The systematics of Boulengerula fischeri (Amphibia ...

Zootaxa 2767: 14–24 (2011) www.mapress.com / zootaxa/ Copyright © 2011 · Magnolia Press

ISSN 1175-5326 (print edition)

Article

ZOOTAXA ISSN 1175-5334 (online edition)

The systematics of Boulengerula fischeri (Amphibia: Gymnophiona: Caeciliidae) based on morphological and molecular data DAVID J. GOWER1,4, ANNA PAPADOPOULOU1, THOMAS M. DOHERTY-BONE1, FABIO PUPIN2, DIEGO SAN MAURO1, SIMON P. LOADER3 & MARK WILKINSON1 1

Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK Museo Tridentino di Scienze Naturali, Via Calepina 14, I-38122 Trento, Italy 3 University of Basel, Institute of Biogeography, Department of Environmental Sciences, Basel 4056, Switzerland 4 Corresponding author. E-mail: [email protected] 2

Abstract Eight new specimens are reported for the caecilian amphibian Boulengerula fischeri, a species known in detail previously only from its holotype collected in 1987. The new material was collected at the type locality in Rwanda in 2009, and is used to expand and refine the morphological characterization of the species. Mitochondrial and nuclear gene sequences were used to assess the phylogenetic position of B. fischeri in the context of a recent molecular phylogeny of six of the seven other species of Boulengerula (from Kenya and Tanzania). Among nominal species, only B. denhardti remains to be included in molecular phylogenetic studies of Boulengerula. Boulengerula is recovered as monophyletic, with either B. fischeri or (more probably) B. boulengeri sister to all other sampled species. There are at least three deep lineages within Boulengerula: (1) B. boulengeri, (2) B. fischeri, and (3) all other Eastern Arc Mountain and Coastal Forest species from Kenya and Tanzania. The status of Afrocaecilia, a genus erected by Taylor in 1968 to contain all Boulengerula except B. boulengeri, is not yet resolved. Key words: Africa, Afrocaecilia, Albertine Rift, caecilian, DNA, Eastern Arc Mountains, phylogenetics, Rwanda

Introduction Boulengerula fischeri Nussbaum & Hinkel in Fischer & Hinkel, 1992 is one of seven nominate species of this East African genus of caecilian amphibian (Gymnophiona). This species was described on the basis of a single specimen from Rwanda collected in 1987 (Nussbaum & Hinkel, 1994). One additional specimen from an unspecified locality in the Albertine Rift region was reported by Behangana et al. (2009), Herrel & Measey (2010) presented locomotory kinematic data for a single field-collected specimen, and a photograph of a specimen appears on the CalPhotos website (http://calphotos.berkeley.edu). Taylor (1968) partitioned Boulengerula Tornier, transferring all but the type species (B. boulengeri) to his new genus Afrocaecilia. Following Nussbaum & Hinkel (1994), the latter genus is currently regarded as a synonym of Boulengerula (e.g., Wilkinson & Nussbaum, 2006), but monophyly of the species transferred by Taylor (1968) to Afrocaecilia (Loader et al., in press) has placed this in some doubt. Boulengerula fischeri was not known to Taylor, and it was one of only two nominal species (the other being B. denhardti, not considered valid by Taylor) not sampled in Loader et al.’s (in press) molecular phylogeny of Boulengerula. Herrel & Measey (2010: fig 2) depicted B. fischeri as sister to B. taitanus Loveridge (the only other Boulengerula in their tree) but cite only Roelants et al. (2007) as a source though the latter study did not include B. fischeri. During herpetofaunal surveys in Rwanda in 2009, new specimens of B. fischeri were collected that have enabled a reassessment of the systematics of this poorly known species on the basis of both morphological and DNA sequence data.

14

Accepted by S. Carranza: 11 Jan. 2011; published: 17 Feb. 2011

Material and methods Fieldwork. The field study was conducted in October to November 2009 in southwestern Rwanda (Fig. 1), in Cyamudongo Forest and in Nyungwe Forest National Park. Cyamudongo is the type locality of B. fischeri (Nussbaum & Hinkel, 1994), and Cyamudongo Forest is a small (diameter no greater than 3.3 km) patch of montane primary rain forest (including some secondary vegetation) approximately 15 km East of the city of Bukavu (Democratic Republic of Congo). Cyamudongo Forest is a Reserve that was previously connected to the much larger Nyungwe Forest National Park (ca. 970 km2 of mostly montane forest and high altitude woodland), which now lies approximately 8 km to the West. Both Cyamudongo Forest Reserve and Nyungwe Forest National Park are surrounded by deforested habitat, predominantly farmland (including a tea plantation) and small settlements. Herpetofaunal surveys carried out that were relevant to finding caecilians included visual encounter surveys (VES), pitfall trapping (with drift fences), and dedicated digging of soil. VES included rolling logs/stones and searching through leaf litter. Digging was carried out with bladed hoes to depths of up to ca. 30 cm. Pitfall traps each comprised 11 eight litre plastic buckets spaced at 5 m intervals along a ca. 0.5 m high, 50 m long plastic drift fence, and these were checked every morning. The bottom of each bucket was pierced by drainage holes up to ca. 5 mm in diameter. All three methods were employed at Cyamudongo, but substantial digging was not carried out in Nyungwe. Collected caecilian specimens were fixed in ca. 95% ethanol and stored in 70% ethanol in the collections of the Museo Tridentino di Scienze Naturali, Trento, Italy (MTSN) and the Natural History Museum, London, UK (BMNH numbers). Tissue samples (liver) were taken from four specimens prior to fixation and preserved in absolute ethanol.

