sign in sign up
<<
Home , Chamaeleo monachus, Socotra

Molecular Phylogenetics and Evolution 49 (2008) 1015–1018

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier.com/locate/ympev

Short Communication Socotra the forgotten fragment of : Unmasking lizard history with complete mitochondrial genomic data

J. Robert Macey a,b,*, Jennifer V. Kuehl c, Allan Larson d, Michael D. Robinson e, Ismail H. Ugurtas f, Natalia B. Ananjeva g, Hafizur Rahman h, Hamid Iqbal Javed i, Ridwan Mohamed Osman j, Ali Doumma k, Theodore J. Papenfuss b a Department of Biology, Merritt College, 12500 Campus Drive, Oakland, CA 94619, USA b Museum of Vertebrate Zoology, University of California, Berkeley, CA 94720, USA c Department of Evolutionary Genomics, DOE Joint Genome Institute, Lawrence Berkeley National Laboratory, 2800 Mitchell Drive, Walnut Creek, CA 94598-1631, USA d Department of Biology, Box 1137, Washington University, St. Louis, MO 63130-4899, USA e Department of Biology, Sultan Qaboos University, P.O. Box 36, Al Khoud, PC 123, , f Department of Biology, Uludag University, 16059 Bursa, Turkey g Zoological Institute, Russian Academy of Sciences, St. Petersburg 199034, Russia h Zoological Survey Department, Government of Pakistan, Block 61, Pakistan Secretariat, Shahrah-e-Iraq, Karachi, Pakistan i Zoological Survey Department, Government of Pakistan, Kiyani Road, Bhara Kahu, Islamabad, Pakistan j Amoud University, Borama, k Faculty of Sciences, Abdou Moumouni University of Niamey, P.O. Box 10662 Niamey, article info

Article history: Received 4 February 2008 Ó 2008 Elsevier Inc. All rights reserved. Revised 1 July 2008 Accepted 28 August 2008 Available online 7 September 2008

Keywords: Reptilia Chamaeleonidae Gondwana Biogeography Phylogeny

1. Introduction Plate moved northward to Eurasia, Arabia came closer to than formerly expected, allowing to disperse from Arabia to Breakup of the southern Gondwana has been a India. The dynamics of formation of the Red Sea and Gulf of focus of historical biogeography for decades, but studies of Gondw- including nearby endemic may be critical to understanding ana rarely consider the Middle Eastern island of Socotra, located in interactions among Gondwanan land masses as they crossed the the at the triple junction of the Gondwanan plates of Tethys Sea. India, Arabia and . Phylogenetic analysis of 16 complete The faunas of land masses derived from the supercontinent mitochondrial genomic sequences reveals that chameleons of Gondwana provide classic examples of cladogenesis induced by Socotra form an ancient lineage and were separated from relatives geological origins of physical changes that fragmented geographic in North Africa and Arabia around the Oligocene, consistent with distributions (Bossuyt et al., 2006; Briggs, 1995). Phases of Gon- geological models of the isolation of Socotra from Arabia and Africa dwanan biogeography include (1) division of the original super- at this time. Phylogenetic grouping of an endemic Indian chame- into fragments separated by intervening ocean leading leon and two Arabian endemic species suggests that as the Indian to , (2) movement of the fragments across ocean with some land masses attaining close proximity including some transi- tory connections, (3) accretion of some Gondwanan fragments to * Corresponding author. Address: Department of Biology, Merritt College, 12500 Campus Drive, Oakland, CA 94619, USA. Laurasia and to each other to form new land masses, and (4) moun- E-mail address: [email protected] (J.R. Macey). tain barriers forming by tectonic plate collisions. Congruence

