JAE 013 "1999# J[ Appl[ Ent[ 013\ 066Ð072 "1999# Þ 1999\ Blackwell Wissenschafts!Verlag\ Berlin ISSN 9820!1937

Genetic population structure in olive fly Bactrocera oleae (Gmelin): gene flow and patterns of geographic differentiation M[ D[ Ochando and A[ Reyes Departamento de Genetica\ Universidad Complutense\ Madrid\ Spain

Abstract] Four wild Spanish populations of Bactrocera oleae "Gmelin# "Dipt[\ Tephritidae# were studied by gel elec! trophoresis\ in order to _nd out their population genetic structure and the relative importance of gene ~ow in that structure[ A great amount of genetic variability was observed[ In terms of the pattern of the genetic variability in our populations\ the most noticeable characteristics were the similarity of the con_guration of allelic frequencies from locality to locality\ the existence of exceptions and quantitative di}erences within that similarity\ and the heterogeneity among the variable loci[ Gene ~ow estimated from Wright|s method gave a value of Nm  7[8[ However\ the FST statistic in three of the loci ranged from 9[9340 to 9[9519 indicating a substantial degree of di}erentiation among populations "P ³ 9[990#[ The present observations seem to support the hypothesis that although gene ~ow could be high "Nm  7[8#\ {natural| selection\ probably due to agricultural practices\ may be the major factor responsible for the pattern of the genetic variability observed in B[ oleae[

0 Introduction The study of insect pests has changed dramatically in the last two decades[ Genetics\ and more recently molecular biology\ have increased their roles and in~uence in top! ics associated with agricultural pests[ Until recently most applied biologists would probably have doubted the relevance of these areas of knowledge to problems of pest control in agriculture and forestry "MENKEN and ULENBERG\ 0876^ LOXDALE and HOLLANDER\ 0878#[ However\ an ever!increasing number of articles cover! ing biological material about pests and including genetic methods can be found in the specialized literature "ROB! INSON and HOOPER\ 0878^ MCPHERON and STECK\ 0885#[ Unfortunately\ these studies are not numerous enough to cover more than a part of the large range of insect pests[ The olive fruit ~y\ Bactrocera oleae\ is a harmful pest in many temperate countries[ However\ despite its agro! economic importance\ genetic studies are scarce and furthermore\ the genetic structure of its populations is completely unknown[ Only a few mainly {descriptive| studies of these populations are reported in the literature "ZOUROS and KRIMBAS\ 0858^ TSAKAS and KRIMBAS\ 0864^ BUSH and KITTO\ 0868^ TSAKAS and ZOUROS\ 0879^ LOUKAS et al[\ 0874^ reviews of ZOUROS and LOUKAS\ 0878^ and LOUKAS\ 0878^ OCHANDO et al[\ 0883#[ There are two aspects of the biology of B[ oleae that make this species of special biological interest\ in addition to its status as an important pest^ _rst\ it is strictly monophagous\ and second\ its extended com! mensialism to man\ given its almost exclusive depen! dence on olive cultivation\ which puts it at the mercy of agricultural practices[ Taking these considerations into account\ the objec! tive of this study was to investigate the genetic varia! U[ S[ Copyright Clearance Center Code Statement]

bility present in Spanish populations of Bactrocera oleae[ Using gel electrophoresis\ we have attempted to obtain information about the inherited inter! and intra! population variation[ This could\ in turn\ make it poss! ible to gain insights into the genetic structure of the populations of this special insect\ into the gene ~ow in~uence in that structure\ as well as into other possible evolutionary processes involved in the maintenance of its genetic variability[ Understanding the genetic struc! ture of populations of olive ~y can facilitate a more e.cient and environmentally safe _ght against the pest[

1 Material and methods 1[0 Insects Four natural populations of B[ oleae were studied\ namely] QUI collected from central Spain "El Pardo\ Madrid prov! ince#\ an area having a continental climate^ ATA\ from the south!east area "Atajate\ Malaga province# with a mild and relatively humid climate^ and SER and FEL populations\ both from western Spain "Serradilla\ Caceres province\ and Valle de la Serna\ Badajoz province\ respectively#\ which have a dry climate\ extreme in the summer\ and mild in the winter[ Collections were made by harvesting several hundred infested olive fruits and allowing the larvae to pupate in the laboratory[ The ~ies analysed were always those obtained directly from the fruits[

1[1 Electrophoresis techniques Standard techniques for horizontal starch gel electrophoresis and assay of enzymes were applied to adult ~ies\ according to AYALA et al[ "0861# with minor changes[ The bu}er system was] gel bu}er] 65 mM TRIS\ 4 mM citric acid\ pH  7[54^ electrode bu}er] 299 mM boric acid\ 59 mM NaOH\ pH  7[0 "POULIK\ 0846^ AYALA et al[\ 0861#[ Nine enzymatic loci "chosen at random# were studied] alde!