FIGURE 1. Map showing position of six sampling localities in Cyamudongo and Nyungwe Forests, Rwanda.

SYSTEMATICS OF BOULENGERULA FISCHERI

Zootaxa 2767 © 2011 Magnolia Press ·

15

Morphological systematics. Numbers of vertebrae were determined by radiography. Sex was determined by examination of gonads. Skull characters were assessed by dissection. Total length and circumference were measured to the nearest 1 mm using a ruler, and a ruler plus piece of thread, respectively. All other measurements were taken to the nearest 0.1 mm with dial calipers under a binocular dissection microscope. Molecular systematics. Genomic DNA was extracted from B. fischeri liver samples using commercial kits (Qiagen). Three partial mitochondrial (mt) genes were amplified and sequenced for all four specimens: 12S, 16S and cytochrome b (cytb). Cytochrome oxidase subunit 1 (cox1) and the nuclear gene RAG1 were amplified and sequenced for a single B. fischeri sample plus one each of the nominal species (B. boulengeri Tornier, B. changamwensis Loveridge, B. taitanus Loveridge, and B. uluguruensis Barbour & Loveridge) and some possibly undescribed species of Boulengerula included in Loader et al.’s (in press) analysis. The 12S and 16S primers were as reported by Loader et al. (in press); the cytb and cox1 primers were following San Mauro et al. (2004). Details of voucher specimens used for sequencing and GenBank accession numbers are given in Table 1. TABLE 1. Details of voucher specimens and sequences used in molecular analyses. All sequences are new to this study except those underlined, from Loader et al. (in press), San Mauro et al. (2004), Zhang & Wake (2009). Voucher prefixes indicate specimens in the following collections: BMNH, The Natural History Museum, London (UK); CAS, California Academy of Sciences, San Francisco (USA); DNM, Department of National Museums, Colombo (Sri Lanka); NMK, National Museums of Kenya, Nairobi (Kenya); MW field tag for specimens to be deposited in BMNH. aSpecimen used for RAG1 sequence; mt sequences without voucher. Taxon Ichthyophis glutinosus Scolecomorphus vittatus Typhlonectes natans Herpele squalostoma Boulengerula fischeri Boulengerula fischeri Boulengerula fischeri Boulengerula fischeri Boulengerula boulengeri

voucher

12S

16S

cytb

cox1

RAG1

AY456251

AY456251

AY456251

AY456251

AY456256

AY456253

AY456253

AY456253

AY456253

AY456258

BMNH 2000.218

AY154051

AY154051

AY154051

AY154051

AY456260

BMNH 2002.97

FR691652

FR691657

FR691662

FR691667

FR691675

DNM MW1733 BMNH 2002.100 a

BMNH 2008.604

FR691653

FR691658

FR691663

FR691668

FR691676

MTSN 7141

FR691654

FR691659

FR691664

-

-

MTSN 7225

FR691655

FR691660

FR691665

-

-

BMNH 2008.607

FR691656

FR691661

FR691666

-

-

BMNH 2002.95; CAS 168822

AY450613

AY450620

EU200987

GQ244464

FR691677

Boulengerula cf. boulengeri

BMNH 2005.1349

FN652682

FN652714

FN652746

FR691669

FR691678

Boulengerula taitanus

NMK A/3112

AY954504

AY954504

AY954504

AY954504

DQ320062

Boulengerula niedeni

NMK A/4294

FN652691

FN652723

FN652755

FR691670

FR691679

Boulengerula changamwensis

NMK A/4129

FN652690

FN652722

FN652754

FR691671

FR691680

Boulengerula uluguruensis

BMNH 2005.187

FN652684

FN652716

FN652748

FR691672

FR691681

Boulengerula cf. uluguruensis (Nguru)

BMNH 2002.928

FN652675

FN652707

FN652739

FR691673

FR691682

Boulengerula cf. uluguruensis (Nguu)