1055-7903/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2008.08.024 1016 J.R. Macey et al. / Molecular Phylogenetics and Evolution 49 (2008) 1015–1018 between geological events and cladogenesis in molecular phyloge- several species (Blanford, 1881; Günther, 1881). This expe- nies provides a strong foundation for understanding the historical dition confirmed that the chameleon, monachus, oc- assembly of diverse continental and island biotas. curred on the island of Socotra (an erroneous locality for the type Acrodont lizards (families Agamidae and Chamaeleonidae) specimen in the British Museum places it in , Gray, illustrate all four phases of Gondwanan biogeographic history. Dis- 1865). tinct acrodont groups are associated with the Gondwanan tectonic Phylogenetic analysis of C. monachus and its closest relatives on fragments of Afro-Arabia, India, Madagascar, Australia, and those other Gondwanan land masses offers insight into Socotra’s role in now forming parts of Southeast Asia (Macey et al., 2000b); contro- the first two phases of Gondwanan historical biogeography. This versy persists regarding which of these differences trace to the ori- is the first molecular phylogenetic study of chameleons endemic ginal fragmentation of Gondwana (phase 1), versus dispersal to Socotra (C. monachus) and India (C. zeylanicus). We report 16 between fragments in close proximity during phase 2 (Schulte complete mitochondrial genomic sequences and phylogenetic et al., 2003; Hugall and Lee, 2004). Raxworthy et al. (2002) and analyses to resolve the relationships and historical biogeography Townsend and Larson (2002) invoke recurring dispersal of chame- of Socotran, Indian, Mediterranean, Sub-Saharan African, and Ara- leons between Africa and Madagascar during phase 2. Phase 3 in- bian species of the Chamaeleo chamaeleon species group (Fig. 1) cludes introduction of agamid lineages to Laurasia via collision of plus outgroups from Sub-Saharan Africa and Madagascar. Afro-Arabia, Indian and Southeast Asian plates with Laurasia (Ma- cey et al., 2000b), cladogenesis in the Laudakia caucasia species 2. Methods group on the Iranian illustrates phase 4 (Macey et al., 1998, 2000a). We sequenced a complete mitochondrial genome from 13 in- The Socotra Island is a largely overlooked piece of Gon- group taxa (Fig. 1) and two outgroups (C. dilepis and Kinyongia), dwanan historical biogeography. Attached to the Afro-Arabian with partial sequence for a third outgroup, Brookesia, (all but the Plate, this island formed at the end of the Cretaceous (65 million second half of cob, trnP, trnT, presumptive Control Region, trnF, years before present, MYBP) as part of a massif (Beydoun and the first half of rrnS). Laboratory protocol follows Macey and Bichan, 1970). The northern portion of this massif remains in et al. (2006). Arabia in the Dhofar Mountains of southern Oman, but Eocene– DNA sequences were aligned manually. Positions encoding pro- Oligocene (34–41 MYBP) rifting in the Red Sea and teins were translated to amino acids using MacClade 4.03 (Madd- isolated Socotra Island from Arabia (Braithwaite, 1987; Girdler ison and Maddison, 2001) for confirmation of alignment. and Styles, 1974). As Arabia moved northward into Eurasia, Socotra Alignments of sequences encoding tRNAs and rRNAs were con- was isolated in the Indian Ocean. The first biological survey of structed using secondary structural models (Kumazawa and Nish- Socotra in 1880 emphasized the uniqueness of its biota. The first ida, 1993; Macey and Verma, 1997; Gutell and Fox, 1988; Hickson explorations, led by Scottish botanist Sir Isaac Bayley Balfour and et al., 1996a,b; Van de Peer et al., 1994). Regions with extensive partially funded by the British Association for the Advancement length variation were deemed unalignable and excluded from phy- of Science, discovered numerous taxa (Balfour, 1888) and logenetic analyses. In rRNA genes, 134 positions were excluded

40˚ 13 9 12 11

2

20˚ 8 5 7 10 3 4 6 1

0 500 1000 km 0˚ 20˚ 0˚ 20˚ 40˚ 60˚

Fig. 1. Map of North Africa, circum Mediterranean, and Middle East depicting the distribution of ingroup taxa in the genus Chamaeleo. Numbers represent localities of taxa sampled and are cross-referenced with the phylogenetic tree presented in Fig. 2. Shading depicts clades discovered in this study. Samples and clades are: Blue circle, (1) C. monachus (Socotra Island, ); pink, (2) C. zeylanicus (Pakistan); red, (3) C. calyptratus (Yemen), (4) C. arabicus (Yemen), (5) C. arabicus (Oman); green, (6) C. calcaricarens (Somaliland), (7) C. africanus (Savana, Niger), (8) C. africanus (Sahara, Niger); yellow, (9) C. chamaeleon (), (10) C. chamaeleon (Yemen), (11) C. chamaeleon (Cyprus), (12) C. chamaeleon (Anamur, Turkey), and (13) C. chamaeleon (Hatay, Turkey). J.R. Macey et al. / Molecular Phylogenetics and Evolution 49 (2008) 1015–1018 1017

