COBISS: 1.01 HOw LONG DOES EVOLUTION OF THE TROGLOMORPHIC FORM TAKE? ESTIMATING DIVERGENCE TIMES IN ASTyANAx MExICANUS KAKO DOLGO TRAJA EVOLUCIJA TROGLOMORFNIH OBLIK? OCENJEVANJE DIVERGENČNIH ČASOV PRI ASTyANAx MExICANUS Megan L. PORTER1, Katharina DITTMAR2 & Marcos PéREZ-LOSADA3 Abstract UDC 551.44:597 Izvleček 591.542 Megan L. Porter, Katharina Dittmar & Marcos Pérez-Losada: How long does evolution of the troglomorphic form take? Esti-mating divergence times in Astyanax mexicanus Features including colonization routes (stream capture) and the existence of both epigean and cave-adapted hypogean popula-tions make Astyanax mexicanus an attractive system for investi-gating the subterranean evolutionary time necessary for acqui-sition of the troglomorphic form. Using published sequences, we have estimated divergence times for A. mexicanus using: 1) two diferent population-level mitochondrial datasets (cyto-chrome b and NADH dehydrogenase 2) with both strict and relaxed molecular clock methods, and 2) broad phylogenetic approaches combining fossil calibrations and with four nuclear (recombination activating gene, seven in absentia, forkhead, and ?-tropomyosin) and two mitochondrial (16S rDNA and cytochrome b) genes. Using these datasets, we have estimated divergence times for three events in the evolutionary history of troglomorphic A. mexicanus populations. First, divergence among cave haplotypes occurred in the Pleistocene, possibly correlating with fuctuating water levels allowing the coloni-zation and subsequent isolation of new subterranean habitats. Second, in one lineage, A. mexicanus cave populations expe-rienced introgressive hybridization events with recent surface populations (0.26-2.0 Ma), possibly also correlated with Pleis-tocene events. Finally, using divergence times from surface populations in the lineage without evidence of introgression as an estimate, the acquisition of the troglomorphic form in A. mexicanus is younger than 2.2 (fossil calibration estimates) – 5.2 (cytb estimate) Ma (Pliocene). Key words: Astyanax mexicanus, divergence time, troglomor-phy, subterranean, evolution. UDK 551.44:597 591.542 Megan L. Porter, Katharina Dittmar & Marcos Pérez-Losada: Kako dolgo traja evolucija troglomorfnih oblik? Ocenjevanje divergenčnih časov pri Astyanax mexicanus Značilnosti, ki vključujejo tudi kolonizacijske poti in obstoj tako epigejičnih kot hipogejičnih populacij vrste Astyanax mexica-nus, ji omogočajo, da predstavlja zanimiv sistem za proučevanje evolucije in časa, potrebnega za razvoj podzemeljskih troglo-morfnih oblik. Za A. mexicanus smo na podlagi že objavljenih sekvenc ocenili divergenčni čas ob uporabi: 1) dveh različnih populacijskih mitohondrialnih podatkovnih baz (citokrom b in NADH dehidrogenaze 2), obe z natančno in sproščeno metodo molekularne ure, in 2) razširjenega flogenetskega pristopa v kombinaciji s fosilno kalibracijo ter štirimi jedrnimi geni (rekombinacijski aktivacijski gen, »forkhead kontrolni gen« in ?-tropomiozin) in dvema mitohondrialnima genoma (16S rDNA in citokrom b). Ob uporabi navedenih podatkovnih baz smo ocenili divergenčni čas za tri dogodke v zgodovini razvoja troglomorfnih populacij A. mexicanus. Prvič, razhajanje med podzemeljskimi haplotipi se je zgodilo v Pleistocenu, verjetno v odvisnosti od nihanja vode, ki je omogočilo kolonizacijo in posledično izolacijo v novih podzemeljskih habitatih. Drugič, verjetno je v povezavi s pleistocenskimi dogodki pri eni liniji podzemeljskih populacij A. mexicanus prišlo do introgresivne hibridizacije s takratnimi površinskimi populacijami (0.26-2.0 Ma). Z uporabo divergenčnega časa površinskih populacij tistih linij, ki ne kažejo introgresije ocenjujemo, da je troglomorfna oblika A. mexicanus mlajša od 2,2 (ocene fosilne kalibracije) do 5,2 milijona let (cytb ocena) (Pliocen). Ključne besede: Astyanax mexicanus, divergenčni čas, troglo-morfzem, podzemlje, speleobiologija, evolucija. 