9820Ð1937:1999:1392Ð9066 , 04[99:9

067 hyde!oxidase "Ao#\ acid phosphatase dehydrogenase "Aph#\ two esterase loci "Est!0 and Est!1#\ fructokinase "Fk#\ hexok! inase "Hk#\ hydroxybutyrate!dehydrogenase "Hbdh#\ malate dehydrogenase "Mdh# and phosphoglucomutase "Pgm#[ A Drosophila melanogaster monomorphic strain "X8sod#\ was used as control in all gels[ This strain has been used as ref! erence {099| for allele designation in all cases[ A sample size of 099 "or as close as possible# was used in most of the cases[

1[2 Data analysis Three classical statistics to quantify variability were used] P\ polymorphism\ the proportion of polymorphic loci "two criteria of polymorphism were used\ the 84) and the 88) criterion#^ H\ heterozygosity\ the average number of het! erozygous individuals^ and n\ the average number of alleles per locus[ Departures from the HardyÐWeinberg equilibrium were tested by means of chi!square of heterozygote adjustments\ in addition to the calculation of WRIGHT|s "0840^ 0867# FIS and FIT statistics[ The heterogeneity among populations of di}erent geo! graphical origin was measured by calculation of the _xation indices[ These parameters "FIS\ FIT and FST# were introduced by WRIGHT "0832\ 0840# as {inbreeding coe.cients|\ but they have been generalized to describe the genetic structure of populations "WRIGHT\ 0854\ 0867^ NEI\ 0866\ 0876^ NEI and CHESSER\ 0872# and o}er a convenient means of sum! marizing population structure[ Genetic identities and distances through the Nei method "NEI\ 0861\ 0867#\ provide insights into the global divergence between every two populations\ and at the same time\ allow the construction of a dendrogram[ With regard to gene ~ow Wright|s method was used "derived from Wright|s _xation indices\ WRIGHT\ 0832\ 0840\ 0858#[ Thus\ gene ~ow "Nm# was estimated from FST values using the relationship FST 3 0:"3Nm ¦ 0#\ that is to say\ Nm  "0 Ð FST#:3FST\ in which N is the e}ective size of the population and m is the proportion of the population that are migrants[

2 Results Table 0 displays the amount of variability found at each of the nine loci studied in the four populations[ The allelic frequencies\ the proportion of heterozygous indi! viduals "both\ observed and expected from HardyÐ Weinberg equilibrium#\ and the number of genomes sampled\ are shown[ Several features are noticeable in table 0[ First\ variation at loci range from highly poly! morphic\ such as Ao\ Est!1\ and Pgm\ to almost com! pletely monomorphic\ such as Fk and Hk^ second\ the same allele has become _xed or represents the highest frequency in the four populations\ with the exception of Ao in QUI population[ However\ the gene frequencies were not uniform in the di}erent populations studied[ On the other hand\ the observed and expected fre! quencies of heterozygotes do not agree in most cases as the four populations show equilibrium frequencies only for the Pgm locus and for Ao and Hbdh\ two of the populations do so[ In the other cases\ there are 07 out of 15 heterozygous frequencies observed which are far from the HardyÐWeinberg equilibrium[ Moreover\ in all the cases except one "locus Ao in QUI population#\ the lack of agreement is due to a de_ciency of het! erozygotes[

