIOUS () geneious.com [accessed June ]. GB, YC; laboratory work, analysis: JMDS; writing: KAT, with assist- GEREAU, R.E., CUMBERLIDGE, N., HEMP, C., HOCHKIRCH, A., JONES, ance from all authors. T., K ARIUKI, M. et al. () Globally threatened biodiversity of the Eastern Arc mountains and coastal forests of Kenya and Tanzania. Conflicts of interest None. Journal of East African Natural History, , –. HADDAD, N.M., BRUDVIG, L.A., CLOBERT, J., DAVIES, K.F., Ethical standards This project abided by the ethical standards of GONZALEZ, A., HOLT, R.D. et al. () Habitat fragmentation and Oryx, and the ethical guidelines of the British Herpetological Society its lasting impact on Earth’s ecosystems. Science Advances, (2017). Research was carried out under clearance from the South ,e. African National Biodiversity Institute (#0001–2015), under agree- HAGUE, M.T.J. & ROUTMAN, E.J. () Does population size affect ment with the Museums of Malawi, Blantyre, a permit from the genetic diversity? A test with sympatric lizard species. Heredity, Malawian Department of Forestry—Forestry Research Institute, , –. Zomba (14 January 2016), and permission from the Malawian HANSKI,I.() Habitat loss, the dynamics of biodiversity, and a Department of National Parks and Wildlife, endorsed by the perspective on conservation. Ambio, , –. Malawian National Commission for Science and Technology, HOBAN, S., ARNTZEN, J., BRUFORD, M.W., GODOY, J.A., HOELZEL, Lilongwe (22 December 2015). A.R., SEGELBACHER, G. et al. () Comparative evaluation of
Oryx, Page 5 of 6 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 30 Sep 2021 at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000952 6 K. A. Tolley et al.
potential indicators and temporal sampling protocols for PFLÜGER, F.J., SIGNER,J.&BALKENHOL,N.() Habitat loss causes monitoring genetic erosion. Evolutionary Applications, , –. non-linear genetic erosion in specialist species. Global Ecology IUCN () The IUCN Red List of Threatened Species. iucnredlist.org and Conservation, ,e. [accessed February ]. POLZIN,T.&DANESCHMAND, S.V. () On Steiner trees and KEYGHOBADI, N., ROLAND, J., MATTER, S.F. & STROBECK,C.() minimum spanning trees in hypergraphs. Operations Research Among-and within-patch components of genetic diversity respond Letters , –. at different rates to habitat fragmentation: an empirical QGIS DEVELOPMENT TEAM () QGIS. Open Source Geospatial demonstration. Proceedings of the Royal Society B: Biological Foundation Project. qgis.osgeo.org [accessed April ]. Sciences, , –. RUTHERFORD, E.M., ONTANO, A., KANTOR,C.&ROUTMAN, E.J. KRAUSS, J., BOMMARCO, R., GUARDIOLA, M., HEIKKINEN, R.K., () Genetic variation across trophic levels: a test of the HELM, A., KUUSSAARI, M. et al. () Habitat fragmentation correlation between population size and genetic diversity causes immediate and time-delayed biodiversity loss at different in sympatric desert lizards. PLOS ONE, ,e. trophic levels. Ecology Letters, , –. TILBURY, C.R. () A new dwarf forest chameleon from Malawi. LEIGH, D.M., HENDRY, A.P., VÁZQUEZ-DOMÍNGUEZ,E.&FRIESEN, Tropical Zoology, , –. V.L. () Estimated six per cent loss of genetic variation in TILBURY, C.R. () Chameleons of Africa: An Atlas, Including the wild populations since the industrial revolution. Evolutionary Chameleons of Europe, the Middle East and Asia. nd edition. Applications, , –. Edition Chimaira, Frankfurt, Germany. LYAKURWA, J.V., HOWELL, K.M., MUNISHI, L.K. & TREYDTE, A.C. TILMAN, D., MAY, R.M., LEHMAN, C.L. & NOWAK, M.A. () () Uzungwa Scarp Nature Forest Reserve: a unique hotspot Habitat destruction and the extinction debt. Nature, , –. for reptiles in Tanzania. Acta Herpetologica, , –. TOLLEY, K.A., ALEXANDER, G.J., BRANCH, W.R., BOWLES,P.& MACMANES,M.() MacManes Salt Extraction Protocol.Museum MARITZ,B.() Conservation status and threats for African of Vertebrate Zoology Evolutionary Genetics Lab, Berkeley, USA. reptiles. Biological Conservation, , –. mvz.berkeley.edu/egl/protocols/extraction/MacManesSaltExtraction TOLLEY, K.A., MENEGON,M.&PLUMPTRE,A.() Rhampholeon [accessed January ]. chapmanorum.InThe IUCN Red List of Threatened Species . MAIN, D.C., VAN VUUREN, B.J. & TOLLEY, K.A. () Cryptic dx.doi.org/./IUCN.UK.–.RLTS.TA.en diversity in the common flap–necked chameleon Chamaeleo [accessed October ]. dilepis in South Africa. African Zoology, , –. WU, Y., CHEN, Y., CHANG, S.C., CHEN,Y.F.&SHEN, T.J. () MENEGON, M., DOGGART,N.&OWEN,N.() The Nguru Extinction debt in local habitats: quantifying the roles of random Mountains of Tanzania, an outstanding hotspot of herpetofaunal drift, immigration and emigration. Royal Society Open Science, diversity. Acta Herpetologica, , –. , .
Oryx, Page 6 of 6 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International doi:10.1017/S0030605320000952 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.22, on 30 Sep 2021 at 10:00:09, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000952