the evolutionary history of the genus timarcha (coleoptera, chrysomelidae) inferred from...

gdftiarambnpTrdcSMoasptr

CbNdctcfrbN1

Molecular Phylogenetics and EvolutionVol. 14, No. 2, February, pp. 304–317, 2000doi:10.1006/mpev.1999.0712, available online at http://www.idealibrary.com on

1CA

The Evolutionary History of the Genus Timarcha (Coleoptera,Chrysomelidae) Inferred from Mitochondrial COII Gene

and Partial 16S rDNA SequencesJesus Gomez-Zurita, Carlos Juan, and Eduard Petitpierre

Lab. Genetica, Departament de Biologia, Universitat de les Illes Balears, E-07071 Palma de Mallorca, Balearic Islands, Spain

Received March 5, 1999; revised July 20, 1999

gahocHttbP

hMSswt2saT

msmbctddstagfittllm

The apterous genus Timarcha consists of three sub-enera and more than 100 species in its Palearcticistribution, with specialized feeding on few plantamilies. Fifty-four sequences sampled from 31 taxa ofhe genus plus three outgroup leaf beetles were stud-ed for their complete cytochrome oxidase II (COII)nd a fragment of 16S rDNA mitochondrial genes,epresenting a total of about 1200 bp. Phylogeneticnalyses using maximum-parsimony and distanceethods for each gene separately and for the com-

ined data set gave compatible topologies. The subge-us Metallotimarcha consistently appears in a basalosition and is well differentiated from the remainingimarcha, but no clear monophyletic grouping of Tima-chostoma and Timarcha s. str. subgenera can be de-uced from our analysis. Calibration of the molecularlock has been done using the opening of the Gibraltartrait after the Messinian salinity crisis (about 5.5YA) as the biogeographic event causing disjunction

f two particular taxa. Accordingly, the COII evolution-ry rate has been estimated to be of 0.76 3 1028

ubstitution/site/year in Timarcha. Relation betweenhylogeny and host-plant use indicates widening ofrophic regime as a derived character in Timarcha.2000 Academic Press

INTRODUCTION

The leaf beetles of the genus Timarcha (Coleoptera,hrysomelidae) are popularly known as ‘‘bloody noseeetles’’ in Europe and ‘‘strawberry leaf beetles’’ inorth America. They comprise more than 100 speciesistributed in Europe, Turkey, and Transcaucasianountries in West Asia, and North Africa from Moroccoo Libya, and there are two representatives in the westoast of North America. The latter species, distributedrom Vancouver to northern California are possiblyelictual taxa. This peculiar disjunct distribution coulde the product of vicariance after the formation of theorth Atlantic fracture during the Tertiary (Jolivet,
994). The Timarcha species are flightless phytopha- m
304055-7903/00 $35.00opyright r 2000 by Academic Pressll rights of reproduction in any form reserved.

ous insects feeding mostly on herbaceous Rubiaceaend/or Plantaginaceae, depending on the availability ofost plants. However, some taxa feed on Rosaceae (thenly host family for the NorthAmerican taxa), Brassica-eae, Scrophulariaceae, Dipsacaceae, or Asteraceae.erbaceous Rubiaceae are hypothesized to be the ances-

ral host plant for the Old World Timarcha and switcheso Plantago and other plants have probably been forcedy ecological changes in their habitats (Jolivet andetitpierre, 1973; Crowson, 1981).These beetles occur from the high-altitude alpine

abitats to humid forests in Central Europe, littoralediterranean areas, and even flatland regions or

aharian oases. More than a third of the describedpecies occur as endemics in the Iberian Peninsula, inhich they have experienced extensive local diversifica-

ion. Bechyne (1948) divided the tribe Timarchini into8 morphological-biogeographical groups and into fourubgenera, one in the Nearctic (Americanotimarcha)nd three in the Palearctic regions (Timarcha s. str.,imarchostoma, and Metallotimarcha).The subgenus Timarchostoma includes a group oforphological types similar to the species T. goettingen-

is, an alliance presenting many synapomorphies at theorphological and karyological levels. This group has

een defined as the Timarcha goettingensis speciesomplex (Petitpierre, 1970a), with the implicit proposi-ion that an incipient speciation is occurring among theifferent local populations within the group. Patchedistributions caused by low vagility and host-plantpecialization are factors that can restrict gene flow inhese beetles, hence promoting geographical isolationnd speciation. However, to study this attractive modelroup of organisms under an evolutionary perspective,rm hypotheses are needed for the phylogenetic rela-ionships within this clade of closely related taxa or forhe major groups of Timarcha. Adult and larval morpho-ogical (Bechyne, 1948; Steinhausen, 1994) and karyo-ogical (Petitpierre, 1970b, 1976) data have provided

any informative characters for such analysis but, in
any cases, defining the polarity of some of the charac-

tvicwnl

ldfsstm1

plwf(Mkepc

u(erds

PtTomcwg

S

taucaaot(

st

D

adswGlCS7CCGocc5fiCtcSrqBwk

S

etpqsaoc

CfcDgs

t4paaF

305MITOCHONDRIAL PHYLOGENY OF LEAF BEETLE GENUS Timarcha

er states is problematic. Data on the extent of geneticariation present in different populations of these taxan relation to geographical distribution and morphologi-al diversification is also necessary. For these reasons,e have undertaken a study of mitochondrial anduclear DNA sequences from Timarcha in addition to

ife history.This paper presents the first insight into the molecu-

ar phylogeny of Timarcha using the complete mitochon-rial cytochrome oxidase subunit II gene (COII) and aragment of the large ribosomal subunit (16S) geneequences, representing a total of 1.2 Kb. We haveequenced 54 individuals from 31 taxa representing allhree Palearctic subgenera and most (about 75%) of theorphological-biogeographical groupings (Bechyne,

948).The COII gene has been extensively used to address

hylogenetic questions in insects at different taxonomicevels: among orders (Liu and Beckenbach, 1992),ithin an order (Frati et al., 1997), and more success-

ully within genus or species groups, including beetlesBeckenbach et al., 1993; Emerson and Wallis, 1995;

ardulyn et al., 1997, among others). There is a goodnowledge of the structure and function of the proteinncoded by COII (Capaldi, 1990) that enables one toredict which gene regions can suffer substitutionalonstraints during their evolution.Part of the mitochondrial rDNA 16S sequences (16S)

sed in this paper have been reported elsewhereGomez-Zurita et al., 1999). We have expanded consid-rably this preliminary data in order to perform sepa-ate and combined phylogenetic analyses of the twoata sets to robustly ascertain the evolutionary relation-hips within Timarcha.We address the following questions: (1) are the

alearctic subgenera monophyletic groupings?; (2) ishe subgenus Metallotimarcha basal to the remainingimarcha, as has been proposed?; (3) what is the levelf genetic divergence and substitution rate for theajor clades and, more specifically, within the Timar-

ha goettingensis species complex?; and finally, (4) cane infer the history of host-plant affiliation for thisenus?

MATERIAL AND METHODS

ampling

The complete list of taxa analyzed, their sources, andhe classical subgeneric and morphological groupingre given in Table 1. For outgroup comparisons we havesed sequences obtained from the chrysomelids Copto-ephala scopolina and Macrolenes dentipes (Clytrinae)nd a seed beetle, Bruchus sp. (Bruchidae), which someuthors consider as belonging to chrysomelids. Theseutgroup taxa are part of two lineages that are believedo be basal divergences in the family Chrysomelidae
Reid, 1995; Farrell, 1998). Voucher specimens of all D
tudied taxa have been deposited in the UIB Labora-ory of Genetics beetle collection.