M[ D[ Ochando and A[ Reyes

Table 0[ Allelic frequencies\ observed "HO# and expected "HE# heterozygous frequencies\ and sample size "N# at nine allozymic loci in four Spanish populations of Bactrocera oleae Populations ******************* Locus Allele ATA SER FEL QUI * ***************************** Ao 0 9[9599 9[9163 9[9299 9[9173 1 9[5649 9[6592 9[5849 9[3194 2 9[9199 9[9957 9 9[9173 3 9[1249 9[1944 9[1649 9[4116 4 9[9099 9 9 9 9[26990 9[2040 9[3699 9[54800 HO HE 9[4340 9[2678 9[3394 9[4373 N 099 62 099 77 Aph 0 9[7499 9[7294 9[6231 9[6446 1 9[0349 9[0584 9[1547 9[1275 2 9[9949 9 9 9 3 9 9 9 9 HO 9[90991 91 91 9[90031 HE 9[1454 9[1704 9[2892 9[2619 N 099 48 68 77 0 9 Ð 9[9985 Ð Est0 1 9[0043 Ð 9[2831 Ð 2 9[7735 Ð 9[4851 Ð HO 91 Ð 9[06201 Ð 9[1931 Ð 9[3780 Ð HE N 41 Ð 41 Ð 0 9 9[9956 9 9 Est1 1 9[9949 9[9222 9[9949 9 2 9[0149 9[2599 9[0499 9[0085 3 9[7099 9[4556 9[6649 9[7150 4 9[9599 9[9222 9[9699 9[9432 9[01991 9[97991 9[08991 9[02931 HO HE 9[2136 9[4369 9[2619 9[2992 N 099 64 099 81 Fk 0 0 9[8645 9[8899 0 1 9 9[9133 9[9099 9 N 099 30 099 85 Hbdh 0 9[9799 9[9667 9[9769 9[9571 1 9[8199 9[8111 9[7802 9[8037 2 9 9 9[9106 9[9069 9[91991 9[9556 9[0293 9[94572 HO HE 9[0361 9[0324 9[0864 9[0471 N 099 34 12 77 Hk 0 9[9099 9 9 9 1 9[8899 0 0 0 N 099 59 099 85 Mdh 0 9[9599 9[0675 9[0099 9[2957 1 9[1299 9[2817 9[2999 9 2 9[5199 9[3175 9[4749 9[5694 3 9[9899 9 9 9[9116 4 9 9 9[9949 9 91 91 9[90991 91 HO HE 9[4409 9[5290 9[4446 9[3447 N 099 73 099 77 Pgm 0 9 9 9[9941 9 1 9[9949 9[9008 9[9941 9[9094 2 9[9199 9[9959 9[9093 9[9094 3 9[7799 9[7984 9[7847 9[7626 4 9[9849 9[0615 9[9722 9[0942 9[0799 9[1499 9[0660 9[1205 HO HE 9[1051 9[2036 9[0893 9[1142 N 099 73 85 84 0 2

signi_cant at P ³ 9[94 level[1 signi_cant at P ³ 9[990 level[ Ð\ no data[

Genetic population structure in olive fruit ~y

Table 1[ Genetic variation in the Spanish populations of Bactrocera oleae] Heterozygosity "H#\ mean number of alleles per locus "n#\ and polymorphism "P#\ at 84) and 88) criteria P ************ H n 84) 88) * ***************************** ATA 9[9667 2[9 9[6667 9[7778 SER 9[9789 1[8 9[5556 9[6667 FEL 9[0167 2[9 9[6667 9[7778 QUI 9[0251 1[7 9[6667 9[6667 X 9[0966 1[8 9[6499 9[7222 Þ

068

Table 3[ Nei|s genetic identities "I\ above the diagonal# and genetic distances "D\ below the diagonal# in four Spanish populations of Bactrocera oleae ATA SER FEL QUI * ***************************** ATA Ð 9[8701 9[8732 9[8659 SER 9[9089 Ð 9[8740 9[8383 FEL 9[9047 9[9049 Ð 9[8653 QUI 9[9132 9[9419 9[9128 Ð

3 Discussion 3[0 Amount of variability Table 2[ Estimates of F!statistics on variable loci Loci FIS FIT FST * ***************************** Ao 9[9153 9[9757 9[95190 Aph 9[8728 9[8739 9[9937 9[5548 9[5709 9[93400 Est1 Hbdh 9[4729 9[4670 Ð9[9006 Mdh 9[8844 9[8847 9[94800 Pgm 9[0082 9[0130 9[9943 X 9[4512 9[4649 9[9163 Þ 0

signi_cant at P ³ 9[990 level[

Table 1 summarizes the amount of variability using the three usual statistics] H\ n and P[ The most remark! able characteristic in table 1 is the high degree of varia! bility shown\ especially with respect to polymorphism "64[99 or 72[22)# and number of alleles per locus "1[8#[ The mean heterozygosity for all the four populations is 9[0966[ In table 2 the _xation indices for the six variable loci are presented[ Heterogeneity is the predominating fea! ture in this table[ The FIS and FIT values make evident the general de_ciency of heterozygotes "positive values#\ especially in the case of Aph and Mdh loci\ and also in those of Est!1 and Hbdh[ In the case of FST\ what stands out is the high diversity\ from zero values\ or not sig! ni_cantly di}erent from zero "such as is the case for the Hbdh locus\ or those of the Aph and Pgm loci#\ to values of the order of 09 times greater\ which can be considered signi_cant of clear di}erentiation "such as in the case of the Ao\ Est!1 and Mdh loci#[ According to WRIGHT "0867#\ values greater than 9[94 indicate substantial gen! etic di}erentiation among populations[ The genetic distances\ D "NEI\ 0861\ 0867# between all the pairwise combinations of the four populations\ based on the studied loci\ are shown in table 3[ Clearly\ the QUI population is distinct from the other three\ and SER and FEL are the most similar[ The Nm values were also calculated to investigate the possible in~uence of gene ~ow in the distribution of the variability found[ The estimated value is 7[8\ using Wright|s formula[

