Recent fieldwork in southern Tanintharyi revealed the presence of a small Green Crested Lizard in the wet evergreen forest. We generated mtDNA sequence data (ND2) that demonstrates that this population’s nearest relative is Bronchocela rayaensis Grismer et al., 2015 of Pulau Langkawi, northwestern Peninsular Malaysia and Phuket Island. Morphologically the Burmese Bronchocela shares many features with B. rayaensis, which potentially would make this recently described Thai-Malay species a synonym of Bronchocela burmana Blanford, 1878; however, we interpret the genetic and morphological differences to reflect evolutionary divergence and recommend the recognition of both species.

Reptilia , Southeast Asia, Tanintharyi Division, Thailand, Peninsular Malaysia, morphology, molecular phylogeny, synonymy, nomenclature

Bronchocela is a light weight among agamid lizards, rivaling Draco in the lightness and slenderness of its body and limbs. Draco’s slenderness is associated with its gliding locomotion. The slenderness and extremely long tail (~ 3× body length) of Bronchocela seem to be an adaptation for moving on the outer edge of the branches of shrubs and trees.

The first member of the Bronchocela clade to be recognized was Agama cristatella Kuhl, 1820 (now B. cristatella). This species continues to be the most commonly recognized Bronchocela owing to its broad distribution from Peninsular Malaysia to the Philippines and into the Lesser Sunda Islands (Grismer et al. 2015). The first mention of a Bronchocela in Burma (Myanmar) was in 1878 by Blanford. He recognized that a Burmese specimen differed strikingly from B. cristatella and the other species described to that time and named this taxon B. burmana. One major difference was the height, shape, and number of nuchal crest scales. Boulenger (1880) did not believe that this Burmese species was unique and synonymized B. burmana with Calotes cristatellus, a name used then because Boulenger also did not consider Bronchocela Kaup, 1827 to be a unique group even though Bronchocela had more than 50 years of recognition by other major authorities in herpetology. Malkmus (1992) appears to be the first author to use Bronchocela in a modern publication; however, his use was limited to B. cristatella presence in the Mount Kinabalu (Sabah, Borneo) area. Diong and Lim’s (1998) review of B. cristatella is the first formal use of Bronchocela with a full explanation of their decision to revive the generic name. Their decision was based on a confirmation of Scott Moody’s data in his unpublished dissertation demonstrating that cristatella was not closely related to other members of the Calotes clade.

Bronchocela burmana has remained a forgotten name until the present, in part because of Boulenger’s nomenclatural authority, but also because few Bronchocela have been collected in Myanmar owing to its limited occurrence to southern Tanintharyi (Fig. 1) and the failure to examine closely the few specimens from southern Burma. One of us (DGM) has been participating in rapid assessment surveys of proposed national parks in southern Tanintharyi. He first discovered two specimens of Bronchocela in the Lenya NP area in 2015 and one in the Lenya NP Extension in 2016. We likely would have labeled these specimens as B. cristatella without examining them closely if part of the survey did not also include mtDNA barcoding of all collected specimens. When additional sequence data (ND2) indicated that the Lenya specimens were not closely related to B. cristatella, instead to the recently described Bronchocela rayaensis (Grismer et al., 2015) from Pulau Langkawi, northwestern Peninsular Malaysia and Phuket Island, Thailand (Grismer et al. 2016), we decided to make a morphological comparison as well.

Figure 1. 

Distribution of Bronchocela burmana (solid circles) in southern peninsular Myanmar, Taylor’s (1963) two localities for B. cristatella (open squares) in southern Thailand, and B. rayaensis type locality (star) in northwestern Peninsular Malaysia and its newly reported localities (open circles) in Thailand (Grismer et al. 2016). A solid diamond denotes the type locality of B. burmana. The red dashed lines depict the political boundaries between Myanmar-Thailand, Cambodia-Thailand, and Malaysia-Thailand. Base map from: CC BY-SA 3.0, https://commons.wikimedia.org/w/index.php?curid=166887.

We present below our analysis of the molecular and morphological data for the southern Burmese Bronchocela specimens. As our title indicates, these data demonstrate that the Burmese population is unique.