NA Extraction, PCR, and Sequencing

DNA extraction and purification from single individu-ls were performed from abdomen (after removing theigestive tract) or head–thorax homogenates followingtandard techniques (Juan et al., 1993). Primers usedere a modified TL2-J-3037 (58 TAATATGGCA-ATTAGTGCATTGGA 38) and a modified and slightly

onger TK-N-3785 (58 GAGACCATTACTTGCTTT-AGTCATCT 38), following the primer nomenclature ofimon et al. (1994) for the amplification of a fragment of77 bp containing part of the tRNA-Leu and the wholeOII gene. LR-N-13398 (58 CGCCTGTTTATCAAAAA-AT 38) and a modified LR-J-12887 (58 CTCCGGTTT-AACTCAGATCA 38) were used for the amplificationf a fragment of about 550 bp of 16S rDNA. PCRonditions were as follows: 4 min at 95°C followed by 35ycles of denaturation at 95°C for 30 s, annealing at0°C for 1 min, and extension at 72°C for 2 min, with anal single extra extension step at 72°C for 10 min.OII PCR products were sequenced in the two direc-

ions with digoxigenin-labeled primers and performingycle sequencing reaction with the DIG Taq DNAequencing Kit for Standard Cycle Sequencing (Boeh-inger) following manufacturer’s instructions. Se-uences were run in a GATC 1500-System with Directlotting Electrophoresis. For 16S rDNA, PCR productsere cloned using the pMOSBlue blunt-ended cloningit (Amersham) and sequenced as above.

equence and Phylogenetic Analyses

COII sequence alignment was done by hand. At the 38nd of the gene, the Timarcha sequences were difficulto align with those of the outgroups and alignment waserformed by means of the inferred amino acid se-uences. The codon position assignation was by compari-on with those reported in Liu and Beckenbach (1992)nd the inference of the amino acid sequences wasbtained by using the Drosophila mitochondrial geneticode (De Bruijin, 1983).The alignment of the 16S sequences was done with

LUSTAL W, version 1.7 (Thompson et al., 1994). Weound a highly variable region in ingroup/outgroupomparisons between position 241 and position 298.ifferent alignment conditions were tried for this re-ion using MALIGN version 1.99 (Wheeler and Glad-tein, 1992–1994).Sequence analyses (nucleotide composition, substitu-

ion rates, etc.) were performed with test version.0d64 of PAUP* written by David L. Swofford, Li93rogram (Li, 1993), and MacClade v. 3.0.3 (Maddisonnd Maddison, 1992). Estimates of nucleotide diversitynd testing of natural selection with Tajima (1989) andu and Li (1993) tests were done with the program
naSP v. 2.0 (Rozas and Rozas, 1997).

o4crAsKlTnppbasnoctcIssn

wam(Ptmot1tdtwtS

R

aC‘‘wsctfatt

fi(

C

tLfiatlbcA(sSC1qwcOmcp

tm7m1pa

wemmOttN(s

mttpart

306 GOMEZ-ZURITA, JUAN, AND PETITPIERRE

Maximum-parsimony (MP) and distance-based meth-ds for phylogenetic inferences were applied using.0d64 test version of PAUP*. For MP analyses we usedharacters equally weighted, the heuristic algorithm,andom stepwise addition with 25 replicates,CCTRAN, and TBR swapping options. Distance analy-es were performed using neighbor-joining and theimura two-parameter correction. Support for particu-

ar nodes was assessed by 500 bootstrap resamplings.o use COII and 16S data sets in a combined phyloge-etic analysis, we assessed congruence between bothartitions using the ILD test (Farris et al., 1994;artition homogeneity test in PAUP*). This test isased on the incongruence index defined by Mickevichnd Farris (1981), and the distribution of the ILDtatistic is computed after randomization of the origi-al data set partitions, constructing random data setsf the same size as the original and calculating theirorresponding incongruence index. The incongruencehreshold for this test is P , 0.05, although it isonsidered too conservative (Cunningham, 1997). TheLD test was performed using PAUP* with a heuristicearch of 5000 replicates, TBR in effect, MULPARS notelected, and removing all invariant characters (Cun-ingham, 1997).MacClade v. 3.0.3 (Maddison and Maddison, 1992)as used to map chromosome numbers and host plantffiliation on the cladograms and to calculate theinimum shifts expected under accelerated

ACCTRAN) or delayed (DELTRAN) transformations.ermutation tests (T-PTP option in PAUP*) were usedo ascertain whether there are associations with theitochondrial phylogeny and the observed distribution

f character states for both chromosome numbers androphic selection. This was accomplished by performing0000 random permutations of the associated charac-er state of the taxa using the topology of the mitochon-rial phylogeny as a constraint tree. Finally, a distribu-ion of tree lengths was obtained that allows testinghether the original tree length is significantly shorter

han expected under a random model (Maddison andlatkin, 1991).

ate Constancy Test

In order to determine whether there is an evolution-ry rate constancy at different divergence levels ofOII and 16S Timarcha sequences, we applied the

‘relative-rate test’’ by Sarich and Wilson (1973) and the‘two-cluster test’’ by Takezaki et al. (1995). The latter

as computed using the software available at the ftpite of the authors. The two-cluster test is based on theomparisons of average substitution rates for two clus-ers (each with one or more sequences) branching offrom an internal node in a given tree. Under thessumption of a molecular clock, the difference betweenhe accumulation of substitutions per site in each of the
wo clusters is not expected to be significantly different 1
rom zero. The deviation of this difference from zero andts significance can be tested by a two-tailed normal testTakezaki et al., 1995).

RESULTS

OII Gene

All Timarcha species show 685 nucleotides (nts) forhis gene encoding to 228 amino acids. As reported byiu and Beckenbach (1992) for several insects, only therst position of the stop codon, a single T, is present inll studied Timarcha plus the outgroup Bruchus sp. Inhe two Clytrinae representatives (Coptocephala scopo-ina and Macrolenes dentipes) the stop codon seems toe complete (TAA). There is variability for the initiationodon as well; most Timarcha and Bruchus sp. showTT, while the remaining Timarcha species show ATC

both being isoleucine codons). Finally, the Clytrinaeequences presentATG (methionine) as initiation codon.imilar variation has been described previously forOII sequences of several insects (Liu and Beckenbach,992). A comparison of the inferred amino acid se-uences of Timarcha shows 52 variable positions ofhich 26 are conservative (with similar physicochemi-

al properties) changes (French and Robson, 1983).nly 21 of these changes are informative under parsi-ony, and 3 (positions 95, 119, and 156) define unequivo-

ally a group formed by T. espanoli, T. rugosa, and T.unctella.Table 2 summarizes the nt composition and substitu-

ion patterns in both the COII and the studied frag-ent of 16S (see below). The sequences are on average

4.4% A 1 T biased, a well-established fact in insectitochondrial sequences (Clary and Wolstenholme,

985; DeSalle et al., 1987). This is in part caused by areference for codons ending in A or T (A 1 T percent-ge rises to 90.7% at third codon positions; see Table 2).Two to seven individuals from different localitiesere sequenced for nine of the Timarcha species. Thestimated nucleotide diversity p (Table 3) was maxi-um for four sequences of T. intermedia (3.11%) andinimum for five sequences of T. interstitialis (0.91%).n average, nucleotide diversity is relatively high for

hese taxa, i.e., 16 positions of every 1000 differ be-ween two random sequences of a Timarcha species.either Tajima’s D (Tajima, 1989) nor Fu and Li’s

1993) tests show evidence of natural selection for theseequences.Divergences corrected by the Kimura two-parameterethod show a range of 0.30–12.28 (average: 7.36) in

he comparisons within Timarcha, within Timarchos-oma, and between both subgenera. Divergences amongairwise comparisons of species of the two subgenerare in a similar range. Comparisons among the twoepresentatives of the subgenus Metallotimarcha andhe remaining Timarcha taxa give values of 11.36–
6.09 (average: 12.97). About one-third of the total