1 Dept. of Biological Sciences, University of Maryland Baltimore County, Baltimore, MD, USA; e-mail: porter@umbc.edu 2 Dept. of Molecular Biology, University of wyoming, Laramie, wy, USA 3 GENOMA LLC, 50E woodland Hills, Provo, UT 84653-2052, USA Received/Prejeto: 06.12.2006 TIME in KARST, POSTOJNA 2007, 173–182 MEGAN L. PORTER, KATHARINA DITTMAR & MARCOS PéREZ-LOSADA INTRODUCTION Understanding the evolution of the cave form has fasci-nated biologists interested in subterranean faunas since Darwin. Termed ‘troglomorphy’, the suite of progressive and regressive characters associated with cavernicolous animals can be observed in the worldwide convergence of form found in the cave environment, exhibited in similar structural, functional, and behavioral changes across diverse taxonomic groups. Much of the debate over troglo-morphy has centered on the evolutionary mechanisms responsible for character regression, generally argued to be either neutral mutation or natural selection. Several studies, (Gammarus minus - Culver et al., 1995; Astyanax mexicanus – Jefery, 2005) have shown eye degeneration is the result of selection, and, in the case of A. mexica-nus, is caused by the pleiotropic efects of natural selec-tion for constructive traits. Another, less studied, aspect of understanding troglomorphy is the evolutionary time required to gain the cave form. Because it is generally dif-fcult to pinpoint the time of subterranean colonization and isolation from surface ancestors, few troglomorphic species ofer the opportunity for quantitative estimates of the evolutionary time spent in the subterranean realm. Terefore, the time of cave adaptation is thought of in relative terms, where the degree of eye and pigment re-duction indicates the period of cavernicolous evolution and therefore the relative phylogenetic age of each spe-cies (Aden, 2005). In evolutionary studies of cave adaptation, Asty-anax mexicanus has become a model system (Jefery, 2001). Te advantageous features of A. mexicanus as a model system include the existence of both surface and troglomorphic cavefsh populations, with several cave fsh populations having evolved constructive and regressive changes independently (Jefery, 2001). Furthermore, since the discovery of the species in 1936 (Hubbs & Innes, 1936), there has been an extensive amount of research devoted to characterizing developmental, phylogenetic, taxonomic, and biogeographic aspects of the species (Jef-fery, 2001; Mitchell et al., 1977; wiley & Mitchell, 1971;). In terms of being a model system for understanding the evolution of the troglomorphic form, A. mexicanus has at least one additional favorable attribute. Te primary mode of A. mexicanus subterranean colonization is via stream capture, with most of the captured surface drain-ages no longer supporting epigean populations (Mitchell et al., 1977). Tese captures provide discrete coloniza-tion events correlated with divergence time from surface populations and therefore with the time of subterranean evolution. Molecular studies that have looked at A. mexicanus phylogeography indicate that at least two independent invasions of surface Astyanax have occurred (Dowling et al., 2002a; Strecker et al., 2003, 2004). Tese two distinct A. mexicanus genetic lineages consist of cave fsh from La Cueva Chica, La Cueva de El Pachón, El Sótano de yerbaniz, El Sótano de Molino, El Sótano de Pichijumo, and La Cueva del Río Subterráneo (lineage A) and from La Cueva de los Sabinos, El Sótano de la Tinaja, La Cueva de la Curva, and El Sótano de Las Piedras (Lineage B) with diferent evolutionary histories - Lineage A clus-ters with closely related epigean populations while lin-eage B has no closely related epigean counterparts. Te close association of Lineage A to epigean populations (as estimated by mitochondrial markers) is thought to be the result of either recent subterranean colonization or refect recent introgressive hybridization with surface populations, while lineage B is considered to be a more ancient colonization event from surface populations that are extinct in the region (Dowling et al., 2002a; Strecker et al., 2004). Although the evolutionary histories of dif-ferent hypogean A. mexicanus populations are complex, the two lineages ofer the unique opportunity to estimate the divergence time required for the evolution of the tro-glomorphic form based on discrete times of colonization and the previous molecular studies of their phylogeogra-phy. At least one other study has estimated lineage ages in A. mexicanus populations; however, this study was based on a single gene molecular clock estimate and did not specifcally estimate the divergence times of the cave populations (Strecker et al., 2003). Here we use three dif-ferent sets of publicly available sequence data and known fossil calibrations and apply multiple phylogenetic ap-proaches to estimate the age of cave colonization and stream capture events, and to provide an estimate of the time necessary to acquire the troglomorphic form in A. mexicanus. 