Brookesia superciliaris stein, 1985) was applied to assess support for individual nodes Madagascar using 1000 replicates. Decay indices (= ‘‘branch support” of Bremer, Kinyongia fischeri 1994) were calculated for all internal branches using searches that retained suboptimal nodes. All searches were heuristic using 100 Chamaeleo dilepis random additions per replicate or search. Tanzania Museum voucher numbers and locality data for all newly re- C. monachus 1 100 Socotra ported sequences are in GenBank files EF222187–EF222202 (major Gulf of Aden 310 C. zeylanicus 2 extension of sequences refer to Macey et al., 1997, 2000b). India I 100 3. Results 99 65 C. calyptratus 3 Yemen 40 100 66 C. arabicus 4 Arabia Fig. 2 shows phylogenetic relationships among sampled chama- 100 Yemen leons. All branches except one are well supported by either a 99% 166 C. arabicus 5 or 100% bootstrap value and decay-index values of 27–397. Brooke- 100 Red Sea Rift Oman sia and Kinyongia root the tree, with Chamaeleo monophyletic; C. 150 C. calcaricarens 6 dilepis is the sister taxon to a clade comprising the remaining spe- Somaliland 99 South cies. Socotran C. monachus is the sister taxon to a clade comprising 33 C. africanus 7 of 100 Niger, Savana Arabian (C. arabicus, C. calyptratus), Indian (C. zeylanicus), Mediter- 397 Sahara ranean (C. chamaeleon) and two Sub-Saharan African (C. africanus, 100 C. africanus 8 Niger, Sahara C. calcaricarans) species. Indian C. zeylanicus is the sister taxon to 78 C. chamaeleon 9 an Arabian clade (C. arabicus and C. calyptratus). A Sub-Saharan Portugal clade (C. africanus and C. calcaricarens) is the sister taxon to C. cha- 100 C. chamaeleon maeleon, whose circum-Mediterranean samples form a clade with 204 10 Yemen II 100 North one from Yemen. The westernmost C. chamaeleon sample from Por- 27 C. chamaeleon 11 tugal is the sister lineage to a clade comprising other conspecific Cyprus of 100 Sahara samples. The isolated C. chamaeleon population from Yemen in 129 C. chamaeleon 12 Arabia is the sister taxon of a haplotype clade from Cyprus and Tur- 79 Ananmur, Turkey 2 key. Haplotypes from Cyprus and Turkey are nearly identical (only C. chamaeleon 13 Complete mt-genome Hatay, Turkey 11–18 nucleotide differences across 14,551 positions analyzed); Turkish haplotypes are grouped with weak support (bootstrap Fig. 2. Phylogenetic tree from parsimony analysis of 14,551 positions (3249 79%, decay index 2). informative) for the Chamaeleo chamaeleon complex. The tree has 10,924 steps. In group taxa are numbered as in Fig. 1. Taxa representing clades found in Arabia, and 4. Discussion the two sides of the Sahara Desert are represented on the right side. Bootstrap values are presented above branches and decay values are presented below branches. Major tectonic divides are denoted as ‘‘Gulf of Aden” and ‘‘Red Sea Rift”. The deep phylogenetic divergence of Socotran C. monachus from Arrows represent potential dispersal events: (I) Movement from Arabia to India; its closest relatives in Africa, Arabia and the Mediterranean Basin (II) movement from the Mediterranean Region to Arabia. Details are discussed in confirms the prediction that this lineage formed when geological the text. rifting in the Red Sea and Gulf of Aden isolated Socotra Island; initi- ated 35–41 MYBP, but may have occurred later as reflected in DNA divergence (see below). From our phylogenetic results for these cha- from rrnS and 276 from rrnL because of extensive length variation, meleons, we hypothesize a historical sequence of events predomi- particularly in loop regions. In tRNA genes, 159 excluded positions nately shaped by plate tectonics but followed by unique dispersal comprise: the D-loop of trnQ, I, M, W, Y, D, H and T; the D-arm events: (1) early vicariant isolation of a Socotra Island population replacement loop of trnC; the V (variable)-loop of trnS1; and the from African, Mediterranean and Arabian populations, (2) vicariant T-loop