M

T

T

O

M

List of Studied Taxa, with Their TaxonoSampling Localities, and

SubgenusTaxon Code

Morphologicalgroupa

etallotimarcha Motsch.Timarcha metallica (Laich.) MET1 1T. metallica (Laich.) MET2 1

imarchostoma Motsch .T. interstitialis Fairm. INS1 3T. interstitialis Fairm. INS2 3T. interstitialis Fairm. INS3 3T. interstitialis Fairm. INS4 3T. interstitialis Fairm. INS5 3T. goettingensis (L.) GOE 3T. sinuaticollis Fairm. SIN1 3T. sinuaticollis Fairm. SIN2 3T. sinuaticollis Fairm. SIN3 3T. montserratensis Bech. MON 3T. recticollis Fairm. REC 4T. cyanescens Fairm. CYA 5T. perezi Fairm. PER1 6T. perezi Fairm. PER2 6T. perezi Fairm. PER3 6T. perezi Fairm. PER4 6T. perezi Fairm. PER5 6T. perezi Fairm. PER6 6T. perezi Fairm. PER7 6T. geniculata (Germ.) GEN1 7T. geniculata (Germ.) GEN2 7Timarcha sp. TSP 6–7T. gougeleti Fairm. GOU 8T. strangulata Fairm. STR 9T. hispanica H.-S. HIS1 12T. hispanica H.-S. HIS2 12T. hispanica H.-S. HIS3 12T. calceata Perez CAL 14T. fallax Perez FAL 15T. aurichalcea Bech. AUR 15T. granadensis Bech. GRA 15T. intermedia H.-S. INM1 16T. intermedia H.-S. INM2 16T. intermedia H.-S. INM3 16T. intermedia H.-S. INM4 16T. balearica Gory BAL1 17T. balearica Gory BAL2 17T. balearica Gory BAL3 17T. insparsa Rosenh. INP 18T. marginicollis Rosenh. MAG 18T. coarcticollis Fairm. COA 19T. lugens Rosenh. LUG 21

imarcha s. str. Dej.T. maroccana Weise MAC 22T. espanoli Bech. ESP 24T. rugosa L. RUG 24T. pimelioides H.-S. PIM 26T. tenebricosa F. TEN1 27T. tenebricosa F. TEN2 27T. nicaeensis Villa NIC1 27T. nicaeensis Villa NIC2 27T. punctella Mars. PUN1 28T. punctella teleuetica Bech. PUN2 28utgroupsBruchus sp. —Coptocephala scopolina (L.) —Macrolenes dentipes (O1.) —

a

TABLE 1

mic and Morphological Group Placement,Sequence Accession Nos.

Localityb Accession Nos.c

Schlagebachtal, Thuringia, GER AJ236315/AJ236367Foret d’Anlier, Anlier, BEL Y18821/Y18826

Alio, Tarragona, SP AJ236316/AJ236368L’Esquirol, Barcelona, SP AJ236317/AJ231148Garraf, Barcelona, SP AJ236318/AJ236369Caramany, FR AJ236319/AJ236370Sant Pere de Roda, Girona, SP AJ236320/AJ236371Jena, Thuringia, GER AJ236321/AJ231142Vall de Nuria, Girona, SP AJ236322/AJ231156Queixans, Girona, SP AJ236323/AJ236372La Creueta, Girona, SP AJ236324/AJ235373Collformic, Barcelona, SP AJ236325/AJ236374Pla de l’Artiga, Lleida, SP AJ236326/AJ231154Cueto, Cantabria, SP AJ236328/AJ231139Villanubla, Valladolid, SP AJ236329/AJ8078Zaragoza, SP AJ236330/AJ231152Ejulve, Teruel, SP AJ236331/AJ236375Puerto de Linares, Teruel, SP AJ236332/AJ236376Pto. de Cuarto Pelado, Teruel, SP AJ236333/AJ236377El Moncayo, Zaragoza, SP AJ236334/AJ236378Layna, Soria, SP AJ236335/AJ236379Puerto de Ventana, Leon, SP AJ236336/AJ231141Pena Amaya, Burgos, SP AJ236337/AJ236380Puerto de S. Isidro, Leon, SP AJ236338/AJ231160Fiobre-Bergondo, A Coruna, SP AJ236339/AJ236381Pla de l’Artiga, Lleida, SP AJ236340/AJ231157Sao Vicente Cape, Algarve, POR AJ23641/AJ231144Sao Torpes beach, Alentejo, POR AJ236342/AJ236382Pto. de Caracuel, Ciudad Real, SP AJ236343/AJ236383Vegacervera, Leon, SP Y18823/Y18828Pto. de la Carrasqueta, Alicante, SP Y18822/Y18827Tragacete, Cuenca, SP AJ236327/AJ231136Sierra de Guillimona, Granada, SP AJ236344/AJ231143Arenales del Sol, Alicante, SP AJ236345/AJ236384Isla Nueva Tabarca, Alicante, SP AJ236346/AJ231147Rodalquilar, Almerıa, SP AJ236347/AJ231146Nacimiento, Almerıa, SP AJ236348/AJ236385Esporles, Mallorca, SP AJ236349/AJ231137Alcoufar, Menorca, SP AJ236350/AJ236386Puig de Randa, Mallorca, SP AJ236351/AJ236387Pico Veleta, Granada, SP AJ236352/AJ231145Pico Veleta, Granada, SP AJ236353/AJ231150Los Barrios, Cadiz, SP AJ236354/AJ231138Pico Veleta, Granada, SP AJ236355/AJ231149

Tanout-ou-Fillali, Moyen Atlas, MO AJ236356/AJ231151L’Altet, Alicante, SP AJ236357/AJ231140Lengua de Tierra, Nador, MO AJ236358/AJ231155C. da Cugno Vasco, Sicily, IT Y18819/Y18824Planoles, Girona, SP AJ236359/AJ231159Port de la Bonaigua, Lleida, SP AJ236360/AJ231158Mt. Zatta, Genova, IT AJ236361/AJ236388S. Cataldo, Puglia, IT Y18820/Y18825Lengua de Tierra, Nador, MO AJ236362/AJ231153Kelaa des Sahrna, Moyen Atlas, MO AJ236363/AJ236389

Campus UIB, Mallorca, SP AJ236364/AJ231161Campus UIB, Mallorca, SP AJ236365/AJ231162Campus UIB, Mallorca, SP AJ236366/AJ231165

The systematic arrangement of the species analyzed and the morphological group assignation follow the proposal by Bechyne (1948).b Abbreviations correspond to the name of the country of the sampling site: BEL, Belgium; FR, France; GER, Germany; IT, Italy; MOR,orocco; POR, Portugal; SP, Spain.c The first EMBL accession no. refers to COII sequences and the second to the 16S fragment sequences.

potwFtCetis8mn

1

CSaptpa(

s1tsvtaaavgmen

bcmcm1

t3mMissa(hoc1tvass

C

1

wM

TTTTTT


ositions of COII are variable, nearly 70% of whichccur at third codon positions (see Table 2). Transi-ional changes are more abundant than transversions,hich in turn are mostly composed of A–T changes.igure 1 shows a plot relating the proportion of transi-ional changes as a function of divergence in TimarchaOII. As has been repeatedly shown elsewhere (Simont al., 1994, and references therein), there is a drop inhe transitional changes as divergence increases. Thiss, in fact, caused by a monotonic increment of transver-ions, which begin to show a saturation effect at about–10% divergence. Figure 2 shows a plot of synony-ous (Ks) substitutions as a function of the totalumber of changes in pairwise comparisons.