We obtained ND2 sequence data from eight individuals ranging in size from 1,260–1,369 bp that were 98.5–99.8% identical to one another (un-corrected sequence divergence). Our samples have a truncated origin of light strand replication (21 bp) between tRNA^ASN and tRNA^CYS that appears functional (folds using the DNA-Matthews 1999 Energy Model in Geneious) and tRNA^CYS shows a d-arm replacement loop (Macey et al. 1997) consistent with 45% of Draconinae agamids (Seligmann and Labra 2014). Our samples formed a well-supported clade (100%) with each other and were placed sister to Bronchocela rayaensis with 100% bootstrap support (Fig. 2). Our samples ranged from 93.4–95.0% identical (un-corrected sequence divergence) to the B. rayaensis specimens in GenBank (KR053115–KR053116, KY498356). Our analyses recovered the Phuket Island B. rayaensis sister to the Pulau Langkawi samples with 94% bootstrap support (Fig. 2). The rest of our tree was nearly identical to that recovered by Grismer et al. (2015) with the exception of the inclusion of B. cristatella (AF128495) and Aphaniotis fusca (AF128497); the latter was placed sister to Aphaniotis fusca (AF288228).

Figure 2. 

Phylogenetic relationships of Bronchocela. Maximum-likelihood phylogeny based on 1,416 bp of ND2 mtDNA, bootstrap values (based on 1,000 replicates) are shown for the major nodes relevant to this study. Terminals highlighted in blue were recently corrected in GenBank and the B. cristatella from the Philippines (in yellow) are now re-identified as B. cf. marmorata.

All measurements, proportions, and scalation were examined for differences between adult females (n = 5) and adult males (n =3). These samples are small so the significance of our statistical results are indicative of dimorphism but not statistical proof. Among the 33 mensural traits tested, only seven suggest dimorphism (means and ranges compared). Females average smaller (SVL 80.8, 78.8–83.8 mm) than males (SVL 88.2, 84.3–92.7 mm), fourth finger shorter in females (means, 11.0 vs. 12.6 mm 4FingL), fourth toe shorter in females (16.6 vs. 18.6 mm 4ToeL), hindlimb distinctly shorter (74.3 vs. 81.2 mm HindlL), head length less (21.2 vs. 22.6 mm HeadL), length of nuchal crest less (9.6 vs. 12.3 mm NucCrsL), and crest length to orbit diameter also shorter (150% vs. 176% NucCrsL/OrbD). No scalation features display sexual dimorphism.

To summarize the statistical results for measurements and proportions (values, mean and minimum-maximum): dimorphic traits, females’ SVLx = 80.8, 78.8–83.8 mm, 4FingL 11.0, 10.4–11.6 mm, 4ToeL 16.6, 15.1–17.9 mm, HindfL 29.5, 29.0–30.0 mm, HeadL 21.2, 20.3–22.0 mm, NucCrsL 9.6, 9.1–10.1 mm, NucCrsL/OrbD 150, 132–168%; males’ SVLx = 88.2, 84.3–92.7 mm, 4FingL 12.6, 11.3–13.9 mm, 4ToeL 18.6, 18.2–18.8 mm, HindfL 32.3, 30.9–33.1 mm; HeadL 22.6, 21.6–23.3 mm, NucCrsL 12.3, 10.6–13.3 mm, NucCrsL/OrbD 176, 166–184%; monomorphic traits (adults combined) TrunkL 42.8, 40.4–46.6 mm, TailL 276.3, 204–306 mm, ForelL 45.5, 42.2–51.5 mm, ForefL 16.3, 14.7–18.2 mm, HindlL 76.9, 70.6–86.6 mm, HindfL 30.6, 29.0–33.1 mm, HeadW 11.6, 9.5–14.1 mm, HeadH 10.2, 9.3–11.3 mm, OrbD 6.6, 5.9–7.5 mm, TympD 3.2, 2.4–3.5 mm, BSC.dors 0.6, 0.5–0.8 mm, BSC.vntl 1.7, 1.5–2.1mm, NucCrsH 1.4, 0.9–3.0, TrunkL/SVL 51, 47–56%, TailL/SVL 328, 243–361%, ForelL/SVL 54, 50–58%, HindlL/SVL 92, 86–97%, HeadL/SVL 26, 25–27%, HeadW/HeadL 54, 44–62%, HeadH/HeadL 47, 45–50%, ForefL/ForelL 35, 29–40%, HindfL/HindlL 40, 38–43%, ForelL/HindlL 21, 20–23%, 4FingL/4ToeL 67, 60–74%, TympD/OrbD 48, 40–58%, DorsS/VntlS 35, 31–39%, NucCrsL/OrbD 160, 132–184%.