174 TIME in KARST – 2007 HOw LONG DOES EVOLUTION OF THE TROGLOMORPHIC FORM TAKE? ESTIMATING DIVERGENCE TIMES ... METHODS Sequence Data Data were acquired from Genbank (http://www.ncbi. nlm.nih.gov/) from previously published studies of A. mexicanus and characiform fshes (Tab. 1). Tese studies provided three diferent datasets, consisting of: 1) population-level haplotype datasets for the mitochondrial cy-tochrome b (cytb; Strecker et al., 2004) and NADH dehy-drogenase 2 (ND2; Dowling et al., 2002a) genes, and 2) a species-level dataset of four nuclear (recombination ac-tivating gene – RAG2; seven in absentia – sina; forkhead – fh; and ?-tropomyosin - trop) and two mitochondrial genes (16S rDNA and cytb) from representatives within the Otophysi (Calcagnotto et al., 2005). Divergence times from all three data sets were estimated and compared. Species-level Phylogenetic Analyses Te species-level dataset included selected Otophysi, Characiformes, and Characidae sequences (see Tab. 1), and was analyzed using Anotophysi species as outgroups. Representative A. mexicanus cytb haplotype sequences from the Strecker et al., (2004) study were included in the dataset of characiform species to estimate diver-gence times based on fossil calibrations for comparison with population-based estimates utilizing substitution rates. Alignments of protein-coding regions were trivial and were accomplished using amino acid translations. Sequences of the trop gene spanned an intron, which was removed due to signifcant length variation (70-836 bp) leading to ambiguous alignments. Te alignment of the 16s rDNA gene was generated using the E-INS-i accuracy-oriented strategy of MAFFT v.5 (Katoh et al., 2005). All of the individually aligned genes were then concatenated to form a single dataset consisting 3770bp in length. Te concatenated dataset was analyzed with PAUP* 4.0b10 (Swoford, 2000) using maximum parsi-mony and implementing the parsimony ratchet method (Nixon, 1999) using a batch fle generated by PAUPRat with the default parameters for 5000 replicates (Sikes & Lewis, 2001). Divergence time estimation Population analysis. Dates of divergence were inferred for A. mexicanus lineage A and B cave fsh populations using the cytb and ND2 datasets with BEASTv1.4 (Drummond & Rambaut, 2003). Because the cytb and ND2 haplotype datasets were generated from diferent studies, they can-not be combined. Terefore, each dataset was used to independently estimate the divergence times of the A. mexicanus cave-adapted haplotype sequences. Each dataset was analyzed using both strict and relaxed clock models (Drummond et al., 2006) tested under constant and skyline models of population growth. As part of BEAST divergence time estimation, either a calibration point (fossil or geologic) or a gene-specifc substitution rate is required. Because there are no geologic dates cor-responding to A. mexicanus populations invading sub-terranean systems, substitution rates were used. For each gene, the range of substitution rates calculated for other freshwater fsh were used. For cytb, mean substitution rates ranged from 0.005 to 0.017 substitutions/site/mil-lion year (my) (Bermingham et al., 1997; Burridge et al., 2006; Dowling et al., 2002b; Perdices & Doadrio, 2001; Sivasundar et al., 2001; Zardoya & Doadrio, 1999) and for ND2 mean substitution rates ranged from 0.011 to 0.026 substitutions/site/my (Near et al., 2003; Mateos, 2005). Tese independent rates were used to calibrate the rate of evolution of our datasets by either fxing the rate to the lowest and highest value estimated for each gene or using strong prior distributions on the substitution rates. Two independent MCMC analyses 2x107 steps long were performed sampling every 2,000th generation, with a burn-in of 2x106 generations. All the Bayesian MCMC output generated by BEAST was analyzed in Tracer v1.3 (Drummond & Rambaut, 2003). Likelihood-based AhRS method. we used the likeli-hood heuristic rate-smoothing algorithm of (yang, 2004) as implemented in PAML3.14 (yang, 2001). Sequence data were analyzed using the F84+? model. Branches at each locus were classifed into four rate groups accord-ing to their estimated rates. Te oldest known fossil rep-resentatives of major lineages within the Ostariophysi are well established in recent literature (see Briggs, 2005 and references therein), and have been used in recent studies estimating molecular-based divergence times of Otocephalan clades (Peng et al., 2006). Tese fossil representatives were used as calibration points for the AHRS divergence time analysis (Fig. 1, Tab. 2,). Fossil calibrations were accommodated as fxed ages and mapped to the basal node of the clade of interest. Given that most fossils are dated to an age range, the minimum and maximum ages of each fossil were used for diver-gence time estimations under separate analyses. Fossil dates were determined using the 1999 GSA Geologic Time Scale. TIME in KARST – 2007 175 MEGAN L. PORTER, KATHARINA DITTMAR & MARCOS PéREZ-LOSADA tab. 1: taxonomy, gene data, and Genbank accession numbers for sequences used in Characiformes phylogeny reconstruction. Abbreviations of mitochondrial gene sequences: 16S = 16S rdNA, cytb = cytochrome b; abbreviations for nuclear gene sequences: fh = forkhead, RAG2 = recombination activatin g gene, sina = seven in absentia, trop = ?