of trnV, I, M, A, C, Y, D, K, G, H, E, and T. For protein-coding separation of African and Arabian clades, (3) movement from Arabia regions a total of 257 nucleotide positions were excluded with to India prior to speciation within Arabia, (4) north–south specia- the number of omitted aligned positions for each gene in parenthe- tion by vicariance across the Sahara, (5) east–west speciation by ses: nad1 (27), nad2 (6), atp8 (3), atp6 (5), nad5 (135), nad5–nad6 vicariance across the Mediterranean Basin, and (6) dispersal from overlap (16), nad6 (47), and cob (18). A total of 237 non-coding the eastern Mediterranean region to southern Arabia. nucleotide positions were excluded from intergenic regions as fol- Following their isolation from Socotra, Africa and Arabia were lows: trnQ-I (10), trnW-A (10), trnA-N (17), stem-loop trnN-C (29), intermittently connected and disconnected during subsequent presumptive duplicated tRNA genes between trnC-Y in Brookesia movements of these plates. Arabian and Indian chameleons form (132; see Townsend and Larson, 2002), trnY-cox1 (23), trnS2-D the sister taxon to those from Africa and the Mediterranean region. (5), nad4-trnH (4), and cob-trnT (7). Among chameleons, trnP is Sub-Saharan species (C. africanus and C. calcaricarens) form the sis- located in the presumptive Control Region toward trnF and not ter taxon to Mediterranean populations (C. chamaeleon; although adjacent to trnT probably because of replication issues (see Macey these haplotypes form a monophyletic group, their deep divergence et al., 1997a,b, 2004; Kumazawa, 2007); regions between trnT and suggests that C. chamaeleon is a complex of undescribed species). P, and between trnP and F were excluded. Hence, a total of 1063 These results suggest that speciation across the Sahara followed nucleotide positions are excluded from rRNA, tRNA, and protein an initial isolation of Arabia from Africa across the Red Sea Rift. coding regions, as well as intergenic positions (not including the The endemic chameleon of India, C. zeylanicus, is the sister tax- unaligned presumptive Control Region) leaving an alignment of on to the two endemic chameleon species of Arabia (C. arabicus and 14,551 positions (3249 parsimony informative) for phylogenetic C. calyptratus). Hence, the Indian population separated from Ara- analysis. The sequence alignment is available in electronic version bian populations prior to speciation within Arabia but after the on the Molecular Phylogenic and Evolution website. Arabian populations had separated from African populations. Phylogenetic trees were inferred by parsimony using PAUP* Two scenarios could explain this result. One is that chameleons beta version 4.0b8 (Swofford, 2001). Bootstrap resampling (Felsen- moved from Arabia into Laurasia and formed a continuous 1018 J.R. Macey et al. / Molecular Phylogenetics and Evolution 49 (2008) 1015–1018 distribution between southern Arabia and the Indian Subcontinent, References followed by extinction in northern Arabia, southern Iran, and wes- tern Pakistan, where suitable habitat currently exists. The second Beydoun, Z.R., Bichan, H.R., 1970. The geology of Socotra Island, Gulf of Aden. Q.J. Geol. Soc. Lond. 125, 413–446. hypothesis is that as India made its final movement north across Blanford, W.T., 1881. Notes on the lizards collected in Socotra by Prof. I. Bayley the Indian Ocean toward Asia it came closer to Arabia than previ- Balfour. Proc. Zool. Soc. London 40, 464–469. ously assumed (Briggs, 2003) and possibly made contact with Ara- Balfour, I.B., 1888. Botany of Socotra. Trans. Roy. Soc. Edinburgh 31, 1–446. bia, permitting dispersal of chameleons from Arabia to India. Bossuyt, F., Brown, R.M., Hillis, D.M., Cannatella, D.C., Milinkovitch, M.C., 2006. Phylogeny and biogeography of a cosmopolitan frog radiation: Late Cretaceous Because chameleons are climbing they could have been diversification resulted in continent-scale endemism in the family