MW 6638

FN652696

FN652728

FN652760

FR691674

FR691683

Phylogenetic analyses were conducted using the Boulengerula samples sequenced for both mt and nuclear genes plus four outgroup species of caecilian selected as members of proximal lineages on the basis of previous phylogenetic results (Frost et al., 2006; Roelants et al., 2007; San Mauro et al., 2009; Wilkinson et al., 2003; Zhang & Wake, 2009): Herpele squalostoma (Stutchbury), Scolecomorphus vittatus (Boulenger), Typhlonectes natans (Fischer), and Ichthyophis glutinosus (Linnaeus). Three different sets of these outgroups were employed in analyses: all four taxa, all taxa except H. squalostoma (the putative sister group to Boulengerula, see Loader et al., in press), and only H. squalostoma. The 12S and 16S sequences were aligned using ClustalX version 2.0.12 (Thompson et al., 1997; Larkin et al., 2007), and alignment ambiguities and gaps filtered out using Gblocks v. 0.91b (Castresana, 2000), with default parameters in both cases; all other sequences were aligned by eye and checked that they could be translated into unbroken amino acid sequences. Maximum parsimony (MP) analyses of the concatenated dataset were performed in PAUP* 4.0b10 (Swofford, 2002), using heuristic searches, 100 random stepwise addition sequence replicates and TBR branch swapping. Maximum likelihood (ML) searches with 20 replicates were conducted in RAxML 7.2.6 (Stamatakis, 2006) under two different partitioning schemes (data divided either by gene or by gene and codon position), and a separate GTR+Γ model was applied to each partition (following the software’s recommen-

16 · Zootaxa 2767 © 2011 Magnolia Press

GOWER ET AL.

dations and limitations). For both MP and ML analyses, clade support was evaluated using non-parametric bootstrapping with 1000 pseudoreplicates. Bayesian inference (BI) was performed with MrBayes 3.1.2 (Ronquist & Huelsenbeck, 2003) under the same partitioning schemes, applying a separate model to each partition as selected by AIC (or AICc to correct for small sample sizes) in jModeltest (Posada, 2008). The best-fit models were: GTR+Γ+I (cox1), GTR+Γ (12S, 16S, RAG1, 1st codon positions of cox1, 3rd positions of cox1, 3rd positions of cytb, cox1 without 3rd positions, cytb without 3rd positions), SYM (cytb), HKY+ Γ+I (2nd positions of cytb), HKY+ Γ (1st positions of cytb, 1st positions of RAG1), HKY+ I (2nd and 3rd positions of RAG1), and F81+I (2nd positions of cox1). Two independent runs were performed of 10 million generations each, using default parameters for the MCMC and sampling every 1000 generations. Posterior probabilities were calculated after checking convergence diagnostics and discarding the first one million generations (10%) as burnin. All analyses were repeated after removing the third codon positions of cox1 and cytb. Saturation of third codon positions was assessed by visual inspection of transition-transversion plots.

Results Fieldwork. Boulengerula fischeri were encountered at two localities, in mixed farmland (mostly smallholdings of banana and maize) on the edge of Cyamudongo Forest (2˚ 32.288’ S, 28˚ 59.631’ E, 1839 m elevation) and within Cyamudongo Forest (between 2˚ 33.371’ S, 28˚ 59.326 E, 1850 m, and 2˚ 33.382’ S, 28˚ 59.327’ E, 1743 m). Ten specimens were collected by an unknown number of local people digging in soil outside the edge of Cyamudongo Forest, of which six specimens were preserved. Three specimens (two preserved) were collected during VES in the forest, from leaf litter covering a compacted soil path (Fig. 2). It is not known how much digging local people conducted to collect the 10 specimens from outside the edge of the forest, but they reported that they were fairly common. A total of 7.95 (5.2 at Cyamudongo) person hours of digging and 191.1 (99 at Cyamudongo) person hours of VES were undertaken in forest. No caecilians were found during 440 bucket days of pitfall trapping at six sites at altitudes of 1737–2919 m within forest at Cyamudongo and Nyungwe. Details of the search effort and numbers of B. fischeri found at each site are given in Table 2. TABLE 2. Field effort and number of Boulengerula fischeri found (in parentheses) at sites in Cyamudongo (Cyamudongo 1, 2) and Nyungwe (Uwinka, Karamba, Bigugu, Nshili). Location of sites shown in Fig. 1. Site

Altitude (m) Habitat

Dates (2009)

VES hours

Bucket days

Digging hours

1

Cyamudongo 1

1839

farmland

4 Nov

0

0

? (10)

2

Cyamudongo 2

1737–1854

montane forest

13–17 Oct, 2–7 Nov

66.8 (3)

99

5.2

3

Uwinka

2260–2264

montane forest

18 Oct–6 Nov

0

220

0

4

Karamba

1823–1830

montane forest

19–22 Oct

42.2

33

0

5

Bigugu

2899–2919

high altitude wood- 23–27 Oct land

36

44

2.75

6

Nshili

2185–2199

montane forest

46.1

55

0

Total 191.1

Total 440

Total 7.95

28 Oct–1 Nov

Morphological systematics. Some morphometric and meristic data are presented in Table 3. Non-destructive tooth counts for whole wet caecilian specimens are usually estimates (e.g., Gower et al., 2010) that can vary among different researchers (e.g., Wilkinson et al., 2009). With the new material studied here, there were sometimes doubts over whether small teeth at the posterior of each row might have been missed, so that counts should be checked in the future in B. fischeri specimens used to make dry skeletal or cleared and stained preparations, or subjected to x-ray computed tomography. In agreement with Nussbaum & Hinkel (1994), we observed bicusped teeth in the inner rows and only monocusped teeth in the outer rows, though the cusp details of the smallest elements of each row could not be readily examined. None of the new specimens is especially well preserved, all having undergone some dehydration. For the most part, the new specimens closely match the morphology of the holotype reported by Nussbaum & Hinkel (1994).