-tropo 16S cytb fkh myosin. RAG2 sina trop Anotophysi (outgroup) Chanidae Chanos chanos NC004693 NC004693 — — — — Gonorynchidae Gonorynchus greyi NC004702 NC004702 — — — — Kneriidae Cromeria nilotica NC007881 NC007881 — — — — Parakneria cameronensis NC007891 NC007891 — — — — Otophysi (ingroup) CHARACIFORMES Anostomidae Leporinus sp. AY788044 AY791416 AY817370 AY804095 AY790102 AY817252 Chilodontidae Chilodus punctatus AY787997 — AY817325 — AY790056 AY817215 Prochilodontidae Prochilodus nigricans AY788075 AY791437 AY817400 AY804120 AY790133 AY817278 Hemiodontidae Hemiodus gracilis AY788027 AY791405 AY817353 AY804084 AY790086 AY817240 Parodontidae Parodon sp. AY788065 AY791427 AY817390 AY804110 AY790123 AY817269 Serrasalmidae Colossoma macropomum AY788000 AY791386 AY817328 AY804061 AY790059 AY817218 Cynodontidae Hydrolycus pectoralis AY788033 — AY817359 AY804088 AY790091 AY817244 Characidae Acestrorhynchus sp. AY787956 AY791353 AY817288 AY804026 AY790014 AY817181 Aphyocheirodon sp. AY787966 AY791363 AY817298 AY804031 AY790025 — Astyanacinus sp.1 AY787969 AY791365 AY817301 AY804033 AY790028 AY817190 Astyanacinus sp.2 AY787987 — AY817317 AY804051 AY790046 AY817209 Astyanax bimaculatus AY787955 — AY817287 AY804025 AY790013 AY817180 Astyanax mexicanus (Brazil) — AY177206 — — — — Astyanax mexicanus (haplotype AB) -- AY639041 -- - - -- Astyanax mexicanus (haplotype AL) -- AY639051 -- - - -- Astyanax mexicanus (haplotype EA) -- AY639075 -- - - -- Astyanax mexicanus (haplotype FA) -- AY639084 -- - - -- Astyanax mexicanus (haplotype GA) -- AY639089 -- - - -- Astyanax mexicanus (haplotype GB) -- AY639090 -- - - -- Astyanax scabripinis AY787967 — AY817299 — AY790026 AY817188 Brycon hilarii AY787976 AY791370 AY817307 AY804040 AY790035 AY817198 Bryconamericus diaphanus AY787984 AY791375 AY817314 AY804048 AY790043 AY817206 Bryconops sp. AY787985 AY791376 AY817315 AY804049 AY790044 AY817207 Chalceus erythrurus AY787990 AY791379 AY817320 AY804053 AY790049 AY817211 Chalceus macrolepidotus AY787999 AY791385 AY817327 AY804060 AY790058 AY817217 Cheirodon sp. AY787995 AY791382 AY817324 AY804057 AY790054 — Cheirodontops sp. AY787996 AY791383 — AY804058 AY790055 — Creagrutus sp. AY788001 — — AY804062 AY790060 AY817219 Exodon paradoxus AY788013 AY791397 AY817340 AY804072 AY790072 AY817227 Gephyrocharax sp. AY788014 AY791398 AY817341 AY804073 AY790073 AY817228 Hemibrycon beni AY788020 AY791402 AY817346 AY804079 AY790079 AY817234 Hemigrammus bleheri AY788017 — AY817343 AY804076 AY790076 AY817231 176 TIME in KARST – 2007 HOw LONG DOES EVOLUTION OF THE TROGLOMORPHIC FORM TAKE? ESTIMATING DIVERGENCE TIMES ... 16S cytb fkh RAG2 sina trop Hemigrammus erythrozonus AY788023 — AY817349 AY804081 AY790082 AY817236 Hemigrammus rodwayi AY788034 — AY817360 AY804089 AY790092 AY817245 Hyphessobrycon eques AY788022 — AY817348 AY804080 AY790081 AY817235 Inpaichthys kerri AY788039 — AY817365 AY804093 AY790097 AY817248 Knodus sp. AY788041 AY791414 AY817367 AY804094 AY790099 AY817249 Moenkhausia sanctaphilomenae AY788054 — — AY804104 AY790112 AY817261 Mimagoniates lateralis AY788051 AY791420 AY817377 AY804101 AY790109 AY817259 Prodontocharax sp. AY788064 AY791426 AY817389 AY804109 AY790122 — Roeboides sp. AY787994 AY791381 AY817323 AY804056 AY790053 AY817214 Salminus maxillosus AY788080 AY791438 AY817405 AY804124 AY790137 AY817282 Triportheus angulatus AY788082 — AY817407 AY804125 AY790139 AY817283 Ctenolucidae Ctenolucius hujeta AY787998 AY791384 AY817326 AY804059 AY790057 AY817216 Lebiasinidae Nannostomus beckfordi AY788059 — AY817384 — AY790117 AY817265 Crenuchidae Characidium fasciatum AY787992 AY791380 AY817322 AY804055 AY790051 AY817213 Erythrinidae Hoplias sp. AY788031 AY791409 AY817357 AY804087 AY790090 AY817242 Alestidae Arnoldichthys spilopterus AY787968 AY791364 AY817300 AY804032 AY790027 AY817189 Brycinus nurse AY787970 AY791366 AY817302 AY804034 AY790029 AY817191 Phenacogrammus aurantiacus AY788066 AY791428 AY817391 AY804111 AY790124 AY817270 Hepsetidae Hepsetus odoe AY788030 AY791408 AY817356 AY804086 AY790089 AY817241 Citharinidae Citharinus citharus AY787989 AY791378 AY817319 — AY790048 — Distichodontidae Distichodus sexfasciatus AY788012 AY791396 AY817339 AY804071 AY790071 AY817226 Neolebias trilineatus AY788063 AY791425 AY817388 AY804108 AY790121 AY817268 CYPRINIFORMES Cobitidae Misgurnus sp. AY788053 — AY817379 AY804103 AY790111 — Cyprinidae Danio rerio AY788011 — AY817338 AY804070 AY790070 AY817225 Labeo sorex AY788043 AY791415 AY817369 — AY790101 AY817251 Gyrinocheilidae Gyrinocheilus sp. AY788015 AY791399 — AY804074 AY790074 AY817229 SILURIFORMES Callichthyidae Corydoras rabauti NC004698 NC004698 — — — — Loricariidae Ancistrus sp. AY787958 AY791354 AY817290 — AY790016 AY817183 Bagridae Chrysichthys sp. AY787957 AY791355 — — AY790017 AY817193 Heptapteridae Pimelodella sp. AY787953 AY791351 AY817285 — AY790011 AY817178 Ictaluridae Ictalurus punctatus AY788040 AY791413 AY817366 — AY790098 — TIME in KARST – 2007 177 MEGAN L. PORTER, KATHARINA DITTMAR & MARCOS PéREZ-LOSADA Fig. 1: Characiform divergence time chronogram estimated using a representative topology chosen from the set of 867 most parsimonious trees. White branches indicate branches where less than 75% of the most parsimonious trees were topologically congruent. Te grey box indicates the clade of Astyanax mexicanus sequences. Fossil calibration nodes are numbered and correspond to tab. 2. Te major geologic periods are mapped onto the phylogeny. 178 TIME in KARST – 2007 HOw LONG DOES EVOLUTION OF THE TROGLOMORPHIC FORM TAKE? ESTIMATING DIVERGENCE TIMES ... tab. 2: taxonomy and ages of fossils used as calibrations for divergence time estimation. Node # refers to Fig. 1. Taxonomy Otophysi Characiformes Cypriniformes Catostomidae Siluriformes Corydoras Reference Geologic age (MYA) Gayet, 1982 Cavender, 1986 Gayet & Meunier, 2003 Cockerell, 1925 Node # Late Cretaceous (65-99) Paleocene (54.8-65) late Campanian-early Maastrichtian (68.2-77.4) Late Palaeocene (61-65) 4 3 2 RESULTS Population-level divergence time estimations. Estimates of the mean divergence times were not signifcantly dif-ferent between strict and relaxed clock and population growth models and calibration methods of the substitution rate, but confdence intervals under the fxed substitution rate approach were narrower, as expected. Hence only the time estimates under the strict clock model, con-stant population size and minimum and maximum mean substitution rates for both genes are provided. Compar-ing the cytb and ND2 estimates of divergence times for the A. mexicanus A and B lineages show several features. First, the estimated ranges of divergence for cave hap-lotypes within each lineage were similar between genes (cytb and ND2) and lineages (A and B), placing the di-vergence among hypogean populations between 0.141-0.885 Ma for lineage A, and 0.084-0.575 Ma for lineage B (Tab. 3). when comparing the estimates among genes within a lineage, however, the divergence times of hypo-gean and epigean haplotypes are diferent, with cytb esti-mates providing generally older estimates. Species-level divergence time estimation. Using the maximum parsimony ratchet, the selected Characidae, Characiform, and Otophysi sequences generated 867 trees of score 11758. Te 50% majority rule consensus of these trees was similar to the published research that generated the data (Calcagnotto et al., 2005). Because a fully resolved tree with branch lengths is required for AHRS divergence time estimation and because very few branches in the consensus tree collapsed (e.g. were in confict), a random tree from the set of 867 was used (Fig. 1). Te A. mexicanus sequences included in the analysis clustered with other Characidae species, although were not monophyletic with other Astyanax species (A. bi-maculatus and A. scabripinnis). Te divergence time estimates for the representative A. mexicanus cave fsh populations generated using this phylogeny with Oto-physi fossil calibrations agreed well with the estimates of hypogean haplotype divergence from cytb and ND2 us-ing substitution rates (Tab. 3). However, the estimates of cave versus surface population divergence times based on fossil calibrations were in better agreement with ND2 than with cytb estimates. Tis is particularly interesting, as the only gene included in this dataset for A. mexica-nus was cytb. tab. 3: Comparison of divergence time estimates using substitution rates and molecular clock methods for cytochrome b (cytb) and NAdh dehydrogenase 2 (Nd2) mitochondrial genes, and for molecular methods incorporating fossil dates as calibrations. Lineage A cave cave vs. surface Lineage B cave cave vs. surface Lineage A vs. Lineage B Substitution Rates Cytb ND2 Min – Max (Ma) Min – Max (Ma) Fossil Calibration Min – Max (Ma) 0.261 – 0.885 0.588 - 2.00 0.169 – 0.575 1.524 – 5.181 1.741 – 5.922 0.141 - 0.331 0.256 – 0.599 0.084 - 0.196 0.877 – 2.055 1.053 – 2.472 0.2-0.3 0.4-0.5 0.1-0.1 1.7-2.2 1.7-2.2 TIME in KARST – 2007 179 1 MEGAN L. PORTER, KATHARINA DITTMAR & MARCOS PéREZ-LOSADA DISCUSSION Previous molecular studies of A. mexicanus phylogeog-raphy indicate that at least two independent invasions of surface Astyanax have occurred (Dowling et al., 2002a; Strecker et al., 2003, 2004). Our estimates of divergence time from two diferent methods and three diferent datasets are in general agreement about the divergence times among the cave haplotypes in each lineage (Tab. 3). Tese estimates place cave haplotype divergence times in the Pleistocene, when it is suggested that climatic cool-ing of surface waters led to the extinction of Astyanax in North America (Strecker et al., 2004). In particular, our data show an interesting pattern for lineage B haplotypes, which are proposed to be the older of the two lineages. Te recent divergence times estimated for lineage B hap-lotypes (0.084-0.575 Ma) supports the hypothesis that afer the initial colonization event, subterranean routes of colonization were associated with fuctuating ground-water levels in the Pleistocene (Strecker et al., 2004). Te fact that estimated times of within lineage divergence were similar also suggests that the divergence of subter-ranean haplotypes in both lineages were infuenced by the same processes. In order to determine the evolutionary age of the subterranean lineage, and therefore estimate the time re-quired for evolution of the troglomorphic form, the di-vergence of the hypogean haplotypes from epigean popu-lations is needed. However, the estimates from our three datasets did not agree, with cytb molecular clock meth-ods estimating older divergence times than either ND2 or fossil calibrated estimates. Some of the discrepancy is due to the fact that diferent sets of surface popula-tions were sampled in each study (Dowling et al., 2002a; Strecker et al., 2004). For example, the most closely relat-ed surface population in the cytb study were from Belize (Strecker et al., 2004) while there were no closely related surface populations to lineage B haplotypes in the ND2 study (Dowling et al., 2002a). However, this makes the older cytb estimates even more notable because lineage B haplotypes have no evidence of introgressive hybridiza-tion with surface populations. If we consider just lineage B hypogean divergence from surface ancestors as an es-timate of subterranean evolution, the estimated time for acquisition of the troglomorphic form is 0.877-2.055 Ma (quaternary – Tertiary boundary) based on ND2 and fossil calibrations, while it is 1.524-5.181 Ma (Pliocene) based on cytb. Although the estimates of divergence times among the three diferent datasets did not agree, comparison of estimates between the lineages show that lineage A diverged from surface ancestors more recently than lineage B (Tab. 3). Tis more recent divergence from epigean populations is congruent with previous hypoth-eses, that either lineage A populations represent a more recent subterranean invasion, or that they are an older invasion masked by more recent mitochondrial intro-gressive hybridization with surface forms (Dowling et al., 2002a). In the few studies that have looked at other markers (allozymes, microsatellites, and RAPDs), it has been suggested that at least Chica and Pachón popula-tions are the result of surface introgression (Avise & Se-lander, 1972; Espinasa & Borowsky, 2001; Strecker et al., 2003). Furthermore, based on the degree of variability in troglomorphic features of each lineage A population, it has been suggested that diferent populations represent diferent degrees and patterns of surface introgression. In order to more accurately determine both the patterns of introgression in the lineage A populations, as well as the underlying relationships of the cave populations to each other in order to estimate subterranean evolution-ary times, studies investigating more types of markers are needed. Previous research of A. mexicanus populations throughout Mexico (including cavefsh lineages A and B) estimated haplotype divergences to range from 1.8 – 4.5 Ma (Strecker et al., 2004). Our estimates suggest that di-vergence times among cave haplotypes and between lin-eage A cave and epigean haplotypes are much younger than this; however, hypogean divergences from surface ancestors in lineage B are concordant with these older dates. Te evolutionary history