H E R P E TOLOGICA L JO U R N A L 21 : 5– 16, 201 1

Molecular  phylogenetics  of  Boulengerula  (Amphibia:   Gymnophiona:  Caeciliidae)  and  implications  for   taxonomy,  biogeography  and  conservation Simon P. Loader1,2, Mark Wilkinson2, James A. Cotton3,4, G. John Measey5,6, Michele Menegon7, Kim M. Howell8, Hendrik Müller2,9 & David J. Gower2 1

Department of Environmental Sciences, Institute of Biogeography, University of Basel, Switzerland 2 3

Department of Zoology, Natural History Museum, London, UK

School of Biological and Chemical Sciences, Queen Mary University of London, UK 4

5

Wellcome Trust Sanger Institute, Cambridge, UK

Applied Biodiversity Research Division, South African National Biodiversity Institute, Cape Town, South Africa

6

Department of Biodiversity and Conservation Biology, University of the Western Cape, Bellville, South Africa 7

Sezione di Zoologia dei Vertebrati, Museo Tridentino di Scienze Naturali, Trento, Italy 8

9

Department of Zoology and Marine Biology, University of Dar es Salaam, Tanzania

Institut für Spezielle Zoologie und Evolutionsbiologie mit Phyletischem Museum, Universität Jena, Germany

Phylogenetic relationships of the East African caeciliid Boulengerula were reconstructed using 12S, 16S and cytb mitochondrial gene sequences for 32 samples from Kenya and Tanzania. The generally well-supported and resolved phylogeny displayed the IROORZLQJUHODWLRQVKLSVDPRQJWKH¿YHQRPLQDWHVSHFLHVVDPSOHG B. boulengeri B. taitanus, B. niedeni  B. changamwensis, B. uluguruensis 7KLVUHVROXWLRQVXSSRUWVDIRUPHUO\SURSRVHGELSDUWLWLRQRIWKHJHQXVDQGGLIIHUVVLJQL¿FDQWO\IURPSUHYLRXV PRUSKRORJLFDO SK\ORJHQLHV 2XU DQDO\VHV LGHQWL¿HG JHQHWLF GLIIHUHQFHV EHWZHHQ VHYHUDO PW'1$ FODGHV WKDW SRWHQWLDOO\ represent undescribed species. If substantiated, the necessary taxonomic revision will have implications for conservation assessments that depend to an important extent upon sizes of distributions. Overall, there is a positive correlation between genetic and geographic distance among and within the main clades. The two lowland, coastal individuals sampled are nested ZLWKLQSULPDULO\PRQWDQHFODGHV'DWLQJDQDO\VHVVXJJHVWVRPHWHPSRUDOO\FRQJUXHQWGLYHUJHQFHVLQBoulengerula, but other divergences happened at different times and over a long period, perhaps extending back to the Oligocene/Eocene. Our results for Boulengerula suggest a role for relative long-term environmental stability in the origins of the Eastern Arc Mountains biodiversity hotspot. Key wordsFDHFLOLDQVPW'1$(DVWHUQ$UF0RXQWDLQV.HQ\D7DQ]DQLD

INTRODUCTION

erecting Afrocaecilia to receive the remaining species. 1XVVEDXP +LQNHO  GHVFULEHGDQHZVSHFLHVB. ¿VFKHUL, carried out a morphological phylogenetic anal\VLVRIWKHJHQXVDQGSODFHGWKHQRQPRQRSK\OHWLF LQ their tree) Afrocaecilia in the synonymy of Boulengerula. :LONLQVRQHWDO  UHPRYHGB. denhardti1LHGHQ from the synonymy of Schistometopum Dermophis) gregorii %RXOHQJHU GHPRQVWUDWHGWKDW1XVVEDXP  +LQNHO¶V  SK\ORJHQHWLFUHVXOWVZHUHQRWUREXVWDQG suggested that synonymy of Afrocaecilia with BoulengerulaZDVSUHPDWXUH0RVWUHFHQWO\0OOHUHWDO   have described an additional species, B. niedeni, which is FXUUHQWO\WKHRQO\,8&1³&ULWLFDOO\(QGDQJHUHG´FDHFLOLDQ ,8&1HWDO  Previous considerations of the molecular systematics of Boulengerula have included no more than three LQGLYLGXDOV :LONLQVRQHWDO)URVWHWDO /RDGHUHWDO5RHODQWVHWDO:ROOHQEHUJ  0HDVH\  =KDQJ  :DNH   ,Q DGGLWLRQ WRWD[RQRPLFVLJQL¿FDQFHDQH[SDQGHGPROHFXODUSHUspective on Boulengerula systematics will contribute to a broader understanding of the biology of the group.

T

he caecilian genus Boulengerula Tornier, 1896 is known from, and presumed to be restricted to, the Eastern Arc Mountains, lowland coastal forests, Albertine 5LIWDQG0DODZL6KLUH+LJKODQGVRI(DVW$IULFD )LJ  Boulengerula is the most speciose African caecilian genus, with the seven recognized species comprising about 39% and 4% of known extant African and global caecilian GLYHUVLW\UHVSHFWLYHO\ :LONLQVRQHWDO0OOHUHW DO:LONLQVRQ 1XVVEDXP,8&1  The centre of Boulengerula diversity is the Eastern Arc 0RXQWDLQV ($0 DJOREDOELRGLYHUVLW\KRWVSRW 0\HUV et al., 2000), from where four species have been described 0OOHUHWDO  ,PPHGLDWHO\ EHIRUH 7D\ORU¶V   PRQRJUDSKLF treatment of caecilians, four species of Boulengerula ZHUHUHFRJQL]HGB. boulengeri Tornier, 1896, B. uluguruensis%DUERXU /RYHULGJHB. changamwensis Loveridge, 1932, and B. taitanus Loveridge, 1935. Taylor  SDUWLWLRQHGWKHJHQXVUHWDLQLQJRQO\WKHW\SHVSHcies B. boulengeri in Boulengerula 7RUQLHU DQG

Correspondence:6LPRQ3/RDGHU,QVWLWXWHRI%LRJHRJUDSK\'HSDUWPHQWRI(QYLURQPHQWDO6FLHQFHV.OLQJHOEHUJVWUDVVH University of Basel, Basel 4056, Switzerland. E-mail: [email protected]

5

S. P. Loader et al .

METHODS Taxon and character sampling Specimens of Boulengerula were obtained by targeted ¿HOGZRUN GLJJLQJVRLOSLWIDOOWUDSSLQJ LQ.HQ\DDQG 7DQ]DQLDEHWZHHQDQG 7DEOH 7LVVXHV JHQerally liver and/or muscle) were preserved in 96% ethanol, ZLWKYRXFKHUVSHFLPHQV¿[HGLQIRUPDOLQDQGVWRUHG LQHWKDQRO6DPSOHVZHUHFROOHFWHGIURPDOO($0 localities where Boulengerula were previously recorded – (DVWDQG:HVW8VDPEDUD1JXUX7DLWD+LOOVDQG8OXJXUX 1XVVEDXP +LQNHO(PPULFK $WWHPSWV were made to collect Boulengerula in places in the EAM where they potentially occur but have not previously been IRXQG\LHOGLQJWKH¿UVWVSHFLPHQVIURP0DOXQGZHDQG 1JXXBoulengerulaZHUHQRWFROOHFWHG DQGUHPDLQXQreported) from Udzungwa, Mahenge, Rubeho, Ukaguru DQG1RUWKDQG6RXWK3DUH. The coastal forests of Tanzania and Kenya were not surveyed as extensively, so that absence of Boulengerula throughout much of these areas is less certain. This study includes representatives of all nominate Boulengerula species except for %¿VFKHUL and B. denhardtiQRQ($0VSHFLHV )LJ NQRZQRQO\ IURPWKHLUKRORW\SHV1RDWWHPSWZDVPDGHWRVDPSOHD reported population of B. changamwensis from the Shire +LJKODQGVRI0DODZL 1XVVEDXP +LQNHO RUD SRSXODWLRQUHFHQWO\IRXQGLQ1JDLDFHQWUDO.HQ\DE\6 6SDZOV SHUVFRPP WKDWPRVWFORVHO\UHVHPEOHVB. denhardtiDPRQJNQRZQVSHFLHV VHHEHORZ %DVHGRQWKH UHVXOWVRISUHYLRXVPROHFXODUDQDO\VHVWKH&HQWUDO$IULcan Herpele is sister to Boulengerula :LONLQVRQHWDO )URVWHWDO/RDGHUHWDO5RHODQWVHW DO DQGVRH. squalostoma was included as an outgroup. The rhinatrematid Epicrionops marmoratus was LQFOXGHGDVDPRUHGLVWDQW HJ5RHODQWVHWDO  second outgroup.

Fig. 1. Map of East Africa, with Eastern Arc mountain chain marked as dark areas. Collection localities of numbered samples are given in Table 1. This map covers the entire known range of Boulengerula. The locations of B. changamwensis (Malawi population), B. denhardti, B. cf. denhardti and %ÀVFKHULnot sampled in this study are also indicated. Boulengerula denhardti was described from an imprecise locality in the Tana River region. Abbreviations for montane blocks are: TH, Taita Hills; NP, North Pare; SP, South Pare; WU, West Usambara; EU, East Usambara; NU, Nguu; NG, Nguru; UK, Ukaguru; UL, Uluguru; ML, Malundwe; RU, Rubeho; UD, Udzungwa; and MA, Mahenge.