17

Nussbaum & Hinkel (1994) reported a short, faint dorsal transverse groove on the first collar that we interpret as a probable superficial crease rather than a ‘true’ transverse groove on the basis that all the new specimens (except MTSN 7225, in which a faint superficial crease is present) lack transverse grooves or creases. Whereas Nussbaum & Hinkel (1994) reported that all the annular grooves were complete except for the last, which was ventrally incomplete, we observed many narrowly incomplete (middorsally and midventrally) grooves in the new material, and the last (posteriormost) groove lies clearly anterior to the vent so that the terminal shield is notably longer than in the holotype (Table 3). All of the new specimens are substantially longer than the holotype (271–329 vs 191 mm), and the range of primary annuli and vertebrae for the species is increased from 186 and 194 to 183–197 and 193–204, respectively. There is no obvious indication of sexual dimorphism in relative head/jaw length or width, or numbers of annuli (or vertebrae) in our small sample, but all five males known for the species are longer than all four females (Table 3). Herrel & Measey (2010: fig. 2) did not present collection or detailed specimen data but reported that the length/width of their single (173 mm long) specimen was 97, which is slightly more attenuate than the 81–95 we recorded in the eight preserved specimens we examined (Table 3). BMNH 2008.604 was skinned, revealing a lack of dorsal exposure of the mesethmoid, as described for the holotype by Nussbaum & Hinkel.

FIGURE 2. Habitat along path within Cyamudongo Forest Reserve where Boulengerula fischeri were found during visual encounter surveys of the leaf litter.

Nussbaum & Hinkel (1994) described B. fischeri as “pinkish in life with the head and neck and the posterior few centimeters of the body a brighter pink than the midbody, light yellowish-tan in preservative”. The new specimens agree generally with this. In life, the new specimens were vivid pink anteriorly with a paler pink head, and becoming more lavender posteriorly so that the last few centimeters are pale lavender and eventually very pale grey-pink (Fig. 3). In addition, there is a prominent but narrow darker middorsal stripe clearest in the mid to posterior part of the body and only weakly indicated or absent both on the anterior part of the body and posteriorly onto the terminal shield. This darker middorsal stripe is also apparent in the preserved material. The photograph presented by Fischer & Hinkel (1992: fig. 99) is perhaps of an already dead specimen because it is dull reddish pink in colour versus the bright pale pink-lavender we observed. Molecular systematics. There is almost no variation in the three mt genes sequenced for four specimens of B. fischeri, only a single variable site for cytb. Visual inspection of transition-transversion plots showed cytb and cox1


GOWER ET AL.

third codon positions to be saturated, even when all outgroups except Herpele were excluded. Details of the support for various clades recovered in different phylogenetic analyses are presented in Table 4, and ML trees for analyses including one and four outgroups are shown in Fig. 4. All our phylogenetic analyses recovered a monophyletic Boulengerula and Boulengerula + Herpele with maximal support (except for parsimony analysis with four outgroups). Within Boulengerula, a clade comprising B. taitanus, B. niedeni, B. changamwensis and B. uluguruensis was recovered with maximal support in all analyses, and relationships within this clade were always exactly as recovered by Loader et al. (in press), with all internal nodes strongly supported (> 90 %). Different partitioning schemes had very little effect on phylogenetic results. Partitioning by gene and codon position was favoured under both Akaike and Bayesian information criteria.

TABLE 3. Morphometric (mm) and meristic data for the holotype (UMMZ – University of Michigan, Museum of Zoology) and eight new specimens of Boulengerula fischeri. Holotype data from Nussbaum & Hinkel (1994). Head, collar and terminal shield lengths measured laterally. Multiple values for tooth counts (separated by a /) are variation encountered by two observers (DJG, MW). TG = dorsal transverse groove. UMMZ BMNH MTSN BMNH 196240 2008.605 7222 2008.607

MTSN BMNH MTSN 7141 2008.604 7223

BMNH MTSN 2008.606 7225

Sex

f

f

f

f

m

m

m

m

m

Total length (TL)

191

292

276

271

301

329

320

315

301

Vertebrae

194

200

200

202

193

193

199

204

203

Primary annuli

186

192

193

197

186

183

191

196

194

Midbody width (MW)