The results for scalation are (values, median and minimum–maximum; adults and juvenile combined): Suplab 10, 9–11, Inflab 10, 9–11, Loreal 5, 5–6, Postm 3, 3–3, Throat 26, 25–30, 3ForefLm 28, 26–32, 4HindfLm 33, 31–38, Midbody 59, 55–67, Dorsal1 7, 6–8, Dorsal2 17, 15–21, NucCrsS 8, 6–9; nuchal spines or scales are predominantly triangular, rarely broadly crescent shaped; middorsal trunk scales are large, keeled but not elevated into a crest.

We attempted to code a few coloration traits but were unsuccessful because preservation had altered life-colors and pattern. The manner of field work and genomic tissue sampling did not permit the recording of coloration in living specimens. The general impression of the Tanintharyi Bronchocela is green with transverse bands of white spots on trunk and alternating bands of green and white on the tail (Fig. 3). Preserved specimens have a broadly mottled pattern of dark rufous brown and dusky turquoise dorsally and laterally. These colors form broad swatches with no consistence of position on the body although the majority (eight of nine) of the specimens have the snout and interorbital (not the superciliary area) dark. The middle third of the tail (dorsally) shows a vague banding, broad bands of medium brown and dusky tan. The venter from the tip of the chin onto the tail base is light turquoise with the chin and throat lighter that chest and abdomen. The temporal area (between orbit and tympanum) is bright rufous in eight of nine specimens although the size of this rufous patch varies from a small spot to the entire area. The eye sheath is dark, the tympanum light. The loreal area also is rufous and size of rufous patch varies from a small patch to the entire area. The ventrolateral neck folds range from uniform dark rufous to uniform turquoise. In two instances of the latter, the posterior edge of the fold bears small, ill-defined rufous spots.

Figure 3. 

Bronchocela burmana Blanford, 1878 from the Lenya area (circa 11.68N 99.42E). A dorsolateral view of a living Burmese Crested Lizard, USNM 587483. Photo by DGM.

For comparative purposes, we extracted the equivalent measurement and scalation data from Grismer et al.’s (2015: tables 5 & 6; 2016: table 1) description of B. rayaensis. Both of the types were adults, holotype a male and paratype a female, and the newly discovered Thai adult female. The following is a summary for the types (see Table 2 for summary of the three known specimens): female’s SVL 85.4 mm, HeadL 20.1 mm; male’s SVL 82.0 mm, HeadL 21.8 mm; monomorphic traits (both types) TrunkL not reported, TailL 305.5, 303–308 mm, ForelL 49.1, 48.2–50.0 mm, ForefL 15.7, 15.7–15.7 mm, HindlL 79, 78–81mm, HindfL not reported, HeadW 12.2, 12.1–12.4 mm, HeadH 10.8, 10.4–11.2 mm, OrbD not reported, TympD not reported, BSC.dors not reported, BSC.vntl not reported, NucCrsH not reported, TrunkL/SVL not reported, TailL/SVL 366, 355–376%, ForelL/SVL 59, 59–59%, HindlL/SVL 95, 91–99%, HeadL/SVL 25, 24–27%, HeadW/HeadL 58, 56–59%, HeadH/HeadL 52, 49–56%, ForefL/ForelL not reported, HindfL/HindlL 37, 37–38%, ForelL/HindlL not reported, 4FingL/4ToeL not reported, TympD/OrbD 48, 46–49%, DorsS/VntlS not reported, NucCrsL/OrbD not reported. For scalation, values are: Suplab 10, 10–11, Inflab 9, 8–10, Loreal 6, 5–7, Postm 3, 3–3, Throat not reported, 4ForefLm 32, 31–33, 4HindfLm 33 (holotype), Midbody 71 (holotype), Dorsal1 6, 5–8, Dorsal2 not reported, NucCrsS 11, 9–13; nuchal spines or scales are lanceolate; middorsal trunk scales are large, keeled but not elevated into a crest.