Phylogenetics An alignment of concatenated partial 12S, 16S and cytoFKURPHE cytb) sequences was assembled, based mostly on newly generated data. An outline of the methods for extraction, amplification and sequencing are given in *RZHUHWDO  DQG:LONLQVRQHWDO  'HWDLOV RIYRXFKHUVSHFLPHQVDQG*HQ%DQNDFFHVVLRQVDUHJLYHQ in Table 1. 6HTXHQFHV ZHUH DOLJQHG LQLWLDOO\ LQ &OXVWDO; XVLQJ default parameters, were manually adjusted, and had any ambiguously aligned sites removed. The alignment is available from the senior author upon request. To investigate levels of saturation we plotted transition/transversion ratios against numbers of transversions for pairs of sequences. Parsimony analyses were performed with 3$83 E 6ZRIIRUG 0D[LPXPOLNHOLKRRG 0/  analyses were performed with PAUP*4b6 and RAxML 6WDPDWDNLV WKHODWWHUXVLQJWZRSDUWLWLRQVU'1$ 6DQG6 DQGcytb)RU0/DQDO\VHVZHXVHGEHVW¿W PRGHOVDVGHWHUPLQHGE\0RGHOWHVW 3RVDGD &UDQGDOO XVLQJWKH$NDLNH,QIRUPDWLRQ&ULWHULRQ $,&  Empirical base frequencies were used in all analyses. PAUP* searches were heuristic, with 10 random addition sequence replicates and tree bisection recombination

Although caecilian natural history is generally little studLHGDQGSRRUO\XQGHUVWRRG *RZHU :LONLQVRQ  Boulengerula species have been the subject of pioneerLQJTXDQWLWDWLYH¿HOGHFRORJLFDOVWXGLHV *DUERULHDX  0HDVH\*RZHUHWDO0DORQ]D 0HDVH\ 0HDVH\ )XUWKHUPRUHB. taitanus is noteworthy for its remarkable reproductive biology including PDWHUQDOGHUPDWRSKDJ\DQGDOORSDUHQWDOFDUH .XSIHUHW DO )LQDOO\DVDGLVWLQFWLYH($0OLQHDJH Boulengerula has the potential to contribute to tests of hypotheses explaining the origins of the rich biodiversity of this region. Here we present a phylogenetic analysis of BoulengerulaXVLQJDVXEVWDQWLDOO\H[SDQGHGPW'1$ dataset. 6

B ou l e ng e rul a phylogeny

Table 1. Details of Boulengerula and outgroup samples used in analyses plus GenBank accession codes. Collection abbreviations: BMNH (Natural History Museum, London, UK), KBM (Kate McQuaid Field Series), JM (John Measey )LHOG6HULHV 0761 0XVHR7ULGHQWLQRGL6FLHQ]H1DWXUDOL7UHQWR,WDO\ 0: ÀHOGVHULHVWREHGHSRVLWHGLQ%01+  NMK (National Museum of Kenya), UMMZ (University of Michigan Museum, Ann Arbor, USA), and UTA (University of 7H[DV$UOLQJWRQ86$ 9RXFKHUVZHUHLGHQWLÀHGWKURXJKFRPSDULVRQVZLWKSXEOLVKHGGHVFULSWLRQVDQGW\SHPDWHULDO FR = Forest Reserve, NR = Nature Reserve. GenBank accession codes in italics are from previous studies. Type localities (or likely within 20 km) for nominal species indicated by asterisks. (Taita Hills: Kenya; Uluguru, Malundwe, Nguru, Nguu, West and East Usambara: Tanzania.) Voucher

Species

Locality

)RUHVWUHVHUYH

800=

Epicrionops marmoratus

Ecuador

&RWRSD[L

12S AY101206

16S AY101226

cytb AY101246

UTA 38889

Herpele squalostoma

&DPHURRQ

Mundemba

)1

)1

)1

1

10.$

B. taitanus

Taita Hills

1JDQJDR)5

)1

)1

)1

2

10.$

B. taitanus

Taita Hills

:XQGDQ\L

AY450614

AY450621

EU200986

3

10.$

B. taitanus

Taita Hills

:XQGDQ\L

)1

)1

)1

4

10.$

B. taitanus

Taita Hills

:XQGDQ\L

)1

)1

)1

5

10.$

B. taitanus

Taita Hills

:XQGDQ\L

)1

)1

)1

6

JM 228

B. taitanus

Taita Hills

&KDZLD)5

FN652689

FN652721

FN652753

7

JM 849

B. taitanus

Taita Hills

Kasigau

)1

)1

)1

8

10.$

B. niedeni

Taita Hills

Sagala*

)1

)1

)1

9

10.$

B. niedeni

Taita Hills

Sagala*

)1

)1

)1

10

10.$

B. changamwensis

&RDVWDO.HQ\D

&KDQJDPZH

)1

)1

)1

11

%01+

B. uluguruensis

Uluguru

0NXQJZH)5PDVO

)1

)1

)1

12

%01+

B. uluguruensis

Uluguru

0NXQJZH)5PDVO

)1

)1

)1

13

%01+

B. uluguruensis

Uluguru

0NXQJZH)5PDVO

)1

)1

)1

14

%01+

B. uluguruensis

&RDVWDOIRUHVW Tanzania

.D]L]XPEZL)5PDVO

)1

)1

)1

15

%01+

B. uluguruensis

Uluguru

Tegetero, 1000 m asl*

)1

)1

)1

16

JM 966

B. uluguruensis

Uluguru,

Tandai Village

)1

)1

)1



KBM 003

B. cf. uluguruensis

Malundwe

)1

)1

)1

18

%01+

B. cf. uluguruensis

1JXUX

.RPERUR1JXUX6RXWK)5

)1

)1

)1

19

0761

B. cf. uluguruensis

1JXUX

3HPED1JXUX6RXWK)5

)1

)1

)1

20

%01+

B. cf. uluguruensis

1JXUX

.RPERUR1JXUX6RXWK)5

)1

)1

)1

21

%01+

B. cf. uluguruensis

1JXUX

.RPERUR1JXUX6RXWK)5

)1

)1

)1

22

0:

B. cf. uluguruensis

1JXX

1JXX1RUWK)5

)1

)1

)1

23

0:

B. cf. uluguruensis

1JXX

1JXX1RUWK)5

)1

)1

)1

24

%01+

B. cf. boulengeri

:HVW8VDPEDUD

Lushoto

)1

)1

)1

25

%01+

B. cf. boulengeri

:HVW8VDPEDUD

0D]XPEDL)5

)1

)1

)1

26

JM 150

B. boulengeri

East Usambara

6KDPEDQJHGD$PDQL15

)1

)1

)1



%01+

B. boulengeri

East Usambara

1LOR)5

)1

)1

)1

28

%01+

B. boulengeri

East Usambara

$PDQL.ZDPNRUR)5

AY450613

AY450620

EU200987

29

%01+

B. cf. boulengeri

:HVW8VDPEDUD

0D]XPEDL)5

)1

)1

)1

30

%01+

B. cf. boulengeri

:HVW8VDPEDUD

0D]XPEDL)5

)1

)1

)1

31

%01+

B. cf. boulengeri

:HVW8VDPEDUD

$PEDQJXOD)5

)1

)1

)1

32

%01+

B. cf. boulengeri

:HVW8VDPEDUD

$PEDQJXOD)5

)1

)1

)1

branch swapping. RAxML searches employed 200 runs on distinct random starting trees. Bayesian analysis was SHUIRUPHGXVLQJ0U%D\HV +XHOVHQEHFN 5RQTXLVW 2001) both with and without partition using parameters HVWLPDWHGE\0RGHOWHVW'DWDZHUHDQDO\VHGLQUXQVZLWK 2,000,000 generations, trees sampled every 1000 gen-

HUDWLRQV:HXVHG7UDFHU 5DPEDXW 'UXPPRQG  WRFKHFNWKDW0&0&UXQVKDGUHDFKHGVWDWLRQDULW\ DQGWKH¿UVWWUHHVZHUHGLVFDUGHGDV³EXUQLQ´ 7RWHVWZKHWKHUWKHGDWDKDYHVLJQL¿FDQWO\PRUHKLHUarchical structure than expected by chance alone we used DSHUPXWDWLRQWDLOSUREDELOLW\ 373 HPSOR\LQJUDQ7

S. P. Loader et al .

Table 2. Support for nodes a–w labelled in Fig. 2 (B = bootstrap, MP= maximum parsimony, ML = maximum likelihood, BPP = Bayesian posterior probability, DI = decay index).