3.2

3.4

3.0

3.5

3.6

3.6

3.3

3.2

Midbody circumference

12

12

12

12

13

13

12

12

TL/MW

91.3

81.2

90.3

86.0

91.4

88.9

95.5

94.1

Head length to corner of mouth (CM)

3.7

3.8

3.7

3.9

4.1

4.0

3.8

3.7

Head length to first collar (HL)

4.8

4.6

4.8

4.9

5.4

4.9

4.8

4.9

HL as % of TL

1.64

1.67

1.77

1.63

1.64

1.53

1.52

1.63

Lower jaw length to CM

2.9

2.7

2.9

3.0

3.2

3.0

2.8

3.0

Head width between CM and first collar

3.0

2.9

2.9

3.0

3.2

3.0

3.0

3.0

HL/HW

1.60

1.59

1.66

1.63

1.69

1.63

1.60

1.63

Length of first collar

1.6

1.5

1.5

1.7

1.6

1.6

1.7

1.5

Length of second collar

1.8

1.7

1.7

1.7

2.0

1.9

1.7

2.0

TGs on first collar

1

0

0

0

0

0

0

0

0

TGs on second collar

1

1

1

1

1

1

1

1

1

1.0

1.0

1.0

1.0

1.2

1.1

1.0

1.1

1.6

1.7

1.7

1.7

1.8

1.8

1.8

1.7

Distance between tentacles

2.2

2.1

2.3

2.4

2.5

2.3

2.3

2.3

Distance between nostrils

1.2

1.1

1.2

1.3

1.4

1.2

1.1

1.1

Distance between tentacle and nostril 1.0

1.2

1.1

1.2

1.2

1.3

1.2

1.3

1.3

Distance between tentacle and lip

0.4

0.2

0.2

0.2

0.2

0.2

0.2

0.2

0.2

Length of terminal shield

1.5

4.2

4.4

3.3

3.5

5.0

4.5

4.5

5.1

Denticulations surrounding vent

10

13

9

11?

10

?

9

?

11

Premaxillary-maxillary teeth

20

20

20

18

20/21

22/23

18

19

20

Vomerine teeth

5

5

5

6

6

7

6

6/7

5

Palatine teeth (left, right)

5,6

4,4

4,4

5,4

6,6

5,6

3,4

3,3

4,5

Distance between snout tip and lip Distance between tentacle and CM

1.4

Dentary teeth

16

18

16/17

18/19

20/22

19/20

17

18

18/19

Inner mandibular teeth

2

1 or 2

2

2

3

4

2

2

2



19

FIGURE 3. A specimen of Boulengerula fischeri in life.

In almost all analyses including four or three outgroups (except MP analysis excluding third codon positions of mt protein-coding genes), B. fischeri is recovered as sister to all remaining Boulengerula (e.g., Fig. 4a) but mostly with only moderate support (Table 4), and this support is reduced when third codon positions are removed for the mt protein-coding genes. In contrast, the analyses with Herpele as sole outgroup all recover B. boulengeri as sister to all other Boulengerula (e.g., Fig. 4b), and support for this relationship does not notably decrease with the removal of third codon positions for mt protein-coding genes (Table 4). The pairing of B. fischeri and B. boulengeri to the exclusion of all other Boulengerula is the least well supported resolution of the three main Boulengerula lineages in all analyses (Table 4).

Discussion Although local people reported that B. fischeri were not difficult to find in farmland outside the edge of one section of the forest at Cyamudongo, abundances within the forest as assessed by digging (0 per person hour), pitfall trapping (0 per bucket day) and VES (0.016 per person hour overall; 0.045 in Cyamudongo Forest) were zero to low. Very little time (< 10 person hours) was spent digging in forest habitats, and so we have no sound basis for comparing the abundance of B. fischeri with reported data for congeners (Gower et al., 2004; Malonza & Müller, 2004; Measey, 2004; Measey & Barot, 2006; Measey et al., 2006). Where reported, other species of Boulengerula have been found almost exclusively by digging in soil (and occasionally under logs/stones and in leaf litter) but not on the surface and rarely by pitfall trapping (Gower et al., 2004; Malonza & Measey, 2005; Measey, 2004; Measey et al., 2006; Measey & Barot, 2006; DJG, SPL & MW unpublished data), consistent with the hypothesis that they spend the vast majority of time within the soil. It might be noted in the present case that there is a chance that B. fischeri could have escaped pitfall buckets through drainage holes. Although the three B. fischeri found in forest were in leaf litter, this covered a path of fairly compact soil and we do not know how much time this species might normally spend in such a microhabitat. Herrel & Measey (2010) reported that B. fischeri had the least amount of skinvertebral independence during locomotion among their sampled caecilian species, and noted that this is possibly associated with a comparatively lower burrowing ability. We have no reason to think that B. fischeri does not resemble other Boulengerula in most commonly occurring in soil (supported by locals collecting B. fischeri outside the edge of Cyamudongo Forest by digging).