6S rDNA Sequences

The same species and individuals used for obtainingOII data were sampled for a fragment of 16S rDNA.equence multialignment was unambiguous except forhighly variable segment between position 241 and

osition 298, in particular in the comparisons involvinghe outgroups. The exclusion of this region in thehylogenetic analysis resulted in loss of informationnd subsequent lack of resolution in the 16S treessensu Gatesy et al., 1993). Prediction of the secondary

TAB

Summary of Nucleotide Compositio

Marker A% C% G% T%Totalsites

OII 36.37 14.36 11.22 38.05 6851st Pos. 34.97 15.57 19.66 29.81 2292nd Pos. 26.87 19.68 12.47 40.98 2283rd Pos. 47.28 7.83 1.48 43.41 228

6Sa 39.40 15.52 9.12 35.97 496Stem 38.14 16.99 10.79 34.07 302Loop 41.35 13.22 6.52 38.92 194

Note. Ts, transitions and Tv, transversions calculated as minimuma Excluding the G3 loop (positions 255–281) ambiguously aligned.

TABLE 3

Nucleotide Diversity (p) 6 Standard Errors (SDE)ithin Timarcha Species for the Two Mitochondrialarkers Studied

SpeciesSample

size

COII 16S

n p SDE (p) n p SDE (p)

. balearica 3 654 0.0214 0.0083 491 0.0068 0.0032

. hispanica 3 685 0.0039 0.0018 489 0 0

. intermedia 4 655 0.0099 0.0024 491 0.0065 0.0015

. interstitialis 5 679 0.0091 0.0022 488 0.0025 0.0010

. perezi 7 675 0.0227 0.0046 488 0.0076 0.0021

. sinuaticollis 3 675 0.0311 0.0105 490 0.0109 0.0045

iNote. n, Number of nucleotide positions analyzed.

tructure based on that proposed for Drosophila yakuba6S rRNA (Gutell and Fox, 1988) allowed determina-ion of the nt homologies of the stems in the ingroupubset. This is particularly important in the case of theariable segment, reducing the ambiguity to 27 nts ofhe G3 loop, following the nomenclature by De Rijk etl. (1996). To assess the effect of different alignmentlternatives involving ingroup/outgroup comparisonsnd bearing in mind the secondary structures for thisariable region, we used MALIGN, assigning differentap-to-nt change costs from 1:1 to 20:1. All the align-ents obtained gave essentially the same phylogenetic

stimate with respect to resolution and support of theodes.The A 1 T content is similar to that of COII (75.4%)

ut is lower in stems than in loops (see Table 2). Thisould be related to selective constraints to maintain ainimum of the more stable G–C pairs in stems

ompared to the loop regions. Intraspecific variationeasured by p is on average three times lower in the

6S fragment than in COII (Table 3).Interspecific divergence (corrected by the Kimura

wo-parameter method) ranges from 0 to 6.9% (average.5%) in the ingroup comparisons, excluding Metalloti-archa, and 4.6–7.3% (average 5.7%) among Timarcha–etallotimarcha comparisons. Thus, divergence in 16S

s between two and three times lower than that ob-erved for COII. For all the ingroup pairwise compari-ons, 19.6% of the positions are variable, of whichlmost two-thirds are phylogenetically informativeTable 2). Nevertheless, variation on the 16S segment iseterogeneous along the sampled sequence and levelsf variation in different regions can be related to theirorresponding A 1 T content (Han and McPheron,997). No clear-cut relationship has been found be-ween transition/transversion ratio and nucleotide di-ergence for 16S in our data (not shown). In addition,nalysis of the substitutions present among Timarchahows that, in the G3 stem, pairs A–T in some taxa areubstituted by pairs T–A in others, which can be

2

for COII and 16S Data of Timarcha

ariablesites

Informativesites Ts Tv Tv AT Ts 1 Tv

259 182 144 52 41 6349 35 30 12 7 729 4 20 8 4 1

181 143 94 32 30 5597 66 43 31 27 2354 40 18 20 17 1643 26 25 11 10 7

erved from the data matrix.

LE

n

V

obs

nterpreted as the effect of compensatory mutations.

Sh

P

scC

otCtatat4cauttstlNltibitcnctsR5

C . Co

ta


ome stems show noncanonical G–T pairs which alsoelp in maintaining stem stability (Gutell et al., 1994).

hylogenetic Analysis

COII and 16S data were used separately to recon-truct the gene phylogenies. Figure 3a shows a strictonsensus obtained under parsimony analysis usingoptocephala, Macrolenes, and Bruchus sequences as

FIG. 1. Proportion of transitions among all changes in pairwise coomparisons at different taxonomic levels are indicated with symbols

FIG. 2. Relationship between number of substitutions that areransitions or transversions and synonymous rate (Ks) calculated

Lccording to Li (1993) in COII of Timarcha.

utgroups. In this analysis 558 equally parsimoniousrees were recovered (one single island of 1073 steps;.I. 5 0.468; R.I. 5 0.650). Neighbor-joining (NJ) COII

rees obtained using divergence estimates correctedssuming several substitutional models rendered essen-ially the same topology (not shown). Using the samessumptions as above, the 16S sequences gave the MPree shown in Fig. 3b (one single island of 290 MP trees,26 steps; C.I. 5 0.629; R.I. 5 0.758), which also isompatible with the NJ analysis. As mentioned above,lternative alignments of the variable region obtainedsing a range of gap-to-nt change costs gave the sameopology. As evidenced by the COII topology, many ofhe basal nodes obtained in the 16S tree were weaklyupported by bootstrap resampling. A comparison ofhe COII- and 16S-based topologies shows that there isittle conflict between the two phylogenetic hypotheses.amely, T. intermedia and T. lugens appear monophy-

etic for COII but not for 16S, and conflicts for T.enebricosa–T. nicaeensis and T. punctella–T. rugosanterrelationships are also present in the two gene-ased phylogenies. All other major clades are the samen the two topologies and the bootstrap supports forhem are similar. Furthermore, the ILD test has beenomputed, rendering a P value of 0.195, which isonsignificant; so we concluded that both data sets areombinable. Figure 4 shows a parsimony tree based onhe combined data set and computed as a strict consen-us of the 96 MP trees obtained (981 steps; C.I. 5 0.460;.I. 5 0.705), in which bootstrap values higher than0% are shown for both parsimony and NJ analyses.

arisons of Timarcha COII as a function of total nucleotide divergence.mparisons with outgroups are shown in a different scale for clarity.

mp

ong-branch attraction has been described elsewhere

FIG

.3.

CO

II(a

)an

d16

SrD

NA

(b)

MP

stri

ctco

nse

nsu

str

ees

con

stru

cted

usi

ng

equ

alw

eigh

tsfo

ral

lch

arac

ters

,wit

hT

BR

and

AC

CT

RA

Nin

effe

ct.

Th

etr

ees

repr

esen

tth

eph

ylog

enet

icre

lati

onsh

ips

infe

rred

from

each

mar

ker

usi

ng

Cop

toce

phal

afl

oral

isan

dM

acro

len

esd

enti

pes

asou

tgro

ups

con

side

red

mon

oph

ylet

ic.C

onfi

den

cele

vels

oftr

eebi

furc

atio

ns

wer

eas

sess

edby

500

boot

stra

pre

plic

ates

;val

ues

hig

her

than

50%

are

indi

cate

dfo

rea

chn

ode

both

for

MP

(firs

tva

lue)

and

NJ

(sec

ond

valu

e)an

alys

es.


T(blc


FIG. 4. Phylogenetic hypothesis for the evolutionary relationships in Timarcha based on the combination of COII and 16S rDNA data sets.he tree was reconstructed under a maximum parsimony criterion and is the strict consensus of the 96 equally parsimonious trees obtained

981 steps; C.I. 5 0.460; R.I. 5 0.705). Sequences of T. metallica were specified as outgroup (see text for details). Robustness was assessed byootstrap resampling (500 replicates) and values higher than 50% are given for the MP (first value) and NJ (second value) analyses. Circledetters designate nodes and species groups discussed in the text. Vertical bars with different patterns show the current subgeneric
lassification for the species studied.