GRPL]DWLRQVRIWKHGDWD )DLWK &UDQVWRQ$UFKLH 1989). The partition homogeneity/incongruence-length difference test implemented in PAUP* was used to deterPLQHLIGLIIHUHQWSDUWLWLRQVRIWKHGDWDKDYHVLJQL¿FDQWO\ GLIIHUHQWVLJQDOV6XSSRUWIRUFODGHVZDVTXDQWL¿HGZLWK Bayesian posterior probabilities, bootstrap proportions )HOVHQVWHLQ EDVHGRQSVHXGRUHSOLFDWHVDQG GHFD\ LQGLFHV %UHPHU   GHWHUPLQHG E\ HQIRUFing converse topological constraints. PAUP* was used to implement a priori .LVKLQR±+DVHJDZD .LVKLQR  +DVHJDZD DQG7HPSOHWRQ 7HPSOHWRQ WHVWV XQGHU0/DQGSDUVLPRQ\UHVSHFWLYHO\ RIQXOOK\SRWKHVLVRIQRQVLJQL¿FDQWGLIIHUHQFHVEHWZHHQRSWLPDODQG VHOHFWHGVXERSWLPDOWUHHVUHÀHFWLQJSULRUWD[RQRPLFELRgeographic and phylogenetic hypotheses.

a b c d e f g h i j k l m n o p q r s t u v w

Biogeography and dating 5RHODQWVHWDO  HVWLPDWHGWKDWWKHVSOLWEHWZHHQ Herpele and BoulengerulaRFFXUUHG0D FRQ¿GHQFHLQWHUYDOV±0D EDVHGRQDEURDGVDPSOLQJ RIDPSKLELDQVPXOWLSOHJHQHWLFPDUNHUV LQFOXGLQJQXclear genes) and 22 calibration points. In the absence of any primary calibration points, we used this estimate as a single secondary point calibration for a dating analysis of the unpartitioned ingroup data using BEAST 1.4.6 'UXPPRQG 5DPEDXW ZLWKGHIDXOWVHWWLQJV7ZR independent chains each of 10 million generations were run under the exponential uncorrelated clock models of 'UXPPRQGHWDO  XVLQJ
Unpartitioned B MP B ML BPP 100 98 100 100 100 100 100 100 100 96 92 100 63  91 86 89 100 86  99 100 96 100 98 96 100 99 98 85 100 100 100 99 99 100 96 98 100 99 91 100 100 100 100 56   99 94 100 90 83 100  81  100 100 100 95 83 89 100 100 100 68 62 85

Partitioned BPP B ML 100  100 100 100 100 100 99 91 85 100 91 99 91 100 98 100 100 83 100 100 100 100 100 100 100 100 100 100 100   100 96 100 100 96 91 100 100   100 100 89 

', 5 31 41 10 0 0 0 0 0 0 8 0 8 11 30 0  6    59 4

EHWZHHQJHRJUDSKLFGLVWDQFHV P DQGJHQHWLF XQFRUUHFWHG S  GLVWDQFHV ZHUH FRQGXFWHG XVLQJ 0$17(/ &DYDOFDQWL ZLWKSHUPXWDWLRQV

RESULTS 7KHFRQFDWHQDWHGDOLJQPHQWFRPSULVHVVLWHV 12S, 423 16S and 633 cytb. Of these, 823 are constant, 156 variable but parsimony uninformative, and 381 parVLPRQ\LQIRUPDWLYH7KHGDWDDSSHDUQRQUDQGRP 373   DQGQRWVLJQL¿FDQWO\KHWHURJHQHRXV ,/'WHVWP  0.55). Plots reveal little evidence of saturation within any partition, including 3rd position sites in cytb. The optimal unpartitioned ML phylogeny is shown in )LJZLWKFODGHVXSSRUWYDOXHVIRUGLIIHUHQWDQDO\VHV reported in Table 2. Analyses using various methods and parameters yielded trees with almost identical and generally well-supported relationships. The optimal ML and %D\HVLDQWUHHV IRUXQSDUWLWLRQHGDQGSDUWLWLRQHGGDWD  have topologies that are identical to each other and to one RIWKH¿YHPRVWSDUVLPRQLRXVWUHHV 037V 7KHUHPDLQing MPTs differ only in the relationships among the three B. boulengeriVDPSOHVLQFODGHT )LJ DQGLQZKHWKHU 8

B ou l e ng e rul a phylogeny

Fig. 2. Maximum likelihood tree from unpartitioned data. (–ln likelihood = 6666.26813). Nucleotide frequencies: A = 0.34310, C = 0.2508, G = 0.1447, T = 0.26140. Number of substitution parameters = 6; rate matrix = 1.00, 3.2018, 1.000, 1.000 and 8.6411; gamma shape parameter = 0.8971; proportion of invariant sites = 0.3935). Letters a–w label nodes referred to in Tables 2–4. Quantitative support values for nodes are given in Table 2. 6KDGHGER[HVLQGLFDWHERXQGVRIÀYHQRPLQDWHVSHFLHV

VDPSOHRU :8VDPEDUDB. cf. boulengeri) alone is sister to clade q. Monophyly of Boulengerula is not RYHUZKHOPLQJO\VXSSRUWHG 7DEOH EXWZHGRQRWGRXEW it on the basis of more extensive molecular phylogenetic DQDO\VHVRIFDHFLOLDQPWDQGQXFOHDU'1$QRWUHSRUWHG here. Each of the four nominate species represented by PXOWLSOHLQGLYLGXDOV B. boulengeri, B. niedeni, B. taitanus and B. uluguruensis) were recovered as strongly supported monophyla. The basal split in the ingroup is between B. boulengeri and all other lineages, consistent ZLWK7D\ORU¶V  SDUWLWLRQLQJRIBoulengerula. The recently described B. niedeni is sister to its geographically and phenotypically closest neighbour B. taitanus,

and these are sister to a clade comprising B. changamwensis and B. uluguruensisSOXV0DOXQGZH1JXXDQG 1JXUXSRSXODWLRQVRIB. cf. uluguruensis. Pairwise uncorrected distances were calculated for DOOWHUPLQDOV VHH6XSSRUWLQJ2QOLQH0DWHULDO 'LIIHUences between all currently recognized species exceed 3DLUZLVHGLVWDQFHVEHWZHHQB. boulengeri and other Boulengerula ! DUHVLPLODUWRWKRVHEHWZHHQBoulengerula and Herpele ! 8OXJXUXB. uluguruensis are 4% and 2% different from B. cf. uluguruensis from 1JXX1JXUXDQGIURP0DOXQGZHUHVSHFWLYHO\/HVV VXEVWDQWLDOGLIIHUHQFHV ± H[LVWDPRQJB. boulengeri from various Usambara localities. 9

S. P. Loader et al .

Table 3. Results of topological tests of a priori phylogenetic hypotheses (T = Templeton Test, KH = Kishino– +DVHJDZD7HVW   VLJQLÀFDQWO\EHWWHUWKDQWKHDOWHUQDWLYH  Z VLJQLÀFDQWO\ZRUVHWKDQWKHDOWHUQDWLYH Nodes are labelled in Fig. 2. Hypothesis Monophyletic Boulengerula taitanus Monophyletic Boulengerula niedeni Monophyly of Taita Hills Boulengerula Monophyletic Boulengerula uluguruensis Monophyletic Boulengerula boulengeri Monophyletic Usambara Boulengerula 1XVVEDXP +LQNHO¶VPRUSKRORJLFDOSK\ORJHQ\ Monophyletic coastal Boulengerula Monophyletic EAM montane Boulengerula

1RGH b c d h q v

Judged by the KH and Templeton tests, the best trees FRQVWUDLQHGWREHFRQVLVWHQWZLWK1XVVEDXP +LQNHO¶V   DQG :LONLQVRQ HW DO¶V   PRUSKRORJLFDO phylogenies of Boulengerula DUHVLJQL¿FDQWO\VXERSWLPDOIRUERWK0/DQGSDUVLPRQ\ P  VHH7DEOH  2WKHUVLJQL¿FDQWO\VXERSWLPDOUHVROXWLRQVDUHWKRVH

T P<0.001* P<0.0001* P<0.05* P=0.103 P<0.05* P<0.0001* P<0.0001w P<0.0003w P<0.0001w

KH P<0.01* P<0.0001* P=0.124 P=0.053 P= P<0.0001* P<0.0001w P<0.0001w P<0.0001w

EHVWWUHHVLQFOXGLQJDQRQPRQRSK\OHWLF³Afrocaecilia´ QRQboulengeri Boulengerula) P<0.001), paraphyletic Usambara B. boulengeri P<0.0001), monophyletic coastal P DQGPRQRSK\OHWLF($0VDPSOHV P<0.0001). 0RQRSK\O\RI:HVW8VDPEDUDB. boulengeri cannot be UHMHFWHG P! 

Table 4. Bayesian estimates of divergence dates (Ma) with different priors assuming divergence of Herpele and BoulengerulaDW0D/HWWHUVLQÀUVWFROXPQUHIHUWRQRGHVRQ)LJDQG)LJ FRQÀGHQFHFUHGLELOLW\ LQWHUYDO$EVROXWH LQWHUYDOUHVFDOHGWRWDNHLQWRDFFRXQWWKHFRQÀGHQFHLQWHUYDO ²0D RIWKHVLQJOH secondary calibration point (Roelants et al., 2007).

a b c d e f g h i j k l m n o p q r s t u v w

Mean  15.18 3.23 41.28 4.39 2.15  9.94  1.50  12.92 30.42  62.21 10.56  18.35 10.09 1.59  34.25 90.41

3'$ 95% 5.69–9.45 11.24–18.68 ± ± 3.25–5.40 1.59–2.64 ± ± 13.53–22.48 1.11–1.84 1.31–2.18 ± ± 33.12–55.05 ± ± ± 13.59–22.59 ± ± 13.25–22.02 25.36–42.15 ±

Absolute ± 8.35–23.28 ± ± ± 1.18–3.3 3.39–9.46 5.46–15.24 10.05–28.02 0.82–2.3 ± ± ± 24.59–68.6 34.2–95.41 5.81–16. 2 2.95–8.24 10.09–28.14 ± ± ± 18.83–52.523 ±

Mean 4.64 9.65 2.30  2.91 1.43    1.06 1.20 8.23 20.48 31.65  6.63 3.54  6.08 1.03 11.04 23.85  10

Yule 95% ± ± ± 20.64–34.30 2.16–3.58 ± 2.94–4.89 ± ± ± 0.89–1.48 6.09–10.13 15.16–25.20 23.43–38.95 35.16–58.43 4.91–8.16 2.62–4.36 8.66–14.40 ± ± 8.18–13.59 ± 56.53–93.95

Mean   1.40 20.49 1.68 0.81  3.96  0.58  5.32 14.62 24.45 39.89 3.92 2.12  3.53 0.56   68.88

&RDOHVFHQW 95% 2.00–3.32 ± ± ± ± 0.60–0.99 ± 2.94–4.88 ± ± 0.56–0.94 3.94–6.55 ± 18.10–30.08 29.54–49.09 2.90–4.82 ± 5.54–9.21 2.61–4.34 0.42–0.69 5.18–8.62 12.85–21.36 ±

B ou l e ng e rul a phylogeny

Fig. 3. Ultrametric tree from Bayesian relaxed clock analysis using exponential uncorrelated clock PDA priors. PDA, Yule and coalescent prior scales based on point calibration of Herpele–Boulengerula split at 96.7 Ma. Letters DERYHEUDQFKHVLGHQWLI\QRGHVUHIHUUHGWRLQ7DEOHV²DQG)LJ%DUVLQGLFDWHFRQÀGHQFHLQWHUYDOV VHH Table 4).