GOWER ET AL.

FIGURE 4. Best trees from ML analysis of mt (12S, 16S, cytb, cox1) and nuclear (RAG1) DNA sequences. Third codon positions for mt protein-coding genes (cytb, cox1) excluded, and data partitioned by gene and codon position. a) with four outgroups. b) with Herpele squalostoma as a single outgroup.



21

TABLE 4. Quantitative support for selected clades in Maximum Likelihood (ML), Bayesian (MrB) and Maximum Parsimony (MP) analyses. Number of partitions, where relevant, given in parentheses. “B. taitanus” = all sampled Boulengerula except B. fischeri and B. boulengeri. all positions

without 3rd positions

ML (11)

ML (5) MrB (11)

MrB (5)

MP

ML (9) ML (5)

MrB (9)

MrB (5)

MP

Boulengerula + Herpele

100

100

100

100

82

100

100

100

100

97

Boulengerula

100

100

100

100

84

100

100

100

100

98

B. boulengeri + B. taitanus

74

67

94

92

56

34

43

42

47

27

B. fischeri + B. taitanus

15

19

4

4

19

31

31

28

30

21

B. fischeri + B. boulengeri

11

14

3

5

19

31

26

31

23

50

Boulengerula

100

100

100

100

100

100

100

100

100

100


73

66

87

88

74

52

53

52

60

77


14

22

7

7

14

24

27

24

24

12


13

13

6

6

12

24

20

24

16

11


23

28

5

3

33

23

20

23

6

31


74

65

95

97

52

72

72

72

94

54


4

7

0

0

15

5

5

5

1

15

1. Four outgroups

2. Three outgroups (no Herpele)

3. One outgroup (Herpele)

Boulengerula fischeri is still known from only a very small area (Cyamudongo) and there has been little if any dedicated searching beyond this, so it should probably retain its current IUCN listing as Data Deficient (Wilkinson & Loader, 2004). However, the occurrence of this species in disturbed, agricultural habitat, and the probability that it also occurs more widely, including within the much larger and better protected Nyungwe Forest National Park (Behangana et al., 2009 did not survey Cyamudongo Forest, so their specimen of B. fischeri was from elsewhere) offer some hope that it might eventually qualify for Least Concern status. Jones et al. (2006) reported sexual dimorphism in B. boulengeri, in which males have relatively longer jaws and are wider across the back of the head, but this is not apparent in our small sample of B. fischeri. The new range in numbers of primary annuli (183–197) for the available sample of B. fischeri still serves to distinguish it from the congener with the second highest number, 161 in the single reported specimen of B. denhardti (Wilkinson et al., 2004). The presence of splenial teeth and lack of mesethmoid exposure on the dorsal surface of the skull roof are features that align B. fischeri with Taylor’s (1968) concept of Afrocaecilia, and distinguish it from the single species classified by Taylor in Boulengerula (B. boulengeri). A monophyletic Boulengerula was not recovered in Zhang & Wake’s (2009) molecular phylogenetic analyses, and recovered with only moderate support by Loader et al. (in press). However, the monophyly of Boulengerula is recovered strongly in our analyses of mitochondrial and nuclear genetic data. Although whether B. boulengeri or B. fischeri is sister to all other Boulengerula sampled here is not overwhelmingly clear, the molecular data provide strong support for three ‘major’ lineages within Boulengerula, based on long branches stemming from the two basalmost divergences – (1) B. fischeri, (2) B. boulengeri, and (3) all other Eastern Arc Mountain and Coastal Forest species from Kenya and Tanzania. Moderate support for B. fischeri as sister to all other sampled Boulengerula (Fig. 4a) comes only from analyses including distant outgroups and/or (saturated) third codon positions of mt protein-coding genes (Table 4), suggestive of ‘long-branch’ attraction. For this reason, we prefer the hypothesis that B. boulengeri is sister to all other sampled Boulengerula (Fig. 4b) – it is a moderately well supported hypothesis in analyses including only the most proximate available outgroup to Boulengerula (Herpele), whether or not mt third codon position sites are included.


GOWER ET AL.