(etedrrftincas

cimbP1Imhr(1

P

stomnvKacldc(mebpcnaatTaa

sssI

ntmcsptotars1toserga2cflsc1

cCictsvhiclpgtspo

cegeb


Hendy and Penny, 1989) as a problem in phylogeneticstimation. In our sequences, long divergences leadingo the outgroups could cause distortion. To avoid thisffect and as previous single-data-set analyses firmlyemonstrate the basal placement of T. metallica withespect to the other Timarcha sampled taxa, we haveooted this tree accordingly. The tree obtained rein-orces most of the supported nodes of single-geneopologies and favors the grouping of T. lugens and T.ntermedia with a high bootstrap support, while it doesot confirm monophyly for T. nicaeensis and T. tenebri-osa. All the nodes weakly supported by single-genenalyses remain unsupported in the combined analy-is.Some clades of the mitochondrial phylogeny (indi-

ated in Fig. 4 with circled letters) are of biologicalnterest and will be discussed below: (A) a group of

orphologically heterogeneous species characterizedy having the highest chromosome numbers in thealearctic Timarcha (2n 5 26–30) (Petitpierre et al.,988); (B) European Timarcha s. str. and southeasternberian and Balearic lineages, well-defined from aorphological point of view; (C) well-supported T.ispanica and T. granadensis clade, species clearlyelated on morphological grounds (Bechyne, 1948); andD) the T. goettingensis species complex (Petitpierre,970a).

DISCUSSION

hylogenetic Inferences

The phylogenetic analyses using COII and 16S datahow clearly the basal position of T. metallica withinhe Palearctic lineage of the genus. This species andther allied taxa (T. gibba, T. corinthia, and T. hum-eli) belong to a distinct clade classified at the subge-

eric rank (Metallotimarcha), based on adult and lar-al morphological characters (Bechyne, 1948; Iablokoff-hnzorian, 1966; Steinhausen, 1994). Our genetic data,lthough limited to only one species of this group, areonsistent with this interpretation, and the subgeneric-evel classification is reinforced by the relatively highivergence found with respect to the remaining Timar-ha (almost 13% for COII and 5.7% for 16S). Bechyne1948), in his Timarcha monograph, and, later, Stock-ann (1966) considered two further Palearctic subgen-

ra: Timarchostoma and Timarcha s. str., which haveeen more controversial (e.g., Stockmann, 1966). Thisroposal was based on several adult morphologicalharacters, some of which are probably of little phyloge-etic value, such as the relative length of male tarsind adult size, among others. The polarity implicitlypplied to character states led to the conclusion thathe clade Timarcha s. str. is relatively recent, whileimarchostoma and Metallotimarcha represent morencestral lineages. This phylogenetic hypothesis, in
ddition to the geographical distribution of the three t
ubgenera, argued for a northeast–southwest expan-ion of the genus with an origin in Central Asia and/orecondary diversification and colonization areas in theberian Peninsula and North Africa (Jolivet, 1994).

The mitochondrial-based phylogenies obtained doot show any evidence for monophyletic Timarchos-oma and Timarcha s. str. groupings. Neither parsi-ony nor distance methods support a Timarcha s. str.

lade, since the phylogenetic relationships of theampled taxa are unresolved (they fall in the basalolytomy), although they are always positioned close tohe root of the tree. The A group (Fig. 4) consists mainlyf species of the Timarcha s. str. subgenus. However,wo of the species included in this group, T. strangulatand T. calceata (subgenus Timarchostoma), have beenelated to other species groups, based on morphologicalynapomorphies and their distribution areas (Bechyne,948). Interestingly, both of them separate clearly fromhese species groups in their karyotypes, which aretherwise very homogeneous and conserved. For in-tance, T. strangulata, a very localized high-altitudendemic species of the Central Pyrenees, has beenelated to the other Pyrenean endemics of the T.oettingensis complex (T. sinuaticollis, T. recticollis,nd others) (Bechyne, 1948), but this species showsn 5 28, contrasting with the 2n 5 20 for this speciesomplex (Petitpierre et al., 1988). Similarly, T. calceata,rom mountain areas in Central Spain, which is morpho-ogically related to T. hispanica among the Iberianpecies, differs from the latter in having a higherhromosome number of 2n 5 30 (Petitpierre, 1970b,976).The mitochondrial relationships among the Timar-

hostoma representatives show three groups of taxa (B,, and D). The group B falls in the basal polytomy and

s characterized by species with shared morphologicalharacter states (shape of mesosternum, shape of prono-um, elytral punctuation, etc.) and 2n 5 20–22 chromo-omes. Another monophyletic clade (C) (100% bootstrapalue) is formed by the T. hispanica–T. granadensisaplotypes. Finally, the sister clade of the former

ncludes all the analyzed species of the T. goettingensisomplex (D), a group highly homogeneous on morpho-ogical and karyological grounds (Bechyne, 1948; Petit-ierre, 1970b, 1976). Interestingly, the T. hispanica–T.ranadensis taxa (group C) share morphological charac-er states with group B (same bifidous mesosternumhape) and with group D (e.g., conspicuous elytralunctuation and non-heart-shaped pronotum, amongthers) (Bechyne, 1948).Chromosomal data gathered from 34 taxa of Timar-

ha (Petitpierre et al., 1988) have been used to hypoth-size an ancestral diploid number of 2n 5 20 for theenus, from which increases occurred in several lin-ages (Petitpierre, 1973). Chromosomal changes haveeen traced on the 50%-majority rule consensus tree for
he combined analysis of COII and 16S in Timarcha

uTc(toTaaclrddiat

s

omDcewerts

T

4vaet

oouat


sing both ACCTRAN and DELTRAN algorithms. The-PTP test shows that the chromosomal data is signifi-antly correlated with the phylogeny of TimarchaP 5 0.0294). The pattern obtained (Fig. 5A) suggestshat from an ancestral 2n 5 20 a shift to 2n 5 22 hasccurred independently in the lineages of T. balearica,. tenebricosa, and T. granadensis (a single shift in thencestor of T. tenebricosa and T. balearica, assumingccelerated transformation algorithm). An increase inhromosomal number to 2n 5 28, subsequently fol-owed by decreases to 2n 5 26 (T. espanoli and T.ugosa) and an increase to 2n 5 30 (T. calceata) iseduced in clade A. Finally, the only documentedecrease from the basic 2n 5 20 refers to T. aurichalcea,n which a translocation involving an autosomal pairnd sex chromosomes has reduced the diploid numbero 2n 5 18 (Gomez-Zurita et al., unpubl.).

The branching order in our phylogenetic estimate

FIG. 5. (A) Character tracing of chromosome numbers in karptimization on a phylogenetic hypothesis represented by the majobtained in the MP analysis. Each chromosome number is treated as asing ACCTRAN. The T-PTP test shows that the trophic selection is siffiliation is given according to Jolivet and Hawkeswood (1995). T.rophisms are not well established.

hows conflicts with the systematic propositions based s

n nonmolecular data. In the phylogeny obtained, theost recent radiation is unequivocally that of the C andlineages, which are considered ancestral in the

lassical systematics of the genus. Furthermore, thisvolutionary scenario is more compatible with south-est–northeast expansion from North Africa and south-ast of Iberia (excluding the more ancient Metallotima-cha lineage), which is the geographical region showinghe deeper mitochondrial divergence and more diver-ity of taxa.