In unconstrained optimal trees, sister lineages are often those that would be predicted from geographical proxLPLW\LQFOXGLQJ(DVW:HVW8VDPEDUD1JXUX1JXX 0DOXQGZH8OXJXUX0DQWHOWHVWVLQGLFDWHVLJQL¿FDQW positive correlation between genetic and geographic GLVWDQFH IRU WKH FRPSOHWH DOLJQPHQW P   IRU ³Afrocaecilia´RQO\DQGIRUB. boulengeri RQO\ P<0.05). 'DWLQJUHVXOWVIURPWKH¿[HGWUHHWRSRORJ\DQDO\VHV )LJ7DEOH DUHYHU\VLPLODUWRWKRVHDOORZLQJWRSRORJ\WRYDU\ QRWVKRZQ 7KH3'$SULRUUHVXOWHGLQ consistently oldest estimates of divergence, with estimates based on coalescent and Yule priors about 20% DQG  \RXQJHU UHVSHFWLYHO\ 1R SULRU DOORZV XV WR reject the null hypothesis of temporal congruence of 1) WKUHHJHRJUDSKLFDOO\DGMDFHQWVSOLWV1JXUX±1JXX QRGH O  (DVW:HVW 8VDPEDUD QRGH U  8OXJXUX±0DOXQGZH QRGHL DQG WKHGHHSHUGLYHUJHQFHRIB. changamwensis QRGHQ DQGWKHVSOLWEHWZHHQWKH7DLWD+LOOVB.

niedeni and B. taitanus QRGHG EXWHDFKSULRUUHMHFWVWKH hypothesis that the latter are contemporaneous with the IRUPHU7KHWRSRORJ\RIWKHWUHH )LJ LPSOLHVWKDWWKH two divergences between coastal forest and EAM Boulengerula, the Kazizumbwi–Uluguru B. uluguruensis split QRGHJ DQGWKHB. changamwensis–B. uluguruensis split QRGHQ ZHUHDV\QFKURQRXVDQGWKHGDWLQJVXJJHVWVWKH former occurred substantially more recently than the latWHU 7DEOH 

DISCUSSION Taxonomy Previous molecular phylogenetic analyses of Boulengerula have not included more than one specimen per species and no more than three nominate species. Our phylogeQHWLFUHVXOWVDUHHQWLUHO\FRQVLVWHQWZLWKSUHYLRXV¿QGLQJV of a deep divergence between B. boulengeri and B. taita11

S. P. Loader et al .

nus or B. uluguruensis :LONLQVRQHWDO)URVWHWDO /RDGHUHWDO5RHODQWVHWDO=KDQJ  :DNH DQGWKHSRVLWLRQRIB. boulengeri outside B. taitanusB. uluguruensis :ROOHQEHUJ 0HDVH\ =KDQJ :DNH 1HZ¿QGLQJVLQFOXGLQJWKHVLVWHU group relationships between B. taitanusB. niedeni and between B. changamwensisB. uluguruensis are also well VXSSRUWHG2XURWKHUPDLQQHZ¿QGLQJVDUHGLYHUJHQWPLtochondrial lineages representing potentially undescribed species of B. cf. boulengeriLQ:HVW8VDPEDUDDQGB. cf. uluguruensisLQ0DOXQGZH1JXXDQG1JXUX 2XUUHVXOWVDUHFRQVLVWHQWZLWK7D\ORU¶V  ELSDUtition of the genus into Boulengerula boulengeri) and Afrocaecilia changamwensis, taitanus, uluguruensis). 1XVVEDXP +LQNHO  ZHUHXQLPSUHVVHGE\WKHWZR FKDUDFWHUVXVHGWRGLDJQRVHWKHVHWZRJURXSV SUHVHQFH absence of inner mandibular teeth with anterior tongue DWWDFKPHQWSUHVHQFHDEVHQFHRIGRUVDOH[SRVXUHRIPHVHWKPRLG7D\ORU1XVVEDXP :LONLQVRQ  and provided morphological phylogenetic evidence of the non-monophyly of Afrocaecilia as additional support for their placement of this genus into the synonymy of Boulengerula. Our new molecular phylogenetic data build RQ:LONLQVRQHWDO¶V  UHDQDO\VLVRI1XVVEDXP  +LQNHO¶V  PRUSKRORJLFDOGDWDLQGHPRQVWUDWLQJWKDW the synonymy of Afrocaecilia with Boulengerula is not supported on phylogenetic grounds. However, we refrain from resurrecting Afrocaecilia, not least because of incomplete molecular sampling. In particular, the position of the as yet unsampled B. denhardti will have important implications for the diagnosis of Boulengerula and Afrocaecilia VKRXOGWKHODWWHUEHUHLQVWDWHG JLYHQWKDWOLNHB. boulengeri, this species lacks inner mandibular teeth and has an anteriorly attached tongue. The close relationship between B. uluguruensis and B. changamwensis is unsurprising given that these are morphologically similar species differentiated from each other mainly in the arrangement of the palatal dentition DQGQXPEHUVRIDQQXOL HJ7D\ORU1XVVEDXP  Hinkel, 1994). These two species also include the only lowland Boulengerula sampled in our study. Boulengerula changamwensis was described originally from FRDVWDOVRXWKHUQ.HQ\D VHH0DORQ]D 0OOHU  but has also been reported from the Shire Highlands of Malawi, some 1400 km to the south, based on a single VSHFLPHQFROOHFWHGLQWKHV 1XVVEDXP +LQNHO  *LYHQWKHJUHDWGLVWDQFHDQGODFNRILQWHUYHQLQJ records, we doubt that the Malawi and Kenya samples represent the same species. However, although montane East African caecilians appear in general to have small ranges, we know much less about range size of their lowland counterparts. The only other lowland Boulengerula we sampled, from coastal forest at Kazizumbwi, is a B. uluguruensis showing little genetic distinction from its Uluguru monWDQHFRQVSHFL¿FV7KLV.D]L]XPEZLVSHFLPHQLVWKHRQO\ extra-Uluguru record for the species. The lack of B. uluguruensis records between Uluguru and the coast can be H[SODLQHGE\ORFDOH[WLQFWLRQDQGRUODFNRIVDPSOLQJWKH latter cannot be rejected because little dedicated caecilian ¿HOGHIIRUWLQ7DQ]DQLDWRGDWHKDVIRFXVHGRQWKHORZODQGV

:LWKLQWKHB. uluguruensis clade, all analyses retrieve the same relationships that show some sign of geographical VWUXFWXULQJDFODGHFRPSULVLQJVDPSOHVIURPWZRORFDOLWLHV 7DQGDL7HJHWHUR WKDWDUHVHSDUDWHGE\DERXWNP is sister to a clade that includes the three samples from 0NXQJZH DERXWNP:RI7DQGDL ZLWKWKHFRDVWDO Kazizumbwi sample from a further 100 km west. The sister pairing of Taita Hills B. taitanus and B. niedeni is unsurprising given their geographic proximity DQGPRUSKRORJLFDOVLPLODULW\ 0OOHUHWDO 7KH single B. taitanus sampled from Kasigau is a well supSRUWHGEXWVRPHZKDWJHQHWLFDOO\GLVWDQW  VLVWHU JURXSRILWVFRQVSHFL¿FVIURPWKHPDLQ7DLWDPRQWDQH block. Kasigau is some 60 km to the south of the main Taita block, much further than the type locality of B. niedeni6DJDOOD 0OOHUHWDO¿J ,QGLYLGXDOV from the Kasigau population are notably darker than B. taitanusIURPWKHPDLQEORFN +0OOHUSHUVREV DQG this population merits more detailed study. 2XU SK\ORJHQHWLF DQDO\VHV UHFRYHUHG WKUHH PW'1$ clades among the Usambara populations. Boulengerula boulengeri was originally described from material from East Usambara, and our East Usambara samples form a clade that can be readily referred to this species. The H[LVWHQFHRIDQXQGHVFULEHGQHZVSHFLHVIURP:HVW8Vambara has been suggested based on morphological data 9HVWHUJDDUG&KDQQLQJ +RZHOO DQGDW OHDVW VRPH:HVW 8VDPEDUD SRSXODWLRQV LQFOXGH VSHFLmens with more annuli and vertebrae than those from East 8VDPEDUD*HQHWLF XQFRUUHFWHGS GLVWDQFHVEHWZHHQ RXU:HVWDQG(DVW8VDPEDUDVDPSOHV XSWRVHH Supporting Online Material) provide some support for WKLVVXJJHVWLRQEXWWKHSDUDSK\O\RIRXU:HVW8VDPEDUD VDPSOHV LQFOXGLQJWKHSDUDSK\O\RIWKHWKUHHVDPSOHV from Mazumbai) and the wide range of genetic distances ±VHH6XSSRUWLQJ2QOLQH0DWHULDO LQGLFDWHD complexity that will require more detailed investigation before any taxonomic action. Other potential new species indicated by our results are B. cf. uluguruensisIURP0DOXQGZH1JXUXDQG1JXX %RWKWKHSK\ORJHQ\ UHFLSURFDOPRQRSK\O\RILQGLYLGXals from these three separate montane areas) and genetic GLVWDQFHV !IURP8OXJXUXB. uluguruensis) suggest that each area might harbour its own endemic species of Boulengerula, and preliminary observations suggest differences also in colour and overall body proportions. Although the known amphibian diversity of the EAM is already high, there are molecular and traditional data that provide strong evidence that the numbers of species and levels of endemism remain substantially underestimated HJ0HQHJRQHWDO%ODFNEXUQ/RDGHUHW DO SDUWLFXODUO\LQWKH1JXUX0RXQWDLQV 0HQHJRQ et al., 2008). $SKRWRJUDSKRIDFDHFLOLDQIURP1JDLD)RUHVW.HQ\D QRWVDPSOHGLQRXUVWXG\ SURYLGHGE\66SDZOVDSSHDUV to represent a Boulengerula, based on body and head shape, lack of externally visible eyes, unsegmented terminal shield and lack of secondary annular grooves. If FRQ¿UPHGWKLVZRXOGEHWKHPRVWQRUWKHUO\UHFRUGIRU the genus, and the only one north of the equator. The inGLYLGXDOKDVDERXWDQQXOL PRUHWKDQDUHNQRZQIRU 12