In a phylogenetic analysis of morphological data, Nussbaum & Hinkel (1994) recovered B. fischeri in a clade including also B. boulengeri, B. changamwensis, and the South American Brasilotyphlus braziliensis. Wilkinson et al. (2004) reported that B. denhardti is also recovered in this clade when it was scored for Nussbaum & Hinkel’s (1994) morphological characters, but they demonstrated also that quantitative support for relationships across the tree is very low with many alternative resolutions not significantly less parsimonious, and argued that the monotypic Brasilotyphlus might not be closely related to Boulengerula. Nussbaum & Hinkel (1994) argued that failure to recover a monophyletic group comprising all Boulengerula except B. boulengeri rendered Taylor’s (1968) bipartition of the genus untenable, and they thus relegated Afrocaecilia Taylor, 1968 to a junior synonym of Boulengerula Tornier, 1896. Wilkinson et al. (2004) argued that this synonymy was premature given the lack of support for Nussbaum & Hinkel’s (1994) phylogenetic hypothesis, and Loader et al. (in press) subsequently found maximal support for the monophyly of Taylor’s Afrocaecilia in molecular phylogenetic analyses. The molecular phylogenetic results presented here potentially complicate the picture in that B. fischeri (a species unknown to Taylor) matches Taylor’s diagnosis of Afrocaecilia rather than Boulengerula in having splenial teeth and lacking dorsal mesethmoid exposure, but might not form a clade with the species that Taylor referred to Afrocaecilia. However, we do not consider this to be fatal for Taylor’s Afrocaecilia because (1) the species separated into that genus (taitanus, changamwensis, uluguruensis) and Boulengerula (boulengeri) are clearly deeply divergent, (2) the positions of B fischeri and B. denhardti are not yet robustly resolved, (3) we interpret the available molecular data to favour the hypothesis that B. boulengeri is sister to all other Boulengerula, and (4) detailed studies of morphology across the group have yet to be carried out. Based on (admittedly rather uncompelling) results of morphological phylogenetic analyses (Nussbaum & Hinkel, 1994; Wilkinson et al., 2004), we predict that the only nominate species of Boulengerula yet to be included in molecular phylogenetic analyses, B. denhardti, does not lie within the taitanus-niedeni-changamwensis-uluguruensis clade. The lack of splenial teeth in B. denhardti is an apomorphic feature suggesting affinity with B. boulengeri, but the high number of annuli and vertebrae are a possible synapomorphy of B. denhardti and B. fischeri. In addition to including B. denhardti in molecular phylogenetic analyses, studies of a greater number of anatomical characters across Boulengerula are needed before a natural and informative classification can be settled upon. Although monophyletic, Boulengerula as currently conceived might not hold together usefully as a single genus given that it is the genus with by far the oldest (possibly > 100 million years) intrageneric splits among known caecilians (Loader et al., in press; Roelants et al., 2007; Zhang & Wake, 2009).

Acknowledgements Fieldwork was funded by the Wildlife Conservation Society Rwanda Program and the Museo Tridentino di Scienze Naturali, Trento. Office Rwandais du Tourisme et des Parcs Nationaux provided the research authorizations and export permissions. For practical assistance and support we thank Michele Menegon, Gill Sparrow, Lakshmi Subrahmanian, Andrew Plumptre, Nerissa Chao, Felix Mulindahabi, Eugene Kayijamahe, Nsengiyunva Barakabuye, Joy Olasunmibo Ogunmakin, Daniel Nyonsaba, Martin Sindikubwabo, Jeremy Nzarora and Mika Nsanzimana.

References Behangana, M., Kasoma, P.M.B. & Luiselli, L. (2009) Ecological correlates of species richness and population abundance patterns in the amphibian communities from the Albertine Rift, East Africa. Biodiversity and Conservation, 18, 2855–2873. Castresana, J. (2000) Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Molecular Biology and Evolution, 17, 540–552. Fischer, E. & Hinkel, H. (1992) Natur und Umwelt Ruandas. Einführing in die Flore und Fauna Ruandas und ihre wichtigsten Biotopsysteme. Ministerium des Innern und für Sport, Rheinland-Pfalz, Mainz, Germany. 452 pp. Frost, D.R., Grant, T., Faivovich, J., Bain, R.H., Haas, A., Haddad, C.F.B., de Sà. R.O., Channing, A., Wilkinson, M., Donnellan, S.C. Raxworthy, C.J., Campbell, J.A., Blotto, B., Moler, P., Drewes, R.C., Nussbaum, R.A., Lynch, J., Green, D.M. & Wheeler, W.C. (2006) The amphiban tree of life. Bulletin of the American Museum of Natural History, 297, 1–370. Gower, D.J., Loader, S.P., Moncrieff, C.B. & Wilkinson, M. (2004) Niche separation and comparative abundance of Boulengerula boulengeri and Scolecomorphus vittatus (Amphibia: Gymnophiona) in an East Usambara forest, Tanzania. African Journal of Herpetology, 53, 183–190.