he Timarcha goettingensis Species Complex

The T. goettingensis species complex (clade D in Fig.) has been defined as a cluster of taxa morphologicallyery plastic but similar to T. goettingensis, which arellopatrically distributed and show a process of incipi-nt speciation (Petitpierre, 1970a, 1973). Most of theaxa are mountainous endemics in the Iberian Penin-

ogically studied species of Timarcha using ACCTRAN character-rule consensus of the 96 equally parsimonious trees (COII 1 16S)gle category. (B) Tracing of host-plant family use of Timarcha beetles

ficantly correlated with the phylogeny of Timarcha (P , 0.001). Plantichalcea, T. coarcticollis, and T. granadensis are excluded as their

yolritysin

gniaur

ula, although T. goettingensis and a few other taxa are

fllcditHdapmiisscTlCsrviTrTcTflppndstrhdshfc1popdtetigtitf

tt

A

1alc‘tcoctaTb

soi1oifCb(bmcys1((a

scbtsioy1tmdsmfb


atland species with larger distribution areas. Karyo-ogically, all the morphs share a highly conservedhromosome set of 2n 5 20 chromosomes with smallifferences that can be attributed mainly to pericentricnversions. The latter could have produced reproduc-ive isolation in some of the taxa (Petitpierre, 1973).erbaceous oligophagy, limited patchy mountainousistribution, and low vagility are factors which shouldccount for small deme sizes and isolation by distancerocesses leading to speciation in these beetles. Theitochondrial sequences obtained for the taxa included

n the complex confirm the monophyly of the group andts sister relationship to the T. hispanica–T. granaden-is clade. Within this complex, at least three differentpecies subgroups that are geographically structuredan be distinguished. T. goettingensis 1 T. recticollis 1. interstitialis (including one haplotype of T. sinuaticol-is) form one of these subgroups (D1), distributed fromentral Europe to the northeast of the Iberian Penin-ula. Divergence within this subgroup is relativelyecent, as exemplified by the 3 and 1% for COII and 16Salues, respectively, between T. goettingensis and T.nterstitialis. Two haplotypes of T. sinuaticollis plus the. monserratensis form another subgroup (D2) of closelyelated haplotypes (COII: 0.74%; 16S: 0.21%), but a. sinuaticollis haplotype is clearly part of the formerlade (is closest to T. interstitialis from Caramany).his could either be the product of introgressive geneow, as the samples come from populations in closeroximity, or be due to retention of an ancestral polymor-hism (Avise, 1994, and references therein). Data onuclear sequences and more detailed sampling shouldistinguish between these two possibilities. Anotherubgroup (D3) is constituted by quite divergent haplo-ypes (maximum of 4.24% for COII) belonging to Ibe-ian endemics reaching the southwest of France. Theaplotypes of this taxa group appear to be paraphyletic,espite the morphological divergence from the nominalpecies from which they originated. Some of the closestaplotypes (T. gougeleti from A Coruna and T. perezi

rom Zaragoza: 1.63% for COII and 0.62% for 16S; T.yanescens from Cueto and T. perezi from Villanubla:.20% for COII and 0.41% for 16S) are found inopulations that are very distant (more than 600 km)r are geographically split by high mountains; so anyresent contact between them probably has to beiscarded. Therefore, the apparent conflict betweenaxonomy and mitochondrial divergence can be betterxplained by past gene flow in relatively recent evolu-ionary time, although we cannot exclude the possibil-ty of lack of congruence between morphology andenotype due to incomplete lineage sorting. The evolu-ionary scenario for the T. goettingensis species complexs compatible with an ancestral lineage having a con-inuous geographic distribution that was progressively
ragmented by orogenic and glaciation events during o
he Pleistocene, producing local diversification, specia-ion, and possibly secondary-contact zones.

Molecular Clock for Dating the TimarchaDiversification

The use of the relative rate test (Sarich and Wilson,973) showed no significant (5% level) deviations fromconstant rate (not shown) at different divergence

evels of COII and 16S Timarcha sequences. Rateonstancy was further assessed by the more stringent‘two-cluster test’’ according to Takezaki et al. (1995). Inhis case, using the total number of substitutionsorrected by the Kimura two-parameter method, somef the basal lineages showed rates that were signifi-antly different from the average. When using onlyransversion rates, statistically significant rate discrep-ncies (5% level) were still obtained for the lineages of. marginicollis, T. pimelioides, T. strangulata, T.alearica, and T. espanoli.As no fossil evidence is known, date estimates for the

eparation of the Timarcha lineages have to be basedn biogeographical considerations. The Messinian salin-ty crisis (about 5.5 MYA) (Hsu et al., 1977; Maldonado,985; Jaeger, 1994) has been invoked to explain therigin of the present-day distribution of several organ-sms in the western Mediterranean region, such as thereshwater adaptation of gobiid fishes or the disjunctarabus (Coleoptera) distributions recently reportedy Penzo et al. (1998) and Pruser and Mossakowski1998), respectively. In the latter study, using flightlesseetles of the genus Carabus and ND1 sequences,olecular clock calibration with the Messinian geologi-

al event gave rates below 1 3 1028 substitution/site/ear (Pruser and Mossakowski, 1998). This rate islower than the estimated mitochondrial average of 2 3028 substitution/site/year for Hawaiian DrosophilaND1) (DeSalle et al., 1987) and several Arthropoda2.3 3 1028 substitution/site/year from RFLP, sequencend DNA–DNA hybridization data; Brower, 1994).Morphological-karyological evidence shows that three

pecies, T. punctella–T. rugosa and T. espanoli, arelearly derived from the same ancestor and their distri-ution is likely to be the product of vicariance caused byhe opening of the Gibraltar strait after the Mesinianalinity crisis. Using this biogeographical event separat-ng the two clades tentatively fixed at 5.3 Myr, we havebtained rate estimates of 0.76 3 1028 substitution/site/ear for COII and 0.45 3 1028 substitution/site/year for6S. These Timarcha estimates are 2.6 to 5 times lowerhan the DeSalle et al. (1987) or Brower (1994) esti-ates, depending on use of COII or 16S sequence

ivergences, respectively. Alternatively, if we apply thetandard 2.3 3 1028 rate, then we should assume auch more recent gene flow across the Gibraltar strait

or both Timarcha and Carabus populations. More dataased on several mitochondrial genes of a range of
rganisms showing different dispersal capacities are

nante

T

mtrpOt1rqrtcg

ipodcna(ooGcwTcdactTndf

TotrscdstsP

tedtrseapt

dtPtasslt

DScMCmavPS

A

B

B

B

B

C

C

C

C


ecessary to assess the significance of these rate discrep-ncies. Also, independent biogeographical evidence iseeded to test whether the Mesinian salinity crisis ishe main paleogeographical event explaining the pres-nt distribution of south Mediterranean organisms.

racing Ancestrality of Host-Plant Affiliation

Herbivorous beetles have been increasingly used asodel organisms to address the question of specializa-

ion on resource use (Kelley and Farrell, 1998, andeferences therein). Several studies estimating thehylogenies of chrysomelid genera, such as Oreina,phraella, and Chrysolina, show the relationship be-

ween phylogeny and food specialization (Funk et al.,995; Mardulyn et al., 1997; Garin et al., 1999). In aecent study, Kelley and Farrell (1998) asked theuestion whether specialized taxa tend to be derived,epresenting an evolutionary ‘‘dead end.’’ They showedhat a Dendroctonus bark-beetle COI phylogeny isompatible with the ancestrality of generalist species,iving rise to specialized taxa in the tips of the tree.Timarcha beetles can be considered as trophic special-

sts, as some species feed on one or a few closely relatedlant species of a particular botanical family, whilethers use single species belonging to different families,epending on their availability. As examples of the lastategory, T. lugens feeds on Plantago nivalis (Plantagi-aceae) and Veronica fruticulosa (Scrophulariaceae) atltitudes below 2700 m, shifting to Alyssum spinosumBrassicaceae) at higher altitudes (Jolivet, 1954). Forther species, such as T. maritima, trophism dependsn the seasonal availability of their two host plants,alium arenarium (Rubiaceae) as the habitual trophic

hoice and Plantago lanceolata (Plantaginaceae) ininter, when the former is not available (Chevin andiberghien, 1968). This strongly suggests that Timar-ha food specialization is dependent on geographicalistribution and hence on availability of host plants,lthough historical factors, such as feeding on chemi-ally similar hosts, have been proposed for other phy-ophagous beetles (Becerra, 1997). This implies that aimarcha species feeding on one plant family is notecessarily more restricted than another species withifferent populations feeding on two or more differentamilies in their distribution area.