B ou l e ng e rul a phylogeny

VWDQWLDOO\ROGHUWKDQWKH3OHLVWRFHQH WR0D DQG 3OLRFHQH WR0D ±WKHWZRFOLPDWLFDOO\YRODWLOH epochs generally cited as most important for the origins of H[WDQW($0GLYHUVLW\ HJ%HUHVIRUGHWDO%RZLH HWDO.DKLQGRHWDO)MHOGVn %RZLH %ODFNEXUQ 0HDVH\ 6XEVWDQWLDOHQYLURQPHQtal change also occurred in Africa during the Miocene 7UDXWKHWDO DQGRWKHUVWXGLHVKDYHLQGLFDWHGWKDW DWOHDVWVRPHRWKHU($0KHUSHWRIDXQDOOLQHDJHVDUH³ROG´ HJRhampholeon0DWKHHHWDOHoplophryne, 9DQ %RF[ODHU HW DO  EUHYLFLSLWLQHV 5RHODQWV HW DONectophrynoides–Churamiti, Van Bocxlaer et DO 7KHVH¿QGLQJVVXJJHVWWKDWDEURDGHUKLVWRULcal perspective will be required to formulate appropriate K\SRWKHVHVRI($0ELRWLFGLYHUVL¿FDWLRQ(YLGHQFHRI LPSRUWDQWSUH3OLRFHQH($0GLYHUVL¿FDWLRQFRPHVDOVR IURPWUDGLWLRQDOGLVWULEXWLRQDOGDWD -HW]HWDO DQG YHU\KLJKOHYHOVRIHQGHPLVP HJ%DVLOHZVN\ 6FKDUII-RKDQVRQ :LOODVVHQ:DUXL  -RFTXp  9DQGHQVSLHJHO  6WDQOH\ HW DO 0DWKHHHWDO7LOEXU\HWDO'DYHQSRUW HWDO0DULDX[HWDO IXUWKHUVXEVWDQWLDWing prolonged history in explanations of the origins and PDLQWHQDQFHRIWKH($0ELRGLYHUVLW\KRWVSRW VHHDOVR %XUJHVVHWDO  High species diversity in the EAM has been contrasted with substantially lower diversity in East African lowland IRUHVWV %XUJHVVHWDO EXWPROHFXODUSK\ORJHQHWLF assessments of lowland genetic diversity are very rare to GDWH)RUBoulengerula, our tree is consistent with EAM origins of lowland populations, with the two lowland individuals nested within different, primarily montane lineages. Similar patterns have been recovered for East $IULFDQOL]DUGV 0DWWKHHHWDO DQGDQJLRVSHUP SODQWV 0|OOHU HW DO  7KH FRPPRQDOLW\ RI WKLV pattern across other East African lineages is another important hypothesis to test in the effort to better understand the EAM hotspot.

any species except the 161 of B. denhardti and 186 of %¿VFKHUL DQGLWVSXUSOLVKFRORXU GDUNHUGRUVDOO\ ZLWK whitish annular grooves readily distinguishes it from B. changamwensis, %¿VFKHUL, B. niedeni, B. taitanus and B. uluguruensis. More detailed study will be required to determine whether this form is distinct also from B. denhardti.

Biogeography 7KHWZRPDLQ DQGQRWPXWXDOO\H[FOXVLYH K\SRWKHVHV put forward to explain the high diversity and local endemism of the EAM biota are long-term environmental stability and habitat fragmentation across an island-like FKDLQRIPRXQWDLQV %XUJHVVHWDO/RYHWWHWDO 2005). Testing these and other possible explanations depends ultimately on empirical data for different lineages occurring in the region. In particular, robust phylogenies and relative and/or absolute estimates of divergence dates should promote the required multi-taxon analyses. Molecular phylogenetic studies of EAM organisms to date HJ 5R\  *UDYOXQG  0|OOHU HW DO  /LQGTTYLVW $OEHUW%RZLHHWDODE %HUHVIRUGHWDO%ODFNEXUQ 0HDVH\ KDYH drawn attention to the presumed importance of Pleistocene DQG3OLRFHQHFOLPDWHÀXFWXDWLRQVLQVKDSLQJGLYHUVL¿FDtion here, but few general patterns have yet emerged. Excluding the position of our two coastal samples and the relationship between B. boulengeri and other lineages, geographical proximity is a good predictor of sister-group relationships in our phylogeny, and this is underlined by our Mantel test results, which corroborate significant positive correlation between geographic and genetic distance within and among clades. This is consistent with short-range dispersal and/or vicariance being a more imSRUWDQWKLVWRULFDOGULYHURIGLYHUVL¿FDWLRQWKDQORQJUDQJH dispersal. )DLOXUHWRUHMHFWWKHK\SRWKHVLVWKDWWKUHHRIWKHPDLQ divergences in the Boulengerula tree occurred contemporaneously is consistent with a possibly regional scale abiotic event that promoted the expansion/fragmentation of moist forest habitat and dispersal/fragmentation of BoulengerulaSRSXODWLRQV VHH)LJDQG7DEOH 7KH divergence between the Taita Hills B. neideni and B. taitanus was not contemporaneous with this putative burst of speciation but occurred substantially earlier, providing evidence of a more complex history than a single cycle RI($0KDELWDWH[SDQVLRQIUDJPHQWDWLRQ*LYHQWKHODFN of robust primary calibration points, inferred absolute GDWHVPXVWEHYLHZHGZLWKFDXWLRQ:LWKWKH3'$SULRU and taking uncertainty in the calibration into account, the basal split within the sampled Boulengerula QRGHZ RFcurred anytime between 49 and 139 Ma, the divergences of nodes i, r and l, if contemporaneous, occurred between 10 and 20 Ma, and the split between the Taita Hills B. neideni and B. taitanus occurred between 22 and 64 0D:LWKWKHRWKHUSULRUVWKHHVWLPDWHGGLYHUJHQFHVDUH younger. These timings can be interpreted tentatively as HYLGHQFHIRULPSRUWDQWGLYHUVL¿FDWLRQGXULQJWKH0LRFHQH WR0D ZLWKWKHSULPDU\VSOLWZLWKLQBoulengerula Boulengerula±³Afrocaecilia´ SRVVLEO\VWUHWFKLQJEDFN to the Palaeocene. These divergence estimates are sub-

Conservation Boulengerula niedeniLVUDQNHGQXPEHUWKUHHLQWKH('*( RI([LVWHQFHDPSKLELDQFRQVHUYDWLRQSURJUDPPH ZZZ edgeofexistence.org). This species’ status as currently the RQO\,8&1³&ULWLFDOO\(QGDQJHUHG´FDHFLOLDQLVEHFDXVH ³«LWKDVDQH[WHQWRIRFFXUUHQFHRIOHVVWKDQNP2, is restricted to one location, and its habitat is undergoing a FRQWLQXLQJGHFOLQHLQTXDOLW\´ ,8&1 $SSUR[LPDWHO\WZRWKLUGVRIFDHFLOLDQVSHFLHVDUH³'DWD'H¿FLHQW´ IRUWKH,8&15HG/LVWODUJHO\EHFDXVHRILQDGHTXDWH taxonomy and particularly scant data on distribution and HFRORJ\ *RZHU :LONLQVRQ )RUBoulengerula, WKUHHRIWKHVHYHQVSHFLHVDUHFXUUHQWO\'DWD'H¿FLHQWDQG whether B. changamwensis, B. denhardti and%¿VFKHUL TXDOLI\IRU/HDVW&RQFHUQRUDWKUHDWHQHGFDWHJRU\ZLOO probably most rapidly be resolved by greatly improving GDWDRQWKHLUGLVWULEXWLRQ7KHFXUUHQW,8&1DVVHVVPHQWV for B. boulengeri and B. uluguruensis /HDVW&RQFHUQ  LQFOXGHPDSVRISRSXODWLRQV :HVW8VDPEDUDDQG1JXX 1JXUXUHVSHFWLYHO\ WKDWPLJKWUHSUHVHQWXQGHVFULEHG species. As well as reducing the range size of the nominate species, formal description of these potential new 13