The molecular (ND2) results support our Tanintharyi specimens as a discrete clade, 5–6.6% different from the closest species Bronchocela rayaensis available. Our samples were placed sister to B. rayaensis, a species known from Pulau Langkawi (island) off the west coast of Peninsular Malaysia and Phuket Island, Thailand (Fig. 2). Our phylogeny recovered the Phuket Island B. rayaensis sister to the Pulau Langkwai samples with 94% bootstrap support. Grismer et al. (2016) recovered it nested among the Pulau Langkwai samples, possibly because of an inaccurate alignment of the tRNA and OL region. Grismer et al. (2016) re-identified previously published B. cristatella samples (KR053113–KR053114; KU326254–KU326255, RMB8882–RMB8883 of Grismer et al., 2015) from Polillo Island, Philippines as B. marmorata, but did not update the taxonomy of those sequences in GenBank. Here, we included all samples from the Philippines from Grismer et al. (2015), the two from Polillo Island and two from Luzon Island, Philippines [KR053111–KR053112, CDS2105 (KU305472) and RMB 9878 (KU315805), respectively]. These samples form a well-supported clade and are 94.4–100% identical to one another for the ND2 sequences. Because there may be more than one lineage of Bronchocela in the Philippines (R. Brown, pers. com.), we refer to these specimens as B. cf. marmorata and they are sister to a clade containing B. shenlong from Peninsular Malaysia and B. cristatella from Malaysia, Borneo, and Indonesia (Pulau Natuna Besar). Though the sister relationship between B. shenlong and B. cristatella received 100% support (Fig. 2), the relationship between them and B. cf. marmorata received much less support, both in our study (51%, Fig. 2) and in that of Grismer et al. (2015, 72% ML bootstrap support); though with fewer individuals analyzed, this relationship received 100% support (Grismer et al. 2016).

We included all specimen in GenBank from Grismer et al. (2015), including outgroups, the B. rayaensis from Phuket Island (Grismer et al. 2016), and an additional Bronchocela cristatella sequence in GenBank (AF128495) and Aphaniotis fusca (AF128497). These specimens were reported as being switched (Zug et al., 2006), which our study confirms. The sequence AF128495 is 99% identical to B. cristatella from Selangor, Malaysia (KR053100–KR053102) and was placed in the B. cristatella clade sister to these specimens from the same state (Selangor) as TNHC57943, the voucher of AF128495 (Fig. 2). The identifications of these sequences were recently corrected in GenBank. The Fig. 5 of Macey et al. (2000b) appears to display the correct tRNA^CYS for these genera; however, our ARWEN model depicts the lower three base-pairs non-binding, extending the d-arm replacement loop to “AAAGTG.”

The tRNAs of Grismer et al. (2015) appear to be improperly annotated because virtually all tRNAs are annotated as transcribing in the light strand direction, whereas tRNA^ALA, tRNA^ASN, tRNA^CYS, and tRNA^TYR transcribe on the heavy strand in all other agamids examined (e.g. Macey et al., 1997), their sequences have “tRNA^ASX” (the transcribed product) in place of tRNA^ASN, and none of the origin of light strand (OL) sequences are identified. Secondary structure inspection of one of their specimens (KR053067) in ARWEN reveals the tRNA^ASN does exists, it is transcribed in the complimentary direction, and their tRNA^ALA as annotated does not form a tRNA, but the bases between tRNA^TRP and tRNA^ASN do form tRNA^ALA. Agamids, in particular, show interesting variation in the relationship between the condition of the OL and the neighboring tRNA’s d-arm replacement loop, which can be useful for the study of evolution of mtDNA replication (Seligmann and Labra 2014) and can be useful as phylogenetic information content (Macey et al. 2000b) if properly annotated.

The morphology of Bronchocela burmana and B. rayaensis are similar; however, the differences indicate adaptive divergence between the southern Burmese and Langkawi Malaysian populations. The Burmese lizards show sexual dimorphism with males larger than females with only a slight overlap. The Langkawi population seems to reverse this dimorphism with females larger than males (Table 2) although that sample includes only one adult of each sex.

The newly discovered Phuket Bronchocela rayaensis is a gravid female yet smaller than the paratype adult female (Table 2). As would be expected, its absolute dimensions are smaller than the typic female and surprisingly also smaller than the typic male. Also several of its body proportions (HeadH/HeadL, TympD/OrbD, ForelL/SVL, HindlL/SVL) are smaller than those of the types. Because the Phuket sample is a single individual, we note these differences, but also note that in B. burmana adult size remains larger, midbody and nuchal crest scales remain fewer, thereby supporting the Phuket individual as B. rayaensis.