of cave adaptation in A. mexicanus is complex. Based on mitochondrial molecu-lar clock estimates, our estimates of divergence times are congruent with previous hypotheses by showing lineage B to be a phylogenetically older subterranean lineage, with more recent divergence among subterranean systems. However, this study also provides quantitative dates for these events. Lineage A populations are estimated to be younger; however, these dates only represent mito-chondrial lineages. Several of the populations in lineage A have been shown to be introgressed with surface forms (Chica, Pachón, and Subterraneo). To our knowledge, the hypothesis of surface introgression has not been investi-gated in the remaining lineage A populations (Molino, Pichijumo, and yerbaniz). Understanding the patterns of introgression in all of the lineage A populations, and estimating the actual subterranean evolutionary time, re-quires investigating additional nuclear markers. 180 TIME in KARST – 2007 HOw LONG DOES EVOLUTION OF THE TROGLOMORPHIC FORM TAKE? ESTIMATING DIVERGENCE TIMES ... CONCLUSIONS Features including colonization routes (stream capture) and the existence of both epigean and cave-adapted hy-pogean populations make A. mexicanus an attractive system for investigating the subterranean evolution-ary time necessary for acquisition of the troglomorphic form. If it is possible to estimate the divergence time of closely related cave versus surface populations, we can estimate the age of subterranean occupancy. Tis same divergence time also has relevancy to geologic proc-esses in the karst system by providing a rough estimate of the age of subterranean stream capture in particular regions. Based on published sequence data, we have esti-mated divergence times for three events in the evolution-ary history of troglomorphic A. mexicanus populations. First, divergence times among cave haplotypes in both lineages occurred in the Pleistocene, possibly correlating with fuctuating water levels allowing the colonization, and subsequent isolation of, new subterranean habitats. Second, in lineage A, A. mexicanus cave populations ex-perienced introgressive hybridization events with surface populations recently. Finally, using divergence times of lineage B from surface populations as an estimate, the acquisition of the troglomorphic form in A. mexicanus is younger than 2.2 (fossil calibration) – 5.2 (cytb) Ma (Pliocene). Given that there are at least 30 caves known to contain populations of A. mexicanus (Espinasa et al., 2001; Mitchell et al., 1977), the number of independent invasions and instances of introgressive hybridization may be even higher than currently understood. In order to fully understand the number of independent invasions, the history of introgression with surface populations, and the divergence times of cave and surface populations, a broader survey of cave fsh populations and of both nu-clear and mitochondrial markers is needed. LITERATURE CITED Aden, E., 2005: Adaptation to darkness. In Culver, D.C., & white, w.B., (eds.), Encyclopedia of Caves, Elsevier Academic Press, pp.1-3. Avise, J.C., & R.K. Selander., 1972: Genetics of cave-dwelling fshes of the genus Astyanax. –Evolution, 26, 1-19. Bermingham, E., McCaferty, S.S., & A.P. Martin., 1999: Fish biogeography and molecular clocks: perspec-tives from the Panamanian isthmus. In Kocher, T.D., & Stepien, C.A., (eds.), Molecular systematics of fshes. Academic Press, San Diego, CA. pp.113-128. Briggs, J.C., 2005: Te biogeography of otophysan fshes (Ostariophysi: Otophysi): a new appraisal. -Journal of Biogeography, 32, 287-294. Burridge, C.P., Craw, D., & J.M. waters., 2006: River cap-ture, range expansion, and cladogenesis: Te genetic signature of freshwater vicariance. –Evolution, 60, 1038-1049. Calcagnotto, D., Schaefer, S.A., & R. DeSalle., 2005: Re-lationships among characiform fshes inferred from analysis of nuclear and mitochondrial gene se-quences. -Molecular and Phylogenetics and Evolution, 36, 135-153. Cavender, T.M., 1986: Review of the fossil history of North American freshwater fshes. In Hocutt, C.H., & wiley, E.O. (eds.), Te zoogeography of North American freshwater fshes. John wiley, New york, pp.699-724. Cockerell, T.D., 1925: A fossil fsh of the family Callich-thyidae. –Science, 62, 317-322. Culver, D.C., Kane, T.C., & D.w. Fong., 1995: Adaptation and natural selection in caves. Harvard University Press, Cambridge, 223p. Dowling, T.E., Martasian, D.P., & w.R. Jefery., 2002a: Evidence for multiple genetic forms with similar eyeless phenotypes in the blind cavefsh, Astyanax mexicanus. -Molecular