S. P. Loader et al .

species would establish taxa with small ranges that might UHTXLUHD³WKUHDWHQHG´FRQVHUYDWLRQDVVHVVPHQW7KLVLV particularly the case for the Malundwe population which, if endemic to the montane forest there, might exist in only DYHU\VPDOO SHUKDSVOHVVWKDQNP2) fragment of habitat. Building upon previous work with further surveying and morphological taxonomic reassessment is going to be crucial for Boulengerula conservation biology as well as underpinning studies of East African caecilian evolution.

relationships of two new species of miniature Arthroleptis $QXUD$UWKUROHSWLGDH IURPWKH(DVWHUQ$UF0RXQWDLQVRI Tanzania. Breviora± %ODFNEXUQ'& 0HDVH\*-  'LVSHUVDOWRRUIURP an African biodiversity hotspot. Molecular Ecology 18, 1904–1915. %RZLH 5&. )MHOGVn - +DFNHWW 6-  &URZH 70 D  0ROHFXODU HYROXWLRQ LQ VSDFH DQG WKRXJK WLPH PW'1$SK\ORJHRJUDSK\RIWKHROLYHVXQELUG Nectarinia olivacea/obscura) throughout continental Africa. Molecular Phylogenetics and Evolution± %RZLH5&.)MHOGVn-+DFNHWW6- &URZH70 E  Systematics and biogeography of double-collared sunbirds from the Eastern Arc Mountains, Tanzania. Auk 121, 660– 681. %RZLH5&.9RHONHU*)MHOGVn-/HQV/+DFNHWW6-  &URZH70  6\VWHPDWLFVRIWKHROLYHWKUXVKTurdus olivaceus species complex with reference to the taxonomic status of the endangered Taita thrush T. helleri. Journal of Avian Biology 36, 391–404. %UHPHU.  7KHOLPLWVRIDPLQRDFLGVHTXHQFHGDWDLQ angiosperm phylogenetic reconstruction. Evolution 42, ± %XUJHVV1'&ODUNH*3 5RGJHUV:$  &RDVWDO IRUHVWVRIHDVWHUQ$IULFDVWDWXVVSHFLHVHQGHPLVPDQGLWV possible causes. Biological Journal of the Linnean Society ± %XUJHVV1'&RUGHLUR1'RJJDUW1)MHOGVn-+RZHOO .0.LODKDPD)/RDGHU63/RYHWW-&0HQHJRQ0 0R\HU'1DVKDQGD(3HUNLQ$6WDQOH\:7 6WXDUW 6  7KHELRORJLFDOLPSRUWDQFHRIWKH(DVWHUQ$UF Mountains of Tanzania and Kenya. Biological Conservation 134, 209–231. &DYDOFDQWL0-  MANTEL, Ver. 1.165LRGH-DQHLUR 'HSWRGH=RRORJLD8QLYHUVLGDGH(VWDGXDOGR5LRGH-DQHLUR $YDLODEOH DW KWWSDFGXIUMEUPDXURELRHQ"$YDLODEOHB Software. &KDQQLQJ$+RZHOO.0  Amphibians of East Africa. ,WKDFD&RUQHOO8QLYHUVLW\3UHVV 'DYHQSRUW75%6WDQOH\:76DUJLV(-'H/XFD': 0SXQJD1(0DFKDJD6- 2OVRQ/(  $QHZ genus of African monkey, Rungwecebus PRUSKRORJ\ ecology, and molecular phylogenetics. Science± 1381. 'UXPPRQG$-  5DPEDXW$   %($67 %D\HVLDQ evolutionary analysis by sampling trees. BMC Evolutionary Biology (PPULFK'  +HUSHWRORJLFDOUHVXOWVRIVRPHH[SHGLWLRQV WRWKH1JXUX0RXQWDLQV7DQ]DQLDMitteilungen aus dem Zoologischen Museum in Berlin± )DLWK'3 &UDQVWRQ36&  &RXOGDFODGRJUDPWKLV VKRUWKDYHDULVHQE\FKDQFHDORQH"2QSHUPXWDWLRQWHVWVIRU cladistic structure. Cladistics± )HOVHQVWHLQ-  &RQ¿GHQFHOLPLWVRQSK\ORJHQLHVDQ approach using the bootstrap. Evolution± )MHOGVn- %RZLH5&.  1HZSHUVSHFWLYHVRQWKH RULJLQDQGGLYHUVL¿FDWLRQRI$IULFD¶VIRUHVWDYLIDXQDAfrican Journal of Ecology± )MHOGVn-%RZLH5&. .LXUH-  7KHIRUHVWBatis, Batis mixtaLVWZRVSHFLHVGHVFULSWLRQRIDQHZQDUURZO\ distributed Batis species in the Eastern Arc biodiversity

ACKNOWLEDGEMENTS :HZRXOGOLNHWRWKDQNWKH7DQ]DQLD&RPPLVVLRQIRU6FLHQFHDQG7HFKQRORJ\ &267(&+UHVHDUFKSHUPLW5&$ 5&$1$5&$  DQG:LOGOLIH'LYLVLRQIRUJUDQWLQJSHUPLVVLRQWRFRQGXFW research in Tanzania and export these specimens. In parWLFXODUZHWKDQN1HER0ZLQD)UHGULFN$PEZHQH/LJDWH -XOLXV.H\\XDQG+01JXOL:HWKDQN'DPDULV5RWLFK DQG3DWULFN0DORQ]DRIWKH1DWLRQDO0XVHXPVRI.HQ\D 0:WKDQNV7DLWD7DYHWDGLVWULFWDQGWKH.HQ\DQ:LOGOLIH Service for issuing collecting permits and their assistDQFH LQ FRQGXFWLQJ WKH ¿HOGZRUN LQ  )LHOG ZRUNFRQGXFWHGLQ.HQ\DE\*-0ZDVFDUULHGRXWXQGHU Kenya’s Ministry of Education, Science and TechnolRJ\UHVHDUFKSHUPLW02(67&7KDQNV WRPDQ\SHRSOH DFNQRZOHGJHGLQSUHYLRXVSDSHUV ZKR DVVLVWHGXVLQWKH¿HOG:HDUHDOVRJUDWHIXOWRPDQ\SHRple and organizations that provided material, support and DGYLFHLQFOXGLQJ)URQWLHU7DQ]DQLD8OXJXUX0RXQWDLQV %LRGLYHUVLW\&RQVHUYDWLRQ3URMHFW(DVW8VDPEDUD&RQVHUYDWLRQ$UHD0DQDJHPHQW3URMHFW/HD&ROOHWW'ZLJKW Lawson and Lucinda Lawson. This work was funded by YDULRXV RUJDQL]DWLRQV LQFOXGLQJ D 1(5& VWXGHQWVKLS 1(56$ WR6/DJUDQWIURPWKH6\VWHPDWLFV$VVRFLDWLRQWR6/D'$3)7JUDQWWR6/0:'* DQG+0D(XURSHDQ&RPPLVVLRQ0DULH&XULH)HOORZVKLS+30)&7WR*-0D'HSDUWPHQWRI =RRORJ\1DWXUDO+LVWRU\0XVHXPVWXGHQWVKLSWR+0 WKH1+00XVHXP5HVHDUFK)XQGWR'*0:DQG6/ DQGWKH3HUF\6ODGHQ0HPRULDO)XQGWR0:DQG*-0

REFERENCES $OGRXV'-  6WRFKDVWLFPRGHOVDQGGHVFULSWLYHVWDWLVWLFV for phylogenetic trees, from Yule to today. Statistical Science 16, 23–34. $UFKLH -:  $ UDQGRPL]DWLRQ WHVW IRU SK\ORJHQHWLF information in systematic data. Systematic Zoology 38, 219–252. %DVLOHZVN\ 3   /HV 2]DHQLQDH G¶$IULTXH HW GH 0DGDJDVFDU &ROHRSWHUD&DUDELGDH Revue de Zoologie et Botanique Africaines 66, 291–314. %DVLOHZVN\3  0LVVLRQHQWRPRORJLTXHGX0XVpH5R\DO GH O¶$IULTXH &HQWUDO DX[ 0RQWV 8OXJXUX 7DQ]DQLD  &ROHRSWHUD &DUDELGDH Revue de Zoologie Africane 90, ± %HUHVIRUG3)MHOGVn- .LXUH-  $QHZVSHFLHVRI DNDODW Sheppardia) narrowly endemic in the Eastern Arc of Tanzania. Auk 12, 23–34. %ODFNEXUQ '   'HVFULSWLRQ DQG SK\ORJHQHWLF