Ranidae. swept by to mangrove swamps on the coast and poten- Syst. Biol. 55, 579–594. tially move between adjacent plates if the gap was not too wide, Braithwaite, C.J.R., 1987. Geology and palaeogeography of the Red Sea Region. In: Edwards, A.J., Head, S.M. (Eds.), Key Environments: Red Sea. Pergamon Press, similar to what has been hypothesized for Africa and Madagascar Oxford, pp. 22–44. in other chameleons (Raxworthy et al., 2002; Townsend and Lar- Briggs, J.C., 1995. Global Biogeography. Elsevier, Amsterdam , New York. p. 452. son, 2002). We favor the hypothesis of a close association of the Briggs, J.C., 2003. The biogeographic and tectonic history of India. J. Biogeogr. 30, 381–388. Arabian and Indian plates over wide-scale extinction of chame- Felsenstein, J., 1985. Confidence limits on phylogenies: An approach using the leons from northern Arabia and southwestern Asia. bootstrap. Evol. 39, 783–791. Among C. chamaeleon populations, the sample from Yemen in Girdler, R.W., Styles, P., 1974. Two stage Red Sea floor spreading. 247, 7– 11. southern Arabia is the sister lineage to a clade of samples from Gray, J.E., 1865. Revision of the genera and species of Chamaeleonidae, with the the eastern Mediterranean (Cyprus and Turkey), and not closely re- description of some new species. Proc. Zool. Soc. London 1864, 465–479. lated to the population in the western Mediterranean region. This Günther, A., 1881. Descriptions of the amphisbaenians and ophidians collected by Prof. I. Baley Balfour on the island of Socotra. Proc. Zool. Soc. London 40, 461– phylogenetic pattern suggests dispersal of C. chamaeleon from the 463. eastern Mediterranean to southern Arabia. Samples from Turkey Gutell, R.R., Fox, G.E., 1988. Compilation of large subunit rRNA sequences presented and Cyprus are nearly identical and may have been transported in a structural format. Nucleic Acids Res. 16s, r175–r269. by human activity. Hickson, R.E., Simon, C., Cooper, A., Spicer, G.S., Sullivan, J., Penny, D., 1996a. Conserved sequence motifs, alignment, and secondary structure for the third Although our phylogenetic results are sound, we point to a dis- domain of 12S rRNA. Mol. Biol. Evol. 13, 150–169. crepancy in our estimate in timing of divergences. Macey et al. Hickson, R.E., Simon, C., Cooper, A., Spicer, G.S., Sullivan, J., Penny, D., 1996b. (1998) provided a linear model for sequence evolution in the region Conserved sequence motifs, alignment, and secondary structure for the third domain of animal 12S rRNA. Mol. Biol. Evol. 13, 150–169. between nad1 and cox1, with 1.3% pairwise change across a million Hugall, A.F., Lee, M.S.Y., 2004. Molecular claims of Gondwanan age for Australian years for agamid lizards. Applying this to our data we reach very agamid lizards are untenable. Mol. Biol. Evol. 21, 2102–2110. young ages that are interpreted as an artifact of mangrove swamps Kumazawa, Y., Nishida, M., 1993. Sequence evolution of mitochondrial transfer RNA genes and deep-branch animal phylogenetics. J. Mol. Evol. 37, 380– being in close proximity during plate tectonic breakup; hence pro- 398. viding an avenue for dispersal and gene flow. For example, Socotra Kumazawa, Y., 2007. Mitochondrial genomes from major lizard families suggest chameleons (13%) would have diverged from others around 10 mil- their phylogenetic relationships and ancient radiations. Gene 388, 19–26. Macey, J.R., Verma, A., 1997. Homology in phylogenetic analysis: Alignment of lion years ago, although we expect sequence saturation at this transfer RNA genes and the phylogenetic position of . Mol. Phylogenet. point and this could correspond to 15 million years. Regardless, it Evol. 7, 272–279. is unlikely that these sequence divergences and phyletic patterns Macey, J.R., Larson, A., Ananjeva, N.B., Fang, Z., Papenfuss, T.J., 1997a. Two novel gene orders and the role of light-strand replication in rearrangement of the are unrelated to tectonic plate movements as