23

Gower, D.J., Wilkinson, M., Sherratt, E. & Kok, P.J.R. (2010) A new species of Rhinatrema Duméril & Bibron (Amphibia: Gymnophiona: Rhinatrematidae) from Guyana. Zootaxa, 2391, 47–60. Herrel, A. & Measey, G.J. (2010) The kinematics of locomotion in caecilians: effects of substrate and body shape. Journal of Experimental Zoology, 313A, 303–309. Jones, D.T., Loader, S.P. & Gower, D.J. (2006) Trophic ecology of East African caecilians (Amphibia: Gymnophiona), and their impact on forest soil invertebrates. Journal of Zoology, London, 269, 117–126. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T.J. & Higgins, D.G. (2007) Clustal W and Clustal X version 2.0. Bioinformatics, 23, 2947–2948. Loader, S.P., Wilkinson, M., Cotton, J.A., Müller, H., Menegon, M., Howell, K.M. & Gower, D.J. (in press) Molecular phylogenetics of Boulengerula (Amphibia: Gymnophiona: caeciliidae) and implications for taxonomy, biogeography and conservation. Herpetological Journal. Malonza, P.K. & Measey, G.J. (2005) Life history of an African caecilian: Boulengerula taitanus Loveridge 1935 (Amphibia Gymnophiona Caeciliidae). Tropical Zoology, 18, 49–66. Malonza, P.K. & Müller, H. (2004) A rediscovery after two decades: The Changamwe lowland caecilian Boulengerula changamwensis Loveridge, 1932 (Amphibia: Gymnophiona: Caeciliidae). Journal of East African Natural History, 93, 57–61. Measey, G.J. (2004) Are caecilians rare? An East African perspective. Journal of East African Natural History, 93, 1–21. Measey, G.J. & Barot, S. (2006) Evidence of seasonal migration in a tropical subterranean vertebrate. Journal of Zoology, 269, 29–37. Measey, G.J., Mejissa, J. & Müller, H. (2006) Notes on the distribution and abundance of the caecilian Boulengerula uluguruensis (Amphibia: Gymnophiona: Caeciliidae) in the Uluguru Mountains, Tanzania. African Journal of Ecology, 44, 6– 13. Nussbaum, R.A. & Hinkel, H. (1994) Revision of East African caecilians of the genera Afrocaecilia Taylor and Boulengerula Tornier (Amphibia: Gymnophiona: Caeciliaidae). Copeia, 1994, 750–760. Posada, D. (2008) jModelTest: Phylogenetic Model Averaging. Molecular Biology and Evolution, 25, 1253–1256. Roelants, K., Gower, D.J., Wilkinson, M., Loader, S.P., Biju, S.D., Guillaume, K. & Bossuyt, F. (2007) Patterns of diversification in the history of modern amphibians. Proceedings of the National Academy of Sciences of the United States of America, 104, 887–892. Ronquist, F. & Huelsenbeck, J.P. (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics, 19, 1572–1574. San Mauro, D., Gower, D.J., Oommen, O.V., Wilkinson, M. & Zardoya, R. (2004) Phylogeny of caecilian amphibians (Gymnophiona) based on complete mitochondrial genomes and nuclear RAG1. Molecular Phylogenetics and Evolution, 33, 413– 427. San Mauro, D., Gower, D.J., Massingham, T., Wilkinson, M., Zardoya, R. & Cotton, J.A. (2009) Experimental design in caecilian systematics: phylogenetic information of mitochondrial genomes and nuclear rag1. Systematic Biology, 58, 425–438. Stamatakis, A. (2006) RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics, 22, 2688–2690. Swofford, D.L. (2002) PAUP*: Phylogenetic Analysis using Parsimony. Version 4.0b. Sunderland, MA: Sinauer Associates. Taylor, E.H. (1968) The caecilians of the World: a taxonomic review. University of Kansas Press, Lawrence, KA, USA. 848 pp. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. (1997) The ClustalX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research, 25, 4876–4882. Wilkinson, M. & Loader, S. (2004) Boulengerula fischeri. In: IUCN 2010. IUCN Red List of Threatened Species. Version 2010.2. . Downloaded on 25 August 2010. Wilkinson, M. & Nussbaum, R.A. (2006) Caecilian phylogeny and classification. In: Exbrayat, J.-M. (Ed.), Reproductive Biology and Phylogeny of Amphibia Volume 3. Gymnophiona. Science Publishers Inc., Enfield, NH, USA, pp. 39–78. Wilkinson, M., Loader, S.P., Gower, D.J., Sheps, J.A. & Cohen, B.L. (2003) Phylogenetic relationships of African caecilians (Amphibia: Gymnophiona): insights from mitochondrial rRNA gene sequences. African Journal of Herpetology, 52, 83– 92. Wilkinson, M., Loader, S.P., Müller, H. & Gower, D.J. (2004) Taxonomic status and phylogenetic relationships of Boulengerula denhardti Nieden, 1912 (Amphibia, Gymnophiona, Caeciliidae). Mitteilungen aus dem Museum für Naturkunde Berlin, Zoologische Reihe, 80, 41–51. Wilkinson, M., Nussbaum, R.A. & Hoogmoed, M. (2009) A new species of Microcaecilia (Amphibia: Gymnophiona: Caeciliidae) from Suriname. Herpetologica, 65, 413–418. Zhang, P. & Wake, M.H. (2009) A mitogenomic perspective on the phylogeny and biogeography of living caecilians (Amphibia: Gymnophiona). Molecular Phylogenetics and Evolution, 53, 479–491.


GOWER ET AL.