14

B ou l e ng e rul a phylogeny

.XSIHU$:LONLQVRQ0*RZHU'-0OOHU+ -HKOH 5  &DUHDQGSDUHQWDJHLQDVNLQIHHGLQJFDHFLOLDQ amphibian. Journal of Experimental Zoology 309A, 460–  /LQGTYLVW& $OEHUW9$  $KLJKHOHYDWLRQDQFHVWU\ for the Usambara Mountains and lowland populations of $IULFDQ YLROHWV Saintpaulia *HVQHULDFHDH  Systematics and Geography of Plants± /RDGHU 63 *RZHU '- 1JDODVRQ :  0HQHJRQ 0  7KUHHQHZVSHFLHVRICallulina $PSKLELD$QXUD Brevicipitidae) highlight local endemism and conservation plight of Africa’s Eastern Arc forests. Zoological Journal of the Linnean Society 160, 496–514. /RDGHU633LVDQL'&RWWRQ-$*RZHU'-'D\--  :LONLQVRQ0  5HODWLYHWLPHVFDOHVUHYHDOPXOWLSOH origins of parallel disjunct distributions of African caecilian amphibians. Biology Letters 3, 505–508. /RYHWW-&0DUFKDQW57DSOLQ- .SHU:  7KH ROGHVWUDLQIRUHVWVLQ$IULFDVWDELOLW\RUUHVLOLHQFHIRUVXUYLYDO DQGGLYHUVLW\",QPhylogeny and Conservation, 198–229. 3XUYLV$*LWWOHPDQ* %URRNV7 HGV &DPEULGJH &DPEULGJH8QLYHUVLW\3UHVV 0DORQ]D3. 0OOHU+  $UHGLVFRYHU\DIWHUWZR GHFDGHVWKH&KDQJDPZHORZODQGFDHFLOLDQBoulengerula changamwensis/RYHULGJH $PSKLELD*\PQRSKLRQD &DHFLOLLGDH). Journal of East African Natural History 93, ± 0DORQ]D3. 0HDVH\*-  /LIHKLVWRU\RIDQ$IULFDQ FDHFLOLDQBoulengerula taitanus &DHFLOLLGDH$PSKLELD *\PQRSKLRQD Tropical Zoology 18, 49–66. 0DULDX[-/XW]PDQQ1 6WLSDOD-  7KHWZRKRUQHG chameleons of East Africa. Zoological Journal of the Linnean Society± 0DWWKHH &$ 7LOEXU\ &5  7RZQVHQG 7$   Phylogenetic review of the African leaf chameleons genus Rhampholeon &KDPDHOHRQLGDH  WKH UROH RI YLFDULDQFH and climate change in speciation. Proceedings of the Royal Society London B± 0HDVH\*-  6XUYH\LQJELRGLYHUVLW\RIVRLOKHUSHWRIDXQD towards a standard quantitative methodology. European Journal of Soil Biology 42, S103–S110. 0HQHJRQ 0 'RJJDUW 1  2ZHQ 1  7KH 1JXUX Mountains of Tanzania, an outstanding hotspot of herpetofaunal diversity. Acta Herpetologica± 0|OOHU0 &URQN4&%  2ULJLQDQGUHODWLRQVKLSV of Saintpaulia *HVQHULDFHDH  EDVHG RQ ULERVRPDO '1$ LQWHUQDO WUDQVFULEHG VSDFHU ,76  VHTXHQFHV American Journal of Botany 84, 956–965. 0RRHUV$‘  +HDUG 6%   ,QIHUULQJ HYROXWLRQDU\ processes from phylogenetic tree shape. Quarterly Journal of Biology± 0OOHU+0HDVH\*-/RDGHU63 0DORQ]D3.   A new species of Boulengerula 7RUQLHU $PSKLELD *\PQRSKLRQD &DHFLOLLGDH  IURP DQ LVRODWHG PRXQWDLQ block of the Taita Hills, Kenya. Zootaxa± 0\HUV10LWWHUPHLHU5$0LWWHUPHLHU&*'D)RQVHFD *$%  .HQW -   %LRGLYHUVLW\ KRWVSRWV IRU conservation priorities. Nature 403, 853–858. 1XVVEDXP5$ +LQNHO+  5HYLVLRQRI(DVW$IULFDQ caecilians of the genera Afrocaecilia Taylor and Boulengerula 7RUQLHU $PSKLELD *\PQRSKLRQD &DHFLOLLGDH  Copeia

hotspot. Journal of Ornithology± )URVW '5 *UDQW 7 )DLYRYLFK - %DLQ 5+ +DDV$ +DGGDG&)%GH6D52&KDQQLQJ$:LONLQVRQ0 'RQQHOODQ6&5D[ZRUWK\&-&DPSEHOO-$%ORWWR %0ROHU3'UHZHV5&1XVVEDXP5$/\QFK- *UHHQ'0 :KHHOHU:&  7KHDPSKLEDQWUHH of life. Bulletin of the American Museum of Natural History ± *DERULHDX2 0HDVH\*-  7HUPLWLYRUHRUGHWULWLYRUH" A quantitative investigation into the diet of the East African caecilian Boulengerula taitanus $PSKLELD*\PQRSKLRQD &DHFLOLLGDH Animal Biology 54, 45–56. *RZHU '- 'KDUQH 0 %KDWWD * *LUL 9 9\DV 5 *RYLQGDSSD92RPPHQ29*HRUJH-6FKRXFKH<  :LONLQVRQ0  5HPDUNDEOHJHQHWLFKRPRJHQHLW\LQ unstriped, long-tailed Ichthyophis $PSKLELD*\PQRSKLRQD ,FKWK\RSKLLGDH DORQJNPRIWKH:HVWHUQ*KDWV,QGLD Journal of Zoology± *RZHU '- .XSIHU $ 2RPPHQ 29 +LPVWHGW : 1XVVEDXP 5$ /RDGHU 63 3UHVVZHOO % 0OOHU +.ULVKQD6%%RLVWHO5 :LONLQVRQ0  $ preliminary molecular phylogeny of ichthyophiid caecilians $PSKLELD*\PQRSKLRQD,FKWK\RSKLLGDH RXWRI,QGLDRU RXWRIVRXWKHDVW$VLD"Proceedings of the Royal Society of London B 269, 1563–1569. *RZHU'-/RDGHU63:LONLQVRQ0 0RQFULHII&%   1LFKH VHSDUDWLRQ DQG FRPSDUDWLYH DEXQGDQFH RI Boulengerula boulengeri and Scolecomorphus vittatus $PSKLELD *\PQRSKLRQD  LQ (DVW 8VDPEDUD IRUHVW Tanzania. African Journal of Herpetology 53, 183–190. *RZHU'- :LONLQVRQ0  &RQVHUYDWLRQELRORJ\RI caecilian amphibians. Conservation Biology 19, 45–55. *UDYOXQG3  0ROHFXODUSK\ORJHQ\RI7RUQLHU¶VFDWVQDNH Crotaphopeltis tornieri), endemic to East African mountain IRUHVWVELRJHRJUDSK\YLFDULDQFHHYHQWVDQGSUREOHPDWLF species boundaries. Journal of Zoological Systematics and Evolutionary Research 40, 46–56. +XHOVHQEHFN-3 5RQTXLVW)  0U%D\HV%D\HVLDQ inference of phylogenetic trees. Bioinformatics±  ,8&1&RQVHUYDWLRQ,QWHUQDWLRQDO 1DWXUHVHUYH  Global Amphibian Assessment. $YDLODEOHDWKWWSZZZLXFQUHGOLVW RUJDPSKLELDQV DFFHVVHG6HSWHPEHU  -HW]:5DKEHN& &ROZHOO5.  7KHFRLQFLGHQFH of rarity and richness and the potential signature of history in centres of endemism. Ecological Letters± -RKDQVRQ.$ :LOOLDVVHQ(  $UH$IULFDQVSHFLHV of HelicopsycheYRQ6LHEROG ,QVHFWD7ULFKRSWHUD +HOLFRSV\FKLGDH  PRQRSK\OHWLF" Tropical Zoology 10, ± .DKLQGR&%RZLH5&. %DWHV-0  7KHUHOHYDQFH of data on genetic diversity for the conservation of Afromontane regions. Biological Conservation± .LVKLQR+ +DVHJDZD0  (YDOXDWLRQRIWKHPD[LPXP likelihood estimate of the evolutionary tree topologies from '1$VHTXHQFHGDWDDQGWKHEUDQFKLQJRUGHULQ+RPLQRLGHD. Journal of Molecular Evolution ± .XSIHU$0OOHU+$QWRQLD]]L00-DUHG&*UHYHQ + 1XVVEDXP 5$  :LONLQVRQ 0   3DUHQWDO investment by skin feeding in a caecilian amphibian. Nature 440, 926–929.