Biology and Evolution, 19, 446-455. Dowling, T.E., Tibbets, C.A., Minckley, w.L., & G.R. Smith, 2002b: Evolutionary relationships of the plagopterins (Teleostei: Cyprinidae) from cyto-chrome b sequences. –Copeia, 2002, 665-678. Drummond, A.J., & Rambaut, A., 2003: BEAST version 1.4 [computer program]. Available: http://evolve. zoo.ox.ac.uk/beast. Accessed 25 November 2006. Drummond, A.J., Ho, S.y.w., Phillips, M.J., & A. Rambaut., 2006: Relaxed phylogenetics and dating with confdence. -PLoS Biology 4, e88. Espinasa, L., & R.B. Borowsky., 2001: Origins and rela-tionship of cave populations of the blind Mexican tetra, Astyanax fasciatus, in the Sierra de El Abra. –Environmental Biology of Fishes, 62, 233-237. Espinasa, L., Rivas-Manzano, P. , & H. Espinosa-Pérez., 2001: A new blind cave fsh population of the genus Astyanax: geography, morphology and behavior. –Environmental Biology of Fishes, 62, 339-344. TIME in KARST – 2007 181 MEGAN L. PORTER, KATHARINA DITTMAR & MARCOS PéREZ-LOSADA Gayet, M., 1982: Consideration sur la phylogenie et la paleobiographie des Ostariophysaries. -Geobios Me-moir, 6, 39-52. Gayet, M., & F.J. Meunier., 2003: Palaeontology and pal-aeobiogeography of catfshes. In Arratia, G., Kapoor, B.G., Chardon, M., & Diogo, R., (eds.), Catfshes, Vol. 2, pp.491-522. Science Publishes, Enfeld, NH. Hubbs, C.L., & w.T. Innes., 1936: Te frst known blind fsh of the family Characidae: A new genus from Mexico. –Occasional Papers of the Museum of Zo-ology University of Michigan, 342, 1-7. Jefery, w.R., 2001: Cavefsh as a model system in evo-lutionary developmental biology. -Developmental Biology, 231, 1-12. Jefery, w.R., 2005: Adaptive evolution of eye degeneration in the Mexican blind cavefsh. -Journal of He-redity, 96, 185-196. Katoh, K., Kuma, K., Toh, H., & T.Miyata., 2005: MAFFT version 5: improvement in accuracy of multiple sequence alignment. -Nucleic Acids Research, 33, 511-518. Mateos, M., 2005: Comparative phylogeography of live-bearing fshes in the genera Poeciliopsis and Poecilia (Poeciliidae: Cyprinodontiformes) in central Mexico. -Journal of Biogeography, 32, 775-780. Mitchell, R.w., Russell, w.H., & w.R. Elliott., 1977: Mexi-can eyeless characin fshes, Genus Astyanax: environment, distribution, and evolution. Texas Tech University Special Publications of the Museum, 12, 89pp. Near, T.J., Kassler, T.w., Koppelman, J.B., Dillman, C.B., & D.P. Philipp., 2003: Speciation in North American black basses, micropterus (Actinopterygii: Centrar-chidae). –Evolution, 57, 1610-1621. Nixon, K.C., 1999: Te Parsimony Ratchet, a new meth-od for rapid parsimony analysis. –Cladistics, 15, 407-414. Peng, Z., He, S., wang, J., wang, w. , & R. Diogo., 2006: Mitochondrial molecular clocks and the origin of the major Otocephalan clades (Pisces: Teleostei): A new insight. –Gene, 370, 113-124. Perdices, A., & I. Doadrio., 2001: Te molecular system-atics and biogeography of the European cobitids based on mitochondrial DNA sequences. -Molecu-lar Phylogenetics and Evolution, 19, 468-478. Sikes, D.S., & P.O. Lewis., 2001: Beta sofware, version 1. PAUPRat: PAUP* implementation of the parsimo-ny ratchet. [computer program]. Available: http:// www.ucalgary.ca/~dsikes/sofware2.htm. Accessed 25 November 2006. Sivasundar, A., Bermingham, E., & G. Ortí., 2001: Population structure and biogeography of migratory freshwater fshes (Prochilodus: Characiformes) in major South American rivers. -Molecular Ecology, 10, 407-417. Strecker, U., Bernatchez, L., & H. wilkens., 2003: Genetic divergence between cave and surface populations of Astyanax in Mexico (Characidae, Teleostei). -Mo-lecular Ecology, 12, 699-710. Strecker, U., Faúndez, V.H., & H. wilkens., 2004: Phylo-geography of surface and cave Astyanax (Teleostei) from Central and North America based on cyto-chrome b sequence data. -Molecular Phylogenetics and Evolution, 33, 469-481. Swoford, D.L., 2000: PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4. Sinau-er Associates, Sunderland, MA. wiley, S. & R.w. Mitchell., 1971: A bibliography of the Mexican eyeless characin fshes of the genus Asty-anax. -Association for Mexican Studies Bulletin, 4, 231-239. yang, Z., 2001: PAML: Phylogenetic Analysis by Maximum Likelihood. University College London, London. yang, Z., 2004: A heuristic rate smoothing procedure for maximum likelihood estimation of species diver-gence times. -Acta Zoologica Sinica, 50, 645-656. Zardoya, R., & I. Doadrio., 1999: Molecular evidence on the evolutionary and biogeographical patterns of European cyprinids. -Journal of Molecular Evolution, 49, 227-237. 182 TIME in KARST – 2007