outlined. vertebrate mitochondrial genome. Mol. Biol. Evol. 14, 91–104. The formation of the Red Sea and Gulf of Aden with their ende- Macey, J.R., Larson, A., Ananjeva, N.B., Papenfuss, T.J., 1997b. Evolutionary shifts in mic faunas may provide a missing key to revealing the interactions three major structural features of the mitochondrial genome among Iguanian of major Gondwanan land masses as they moved across the Tethys lizards. J. Mol. Evol. 44, 660–674. Macey, J.R., Schulte II, J.A., Ananjeva, N.B., Larson, A., Rastegar-Pouyani, N., Sea toward Laurasia. Socotra Island, southern Arabia, and the Horn Schammokov, S., Papenfuss, T.J., 1998. Phylogenetic relationships among of Africa remain one of the least explored regions of the world. Our agamid lizards of the Laudakia caucasia complex: Testing hypotheses of reconstructed history of the chameleon fauna endemic to these fragmentation and an area cladogram for the Iranian Plateau. Mol. Phylogenetic. Evol. 10, 118–131. areas serves as a set of a priori hypotheses to be tested in other Macey, J.R., Schulte II, J.A., Kami, H.G., Ananjeva, N.B., Larson, A., Papenfuss, T.J., co-distributed taxa. Genomic sequencing of numerous biotic 2000a. Testing hypotheses of vicariance in the agamid lizard Laudakia caucasia groups in the region will provide a diverse genetic resource for from mountain ranges on the northern Iranian Plateau. Molecular Phylogenetics and Evolution 14, 479–483. testing historical biogeographic relationships in the Middle East. Macey, J.R., Schulte II, JA., Larson, A., Ananjeva, N.B., Wang, Y., Pethiyagoda, Y., It has not escaped our attention that this may be dynamic, involv- Rastegar-Pouyani, N., Papenfuss, T.J., 2000b. Evaluating trans-Tethys ing plate tectonics that includes dispersal. migration: An example using acrodont lizard phylogenetics. Syst. Biol. 49, 233–256. Macey, J.R., Papenfuss, T.J., Kuehl, J.V., Fourcade, H.M., Boore, J.L., 2004. Phylogenetic Acknowledgments relationships among amphisbaenian based on complete mitochondrial genomic sequences. Mol. Phylogenetic. Evol. 33, 22–31. Macey, J.R., Schulte II, J.A., Fong, J.J., Das, I., Papenfuss, T.J., 2006. The complete Eddy Rubin, James Bristow and Jeffery L. Boore provided sup- mitochondrial genome of an agamid lizard from the Afro–Asian subfamily port. Karen Klitz prepared Fig. 1. Craig L. Hassapakis assisted with agaminae and the phylogenetic position of Bufoniceps and Xenagama. Mol. GenBank submissions. H. Mathew Fourcade assisted with labora- Phylogenetic. Evol. 39, 881–886. tory work. The American Institute for Yemeni Studies facilitated Maddison, W.P., Maddison, D.R., 2001. MacClade, Analysis of Phylogeny and Character Evolution, Version 4.03. Sinauer, Sunderland, MA. fieldwork in Yemen including Socotra Island. This work was per- Raxworthy, C.J., Forstner, M.R.J., Nussbaum, R.A., 2002. Chameleon radiation by formed under the auspices of the US Department of Energy’s Office . Nature 415, 784–787. of Science, Biological and Environmental Research Program and by Schulte II, J.A., Melville, J., Larson, A., 2003. Molecular phylogenetic evidence for ancient divergence of lizard taxa on either side of Wallace’s Line. Proc. Roy. Soc. the University of California, LLNL under contract No. W-7405-Eng- London B 270, 597–603. 48, LBNL No. DE-AC03-76SF00098 and LANL No. W-7405-ENG-36. Swofford, D.L., 2001. PAUP* Phylogenetic Analysis Using Parsimony (* and other methods), Beta Version 4.0b8. Sinauer, Sunderland, MA. Townsend, A., Larson, A., 2002. Molecular phylogenetics and mitochondrial genomic Appendix A. Supplementary data evolution in the Chamaeleonidae (Reptilia, ). Mol. Phylogenetic. Evol. 23, 22–36. Supplementary data associated with this article can be found, in Van de Peer, Y., Van den Broeck, I., de Rijk, P., de Wachter, R., 1994. Database on the structure of small ribosomal subunit RNA. Nucleic Acids Res. 22, 3488– the online version, at doi:10.1016/j.ympev.2008.08.024. 3494.