The molecular phylogenetic hypothesis obtained forimarcha allows inference of the evolutionary historyf host-plant affiliation for this genus. A characterracing of Timarcha family host plants on a majority-ule tree obtained using the combined COII 1 16S dataet is shown in Fig. 5B. In this analysis we haveonsidered only the trophic affiliations that are well-ocumented in the literature, and character status is ofpecialization at the family level. Character optimiza-ion using either accelerated or delayed transformationhowed the clear basal affiliation to Rubiaceae for the
alearctic Timarcha. None of the host-plant shifts from
his food regime appear as single events in the phylog-ny but are products of convergence or reversal. Theerived condition in this genus seems therefore to behe widening of the ecological niche, probably as aesult of selection pressure imposed by differences ineasonal or altitudinal availability of the appropriatedible plants. We suggest that this could be interpreteds transitional food regimes occurring in marginalopulations in the species’ distribution area changingoward new specialized food regimes.

In conclusion, these results show that the mitochon-rial phylogeny of Timarcha allows one to address suchopics as the ancestrality of Metallotimarcha in thealearctic lineages, the compatibility with current sys-

ematic proposals, and the evolution of host-plantffiliation. The use of other sequences with differentubstitution rates, including nuclear sequences, andampling of the New World and Eastern Europeanineages should provide a more complete picture of theime scale and evolution of the genus.

ACKNOWLEDGMENTS

We thank F. Angelini, A. Baselga, G. Costa, J. de Diego, K.esender, F. Fritzlar, J. M. Guzman, F. Murria, P. Oromı, F. Polese, G.abella, J. M. Salgado, P. Verdyck, E. Vives, J. Vives, and S. Zoia forollecting some of the specimens. We also thank C. Garin, D. Jaume,. Palmer, and J. Pons for helpful discussions. A. V. Z. Brower and B.. Emerson kindly read a previous version of the manuscript andade helpful comments. We are indebted to R. DeSalle, B. D. Farrell,

nd an anonymous referee for comments that improved the finalersion of the paper. This work has been financed by Project DGESB96-0090 and J. G.-Z. has been supported by a grant from thepanish Ministry of Education and Culture.

REFERENCES

vise, J. C. (1994). ‘‘Molecular Markers: Natural History and Evolu-tion,’’ Chapman & Hall, New York.ecerra, J. X. (1997). Insects on plants: Macroevolutionary chemicaltrends in host use. Science 276: 253–256.

echyne, J. (1948). Contribution a la connaissance du genre Timar-cha Latr. 12: Etudes phylogenetiques et zoogeographiques (Col.Phytophaga, Chrysomelidae). Sbor. Narod. Mus. Praze 4: 1–62.

eckenbach, A. T., Wei, Y. W., and Liu, H. (1993). Relationships in theDrosophila obscura species group inferred from mitochondrialcytochrome oxidase II sequences. Mol. Biol. Evol. 10: 619–634.

rower, A. V. Z. (1994). Rapid morphological radiation and conver-gence among races of the butterfly Heliconius erato inferred frompatterns of mitochondrial DNA evolution. Proc. Natl. Acad. Sci.USA 91: 6491–6495.apaldi, R. A. (1990). Structure and function of cytochrome c oxidase.Annu. Rev. Biochem. 59: 569–596.hevin, H., and Tiberghien, G. (1968). Existence d’un regime alimen-taire mixte chez Timarcha maritima Perris (Col. Chrysomelidae).Bull. Soc. Entomol. Fr. 73: 203–206.

lary, D. O., and Wolstenholme, D. R. (1985). The ribosomal RNAgenes of Drosophila mitochondrial DNA. Nucleic Acids Res. 13:4029–4045.

rowson, R. A. (1981). ‘‘The Biology of the Coleoptera,’’ Academic
Press, London.

C

D

D

D

E

F

F

F

F

F

F

G

G

G

G

G

H

H

H

I

J

J

J

J

J

J

K

L

L

M

M

M

M

M

P

P

P

P

P

P

P

R


unningham, C. W. (1997). Can three incongruence tests predictwhen data should be combined? Mol. Biol. Evol. 14: 733–740.e Bruijin, M. H. L. (1983). Drosophila melanogaster mitochondrialDNA, a novel gene organization and genetic code. Nature 304:234–241.e Rijk, P., Van de Peer, Y., and De Wachter, R. (1996). Database onthe structure of large ribosomal subunit RNA. Nucleic Acids Res.24: 92–97.eSalle, R., Freedman, T., Prager, E. M., and Wilson, A. C. (1987).Tempo and mode of sequence evolution in mitochondrial DNA ofHawaiian Drosophila. J. Mol. Evol. 26: 157–164.

merson, B. C., and Wallis, G. P. (1995). Phylogenetic relationships ofthe Prodontria (Coleoptera; Scarabaeidae; Subfamily Melolon-thinae), derived from sequence variation in the mitochondrialCytochrome Oxidase II gene. Mol. Phylogenet. Evol. 4: 433–447.

arrell, B. D. (1998). ‘‘Inordinate Fondness’’ explained: Why are thereso many beetles? Science 281: 555–559.

arris, J. S., Kallersjo, M., Kluge, A. G., and Bult, C. (1994). Testingsignificance of incongruence. Cladistics 10: 315–319.

rati, F., Simon, C., Sullivan, J., and Swofford, D. L. (1997). Evolutionof the mitochondrial Cytochrome Oxidase II gene in Collembola. J.Mol. Evol. 44: 145–158.

rench, S., and Robson, B. (1983). What is a conservative substitu-tion? J. Mol. Evol. 19: 171–175.

u, Y.-X., and Li, W.-H. (1993). Statistical tests of neutrality ofmutations. Genetics 133: 693–709.

unk, D. J., Futuyma, D. J., Ortı, G., and Meyer, A. (1995). Mitochon-drial DNA sequences and multiple data sets: A phylogenetic studyof phytophagous beetles (Chrysomelidae: Ophraella). Mol. Biol.Evol. 12: 627–640.arin, C. F., Juan, C., and Petitpierre, E. (1999). Mitochondrial DNAphylogeny and the evolution of host-plant use in Palearctic Chryso-lina (Coleoptera, Chrysomelidae) leaf beetles. J. Mol. Evol. 48:435–444.atesy, J., DeSalle, R., and Wheeler, W. (1993). Alignment-ambigu-ous nucleotide sites and the exclusion of systematic data. Mol.Phylogenet. Evol. 2: 152–157.omez-Zurita, J., Garin, C. F., Juan, C., and Petitpierre, E. (1999).Mitochondrial 16S rDNA sequences and their use as phylogeneticmarkers in leaf-beetles with special reference to the subfamilyChrysomelinae. In ‘‘Advances in Chrysomelidae Biology,’’ (M. L.Cox, Ed.), pp. 25–38. Backhuys Publ. Oegstgeest, The Netherlands.utell, R. R., and Fox, G. E. (1988). A compilation of large subunitRNA sequences presented in a structural format. Nucleic AcidsRes. 16(Suppl.): r175–r269.utell, R. R., Larsen, N., and Woese, C. R. (1994). Lessons from anevolving rRNA: 16S and 23S rRNA structures from a comparativeperspective. Microbiol. Rev. 58: 10–26.an, H.-Y., and McPheron, B. A. (1997). Molecular phylogeneticstudy of Tephritidae (Insecta: Diptera) using partial sequences ofthe mitochondrial 16S ribosomal DNA. Mol. Phylogenet. Evol. 7:17–32.endy, M. D., and Penny, D. (1989). A framework for the quantitativestudy of evolutionary trees. Syst. Zool. 38: 297–309.su, J. K., Montadert, L., Bernoulli, D., Cita, M. B., Erickson, A.,Garrison, R. E., Kidd, B. R., Melieres, F., Muller, C., and Wright, R.(1977). History of the Mediterranean salinity crisis. Nature 267:399–403.