Bronchocela rayaensis possess more nuchal crest scales than B. burmana—with a slight overlap 6–9 vs. 9–13. Third and fourth fingers are nearly equal in B. rayaensis, third finger is about half a claw length longer than fourth finger, and midbody scales 71 in holotype of B. rayaensis vs. 59 (55–67) scales around midbody in our Tanintharyi sample

At this time, we cannot define the distribution of B. burmana in Thailand. The distribution map in Chan-ard et al.’s field guide (2015: p. 94) shows a Thai occurrence of B. cristatella in the Tenasserim range from Khanchanburi Province southward into southern Peninsular Thailand to the Malaysia border. Their morphological description appears to mix characteristics of B. burmana and B. cristatella. We assume that this characterization demonstrates that the distribution of their Thai Bronchocela specimens was a mix of B. burmana and B. cristatella. The Phuket and Khura Bun specimens (Grismer et al. 2016) confirm the presence of B. rayaensis in Thailand, at least south of the Isthmus of Kra. In an earlier review of Thai lizards, Taylor (1963) considered cristatella as a Calotes. His description was based on a specimen from Yala Province and his illustration on an individual from Sangkhla. Both of these specimens are clearly B. cristatella, demonstrating that B. cristatella occurs in southern-most peninsular Thailand. Chan-ard et al.’s (1999) photographic atlas depicts two Bronchocela; the one from Malaysia (left photograph, p. 92) is an adult female B. cristatella, the one from Khao Soi Dao, Chanthiburi is a B. smaragdina, thus, of no assistance in defining the distribution of B. burmana in Thailand or confirming its presence there. For the present, we suggest that the northern half of the distribution of B. cristatella as shown in the Chan-ard et al. map (2015) is B. burmana from the Isthmus of Kra northward, B. rayaensis south of the isthmus, perhaps restricted to the western side of the Thai–Malay Peninsula, and B. cristatella occurs southward in the eastern lowlands into Malaysia.

Bronchocela rayaensis has been recently characterized by Grismer et al. 2015 and 2016 and is not repeated here. The following is a redescription of B. burmana based exclusively on our recently vouchered specimens from southern Tanintharyi.

All or portions of the laboratory and/or computer work were conducted in and with the support of the Laboratories of Analytical Biology facilities of the National Museum of Natural History (NMNH) or its partner labs and lab work was funded by the Global Genome Initiative (NMNH) and lab work was assisted by L. Dickens Jr., A. Ibarra, and B. Cruz of the Youth Engagement thru Science Global Genome (YES!-GG) Program. E. Buring (NMNH HS-Intern) assisted with the tRNA secondary structure analyses. Specimens of Bronchocela from Tanintharyi were collected by two survey teams. The first rediscovery of B. burmana were made in Kawthoung District in 2010 by a Myanmar Herpetology Survey team of rangers from the Myanmar Nature and Wildlife Conservation Division under the supervision of J. V. Vindum. The survey program was country-wide and its main support derived from the Biodiversity program of the National Sciences Foundation (DEB-9971861, DEB-0451832). The second vouchering occurred in 2015 and 2016 as part of Fauna & Flora International’s (FFI-Myanmar Program) biotic assessment of the southern Tanintharyi evergreen forest. Additional funding for the Smithsonian field team derived from the Smithsonian Myanmar Biodiversity Initiative and its support by The Leona M. and Harry B. Helmsley Charitable Trust (http://www.helmsleytrust.org) awarded to Melissa Songer (Smithsonian Conservation Biology Institute). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We thank Fauna & Flora International, Yangon office, specifically M. Grindley, F. Momberg, Nay Myo Shwe, Saw Soe Aung and their field crew (particularly Myint Kaw Thura and Thaw Zin) for in-country logistics and assistance with field surveys. The base map image of Tanintharyi and adjacent areas is the base map from https://commons.wikimedia.org/w/index.php?curid=166887; we appreciate and thank the publisher for the use of the image to depict the known distribution of Bronchocela in that area. We also wish to thank S. Gotte, K. Tighe, and A. Wynn of the USNM collections management team for their always-ready assistance. We thank L. Grismer and J. Moravec for reviewing our manuscript and L. Grismer for quickly providing us with access and consent to publish the sequence of ZMKUR0017, and J. Schulte for following up on correcting AF128495 and AF128497 sequence identifications in GenBank.

Bronchocela burmana.

Myanmar; Tanintharyi Division: CAS247755 – CAS247757, CAS24786, USNM587775 – USNM587776, USNM587483 – USNM587484, USNM587793.

Bronchocela cristatella.

Malaysia; Johore Province: USNM29576, USNM29579; Selangor Province USNM129482, USNM141704 – USNM141706.

Thailand; Songgkhla Province: EHT 349; Yala Province EHT 134; data from Taylor 1963.

Bronchocela rayaensis.

Malaysia; Kedah Province: data from Grismer et al. 2015 and 2016.