15

S. P. Loader et al .

± 3RVDGD' &UDQGDOO.  0RGHOWHVWWHVWLQJWKHPRGHO RI'1$VXEVWLWXWLRQBioinformatics± 5DPEDXW$ 'UXPPRQG$-  Tracer: MCMC Trace File Analyzer, Version 1.1.12[IRUG8QLYHUVLW\RI2[IRUG $YDLODEOHIURPKWWSHYROYH]RR[DFXN 5RHODQWV.*RZHU'-:LONLQVRQ0/RDGHU63%LMX 6' *XLOODXPH .  %RVVX\W )   3DWWHUQV RI diversification in the history of modern amphibians. Proceedings of the National Academy of Sciences of the United States of America± 5R\06  5HFHQWGLYHUVL¿FDWLRQRI$IULFDQJUHHQEXOV 3\FQRQRWLGDHAndropadus) supports a montane speciation model. Proceedings of the Royal Society London B 264, ± 6FKDUII 1   7KH OLQ\SKLLG IDXQD RI HDVWHUQ $IULFD $UDQHDH /LQ\SKLLGDH  ± GLVWULEXWLRQ SDWWHUQV GLYHUVLW\ and endemism. Biological Journal of the Linnean Society ± 6WDPDWDNLV$  5$[0/9,+3&0D[LPXPOLNHOLKRRG based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–2690. 6WDQOH\:7 2OVRQ/(  3K\ORJHQ\SK\ORJHRJUDSK\ and geographic variation of Sylvisorex howelli 6RULFLGDH  an endemic shrew of the Eastern Arc Mountains, Tanzania. Journal of Zoology 266, 341–354. 6ZRIIRUG'/  PAUP*: Phylogenetic Analysis Using Parsimony (and Other Methods). Test Version 4b6. 6XQGHUODQG0DVVDFKXVHWWV6LQDXHU$VVRFLDWHV 7D\ORU(+  The Caecilians of the World. A Taxonomic Review/DZUHQFH.DQVDV8QLYHUVLW\RI.DQVDV3UHVV 7HPSOHWRQ$5  3K\ORJHQHWLFLQIHUHQFHIURPUHVWULFWLRQ endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution  221–244. 7LOEXU\ & 7ROOH\ .  %UDQFK %   $ UHYLHZ RI the systematics of the genus Bradypodion 6DXULD &KDPDHOHRQLGDH ZLWKWKHGHVFULSWLRQRIWZRQHZJHQHUD Zootaxa 1363, 23–38. 9DQ%RF[ODHU,%LMX6'/RDGHU63 %RVVX\W)   Toad radiation reveals into-India dispersal as a source of HQGHPLVP LQ WKH :HVWHUQ *KDWV±6UL /DQND ELRGLYHUVLW\ hotspot. BMC Evolutionary Biology 9, 131. 9DQ%RF[ODHU,5RHODQWV.%LMX6'1DJDUDMX- %RVVX\W )   /DWH &UHWDFHRXV YLFDULDQFH LQ *RQGZDQDQ amphibians. PLoS ONEH 9DQGHQVSLHJHO'  7DLWDVWUHSWXVÀDYLSHV, a new genus and new species for an arboreal millipede from Kenya 'LSRSRGD 6SLURVWUHSWLGDH  Insect Systematics and Evolution±

9HVWHUJDDUG0  An Annotated and Illustrated Checklist of the Amphibians of the Usambara Mountains, with a Tentative Key and the Description of Two New Taxa.&DQG VFLHQWWKHVLV&RSHQKDJHQ=RRORJLFDO0XVHXP8QLYHUVLW\ RI&RSHQKDJHQ :DUXL &  -RFTXp 5   7KH ILUVW *DOOLHQLHOOLGDH $UDQDHDH IURP(DVWHUQ$IULFDJournal of Arachnology ± :LONLQVRQ0/RDGHU63*RZHU'-6KHSV-$ &RKHQ %/  3K\ORJHQHWLFUHODWLRQVKLSVRI$IULFDQFDHFLOLDQV $PSKLELD*\PQRSKLRQD LQVLJKWVIURPPLWRFKRQGULDO U'1$JHQHVHTXHQFHVAfrican Journal of Herpetology 52, 83–92. :LONLQVRQ 0 /RDGHU 63 0OOHU +  *RZHU '-  7D[RQRPLFVWDWXVDQGSK\ORJHQHWLFUHODWLRQVKLSV of Boulengerula denhardti 1LHGHQ  $PSKLELD *\PQRSKLRQD&DHFLOLLGDH . Mitteilungen aus dem Museum für Naturkunde Berlin, Zoologische Reihe 80, 41–51. :LONLQVRQ0 1XVVEDXP5$  &DHFLOLDQSK\ORJHQ\ DQGFODVVL¿FDWLRQ,QReproductive Biology and Phylogeny of Amphibia, Volume 3, Gymnophiona±-0([EUD\DW HG (Q¿HOG6FLHQFH3XEOLVKHUV,QF :ROOHQEHUJ .&  0HDVH\ *-   :K\ FRORXU LQ VXEWHUUDQHDQYHUWHEUDWHV"([SORULQJWKHHYROXWLRQRIFRORXU patterns in caecilian amphibians. Journal of Evolutionary Biology 22, 1046–1056.
Accepted: 28 August 2010

NOTE ADDED IN PROOF *RZHUHWDO  UHSRUWQHZPDWHULDORIBoulengerula ¿VFKHUL and add it to the molecular phylogenetic data set analysed here. *RZHU'-3DSDGRSRXORX$'RKHUW\%RQH703XSLQ )6DQ0DXUR'/RDGHU63 :LONLQVRQ0   The systematics of Boulengerula fischeri $PSKLELD *\PQRSKLRQD&DHFLOLLGDH EDVHGRQPRUSKRORJLFDODQG molecular data. Zootaxa

16

Molecular phylogenetics of Boulengerula

Simon P. Loader1,2, Mark Wilkinson2, James A. Cotton3,4, G. John ... Our analyses identified genetic differences between several mtDNA clades that .... chain marked as dark areas. ..... Judged by the KH and Templeton tests, the best trees.

1MB Sizes 0 Downloads 199 Views

Recommend Documents

Molecular Phylogenetics and Evolution 38
plete mitochondrial genome sequences of 10 species and a »14 kb sequence from an eleventh species ...... try of Nature Protection of Turkmenistan for hosting.

Molecular phylogenetics of the spider infraorder ... - CiteSeerX
logenetic markers (e.g., nuclear protein-coding genes), and the integration of both ..... manual.php/> (last accessed 20.01.06). Sanmartın, I., Ronquist, F., 2004.

Molecular phylogenetics of the spider infraorder ... - CiteSeerX
The following primary results are supported by both Bayesian and parsimony analyses of combined matrices representing ..... and 28S data partitions, the computer program MrModel- .... of 10–365 likelihood units, as compared to constrained.

Molecular Phylogenetics and Evolution 38
A rank-free hierarchical representation of the taxonomy proposed here is shown in ...... Sarich, V.M., Wilson, A.C., 1973. Generation time and genomic evolu-.

Molecular Phylogenetics and the Diversification of ... - Semantic Scholar
Apr 3, 2014 - as a whole does not exhibit a significant signature of density-dependent diversification ..... Digital Distribution Maps of the Birds of the Western ...

Molecular phylogenetics of the spider infraorder ...
sampling and the inclusion of additional character systems (more genes and morphology) are ..... and 28S data partitions, the computer program MrModel- test v.

Molecular phylogenetics of the hummingbird genus ...
spite a good knowledge of higher-level relationships in the family ... Ecology and Evolution, State University of New York, Stony Brook, NY 11794-5245, ..... University of California, Los Angeles; MVZ: Museum of Vertebrate Zoology, University ...

Phylogenetics
Denoting the five species in the original distance matrix by species 1, 2,_ and 5, and using the dij values in the distance matrix, we have Q12=22261. In a similar manner, we ob- tain the other Q values under step 1 in Table 1. Since Q45 is the small

Phylogenetics Theory and Practice of Phylogenetic Systematics.pdf ...
Retrying... Phylogenetics Theory and Practice of Phylogenetic Systematics.pdf. Phylogenetics Theory and Practice of Phylogenetic Systematics.pdf. Open.

Introduction to phylogenetics using - GitHub
Oct 6, 2016 - 2.2 Building trees . ... Limitations: no model comparison (can't test for the 'best' tree, or the 'best' model of evolution); may be .... more efficient data reduction can be achieved using the bit-level coding of polymorphic sites ....

The Troubled Growth of Statistical Phylogenetics
1971, at that meeting in Ann Arbor, Gareth ... onomy Conference meetings, which were among the only .... This was reflected in available software such as PAUP ...

Phylogenetics of the southern African dwarf ...
De- spite the reported phenotypic diversity in this region corresponding to described species (Raw, 1995, 2001), ... Even with a lack of duplicate samples.

Phoronid phylogenetics (Brachiopoda ... - Wiley Online Library
1Smithsonian Environmental Research Center, PO Box 28, 647 Contees Wharf Road, Edgwater, MD ... 24 taxa, phoronid rDNAs were combined with data from brachiopods and distant (molluscan) outgroups. ... E-mail: [email protected].

Introduction to phylogenetics using
Feb 22, 2016 - Three main classes of phylogenetic approaches are introduced, namely distance-based, maximum parsimony, and maximum likelihood methods. We also illustrate how to assess .... and adegenet [2] are here used essentially for their graphics

Optimal Design of a Molecular Recognizer: Molecular Recognition as ...
information channels and especially of molecular codes [4], [5]. The task of the molecular .... Besides the questions regarding the structural mis- match between the ...... Institute of Technology, Haifa, Israel, and the M.Sc. degree in physics from 

Molecular
Mar 21, 2009 - The simulation results indi- cate that for large molecules with small diffusion coefficients (e.g. .... Institute for Science and Education. References.

The Posterior and the Prior in Bayesian Phylogenetics
... [email protected]. 2School of Computational Science, Florida State University, Tallahassee, ... First published online as a Review in Advance on July 21, 2006.

The Posterior and the Prior in Bayesian Phylogenetics
the statistical behavior of the PP, especially in relation to the BP (Alfaro et al. .... clade will be found in the full data analysis and gives systematists a sense of ...

The Posterior and the Prior in Bayesian Phylogenetics
... University, Pullman,. Washington 99164-4236; email: [email protected] ...... An ad hoc approach could also be used to mimic the be- havior of a conservative ...

Molecular mechanism of floral .pdf
... marker push Qpin 1. Icon Archive Website: http://www.iconarchive.com/show/vista-map-markers-icons-by-icons-land/Map-Marker-Push-Pin-1-Right-Pink-icon.