ablokoff-Khnzorian, S. M. (1966). Considerations sur l’edeage desChrysomelidae et son importance phylogenique. L’Entomologiste22: 115–136.

aeger, J.-J. (1994). The evolution of biodiversity among the south-west European neogene rodent (Mammalia, Rodentia) communi-ties: Pattern and process of diversification and extinction. Palaeo-
geogr. Palaeoclimatol. Palaeoecol. 111: 305–336.
olivet, P. (1954). Notes ecologiques preliminaires sur les Chrysomeloi-dea de la Sierra Nevada. Bull. Ann. Soc. Ent. Belgique 90: 69–72.

olivet, P. (1994). Remarks on the biology and biogeography ofTimarcha (Chrysomelidae: Chrysomelinae). In ‘‘In Proceedings ofthe Third International Symposium on the Chrysomelidae, Beijing1992’’ (D. G. Furth, Ed.), pp. 85–97. Backhuys Publ., Leiden.

olivet, P., and Petitpierre, E. (1973). Plantes-hotes connues desTimarcha Latreille (Col. Chrysomelidae). Quelques considerationssur les raisons possibles du trophisme selectif. Bull. Soc. Entomol.Fr. 78: 9–25.

olivet, P., and Hawkeswood, T. J. (1995). ‘‘Host-Plants of Chrysomeli-dae of the World,’’ Backhuys Publ., Leiden.

uan, C., Vazquez, P., Rubio, J. M., Petitpierre, E., and Hewitt, G. M.(1993). Presence of highly repetitive DNA sequences in Triboliumflour-beetles. Heredity 70: 1–8.elley, S. T., and Farrell, B. D. (1998). Is specialization a dead-end?The phylogeny of host use in Dendroctonus bark beetles (Scolyti-dae). Evolution 52: 1731–1743.

i, W.-H. (1993). Unbiased estimation of the rates of synonymous andnonsynonymous substitution. J. Mol. Evol. 36: 96–99.

iu, H., and Beckenbach, A. T. (1992). Evolution of the mitochondrialCytochrome Oxidase II gene among 10 Orders of insects. Mol.Phylogenet. Evol. 1: 41–52.addison, W. P., and Maddison, D. R. (1992). ‘‘MacClade: Analysis ofPhylogeny and Character Evolution, Version 3.0,’’ Sinauer, Sunder-land, MA.addison, W. P., and Slatkin, M. (1991). Null models for the numberof evolutionary steps in a character on a phylogenetic tree. Evolu-tion 45: 1184–1197.aldonado, A. (1985). Evolution of the Mediterranean basins and adetailed reconstruction of the Cenozoic paleoceanography. In ‘‘KeyEnvironments: Western Mediterranean’’ (R. Margalef, Ed.), pp.17–59. Pergamon Press, Oxford.ardulyn, P., Milinkovitch, M. C., and Pasteels, J. M. (1997).Phylogenetic analyses of DNA and allozyme data suggest thatGonioctena leaf beetles (Coleoptera; Chrysomelidae) experiencedconvergent evolution in their history of host-plant family shifts.Syst. Biol. 46: 722–747.ickevich, M. F., and Farris, J. S. (1981). The implications ofcongruence in Menidia. Syst. Zool. 30: 351–370.

enzo, E., Gandolfi, G., Bargelloni, L., Colombo, L., and Patarnello, T.(1998). Messinian salinity crisis and the origin of freshwaterlifestyle in Western Mediterranean gobies. Mol. Biol. Evol. 15:1472–1480.

etitpierre, E. (1970a). Variaciones morfologicas y de la genitalia enlas Timarcha Lat. (Col. Chrysomelidae). P. Inst. Biol. Apl. 48: 5–16.

etitpierre, E. (1970b). Cytotaxonomy and evolution of TimarchaLatr. (Coleoptera: Chrysomelidae). Genet. Iber. 22: 67–120.

etitpierre, E. (1973). ‘‘Estudios Sistematicos, Citogeneticos y Evolu-tivos Sobre el Genero Timarcha (Coleoptera, Chrysomelidae),’’ PhDthesis, Publ. Univ. of Barcelona.

etitpierre, E. (1976). Further cytotaxonomical and evolutionarystudies on the genus Timarcha Latr. (Coleoptera: Chrysomelidae).Genet. Iber. 28: 57–81.

etitpierre, E., Segarra, C., Yadav, J. S., and Virkki, N. (1988).Chromosome numbers and meioformulae of Chrysomelidae. In‘‘Biology of Chrysomelidae’’ (P. Jolivet, E. Petitpierre, and T. H.Hsiao, Eds.), pp. 161–186. Kluwer Academic, Dordrecht, TheNetherlands.

ruser, F., and Mossakowski, D. (1998). Low substitution rates inmitochondrial DNA in Mediterranean carabid beetles. Insect Mol.Biol. 7: 121–128.eid, C. A. M. (1995). A cladistic analysis of subfamilial relationshipsin the Chrysomelidae sensu lato (Chrysomeloidea). In ‘‘Biology,
Phylogeny, and Classification of Coleoptera: Papers Celebrating

R

S

S

S

S

T

T

T

W


the 80th Birthday of Roy A. Crowson’’ (J. Pakaluk and S. A.Slipinski, Eds.), pp. 559–631. Muzeum i Instytut Zoologii PAN,Warszawa.

ozas, J., and Rozas, R. (1997). DnaSP version 2.0: A novel softwarepackage for extensive molecular population genetics analysis.Comput. Appl. Biosci. 13: 307–311.

arich, V. M., and Wilson, A. C. (1973). Generation time and genomicevolution in primates. Science 179: 1144–1147.

imon, C., Frati, F., Beckenbach, A. T., Crespi, B., Liu, H., and Flook,P. (1994). Evolution, weighting, and phylogenetic utility of mitochon-drial gene sequences and a compilation of conserved polymerasechain reaction primers. Ann. Entomol. Soc. Am. 87: 651–701.

teinhausen, W. R. (1994). Larvae of Palaearctic Timarcha Latreille.In ‘‘Novel Aspects of the Biology of Chrysomelidae’’ (P. H. Jolivet,M. L. Cox, and E. Petitpierre, Eds.), pp. 119–125. Kluwer Aca-demic, Dordrecht, The Netherlands.

tockmann, R. (1966). Etude de la variabilite de quelques especesfrancaises du genre Timarcha Latreille (Col. Chrysomelidae). Ann.Soc. Entomol. Fr. (N. S.) 2: 105–126.

ajima, F. (1989). Statistical method for testing the neutral mutationhypothesis by DNA polymorphism. Genetics 123: 585–595.

akezaki, N., Rzhetsky, A., and Nei, M. (1995). Phylogenetic test ofthe molecular clock and linearized trees. Mol. Biol. Evol. 12:823–833.

hompson, J. D., Higgins, D. J., and Gibson, T. J. (1994). CLUSTALW: Improving the sensitivity of progressive multiple sequencealignment through sequence weighting position-specific gappenalties and weight matrix choice. Nucleic Acids Res. 22: 4673–4680.heeler, W., and Gladstein, D. (1992–1994). ‘‘MALIGN: Programand Documentation-version 1.9,’’ Am. Mus. Nat. Hist., NewYork.

the evolutionary history of the genus timarcha (coleoptera, chrysomelidae) inferred from...

Documents