Although only nine loci have been assayed\ these may be considered to be a random sample of the genome[ SINGH and RHOMBERG "0876#\ who have compiled extensive variability and species data\ conclude that doubling the number of studied loci has no signi_cant e}ect on the proportion of polymorphic loci or on the mean heterozygosity "although they do make reference to a higher number of loci#[ Therefore\ taking this opi! nion and also the fact that a large sample size "about 099 individuals# was used\ it is felt that the present data are representative of the real genetic variation in B[ oleae[ The olive fruit ~y is an extremely monophagous spec! ies[ According to some authors "POWELL\ 0860^ LEVIN! TON\ 0862^ GILLESPIE and LANGLEY\ 0863^ YONG\ 0881#\ one would expect a positive correlation between the genetic variation of a species and the degree of its environmental diversity[ However\ comparison of the genetic variability levels present in the populations of the olive fruit ~y studied "tables 0 and 1#\ with those present in other fruit ~ies "Diptera\ Tephritidae#\ pro! vides the evidence that genetic variability in B[ oleae does not follow that expectation[ Thus\ available data on Ceratitis capitata "Wiedeman# "HUETTEL et al[\ 0879^ MORGANTE et al[\ 0870^ GASPERI et al[\ 0875\ 0880^ LOUKAS\ 0878^ MILANI et al[\ 0878^ MALACRIDA et al[\ 0881^ REYES and OCHANDO\ 0883#\ which is an extremely polyphagous species\ show values of het! erozygosity\ polymorphism\ and number of alleles per locus that are much lower\ in general\ than the values obtained on B[ oleae[ Information available on other Tephritidae\ such as Anastrepha fraterculus "Wiedeman# and the genus Rhagoletis "Dipt[\ Tephritidae# "BERLO! CHER and BUSH\ 0871^ MALAVASI and MORGANTE\ 0872^ MCPHERON et al[\ 0877^ MCPHERON\ 0889# once again serves to underscore the idea of the high degree of variability shown by the olive fruit ~y\ especially with reference to polymorphism and number of alleles per locus[ Comparison with other species of the group\ such as Bactrocera cucurbitae "Coquillett#\ which has a wide host range\ or Bactrocera umbrosus "Fabricius# and Bac! trocera albistrigata "de Meijere#\ both with a host range restricted but higher than that of B[ oleae "YONG\ 0877\ 0889\ 0881# also indicates a higher degree of variability in B[ oleae[ The question which subsequently arises is whether these high values are typical of the species or\ on the

079

M[ D[ Ochando and A[ Reyes

contrary\ whether our populations represent an excep! tional case[ Unfortunately\ not much data are available on the genetic variability of natural populations of B[ oleae[ Only a few studies can be found in the specialized literature that can be considered {populational|\ in which at least two populations or various loci have been studied "TSAKAS and KRIMBAS\ 0864^ BUSH and KITTO\ 0868^ TSAKAS and ZOUROS\ 0879^ LOUKAS et al[\ 0874^ reviews from LOUKAS\ 0878^ and ZOUROS and LOUKAS\ 0878^ OCHANDO et al[\ 0883#[ Speci_cally\ the mean values given by LOUKAS "0878# in his com! pilation are n  2[356 and H  9[077^ these values are reduced to n  1[427 and H  9[042\ if the biased high information obtained for Est!A and Est!B loci is elim! inated "the author does not give information on poly! morphism#[ Previous data "OCHANDO et al[\ 0883# on two Spanish populations of B[ oleae and 01 loci reveal values of] H  9[0228\ n  1[5707\ and P  9[6616 or 9[8434 "according to the criteria of polymorphism used#[ All these values are clearly similar to those obtained in the four olive ~y populations analysed[ Therefore\ the high rate of genetic variability observed seems to be typical of this species[ In summary\ the proposed hypothesis as to the poss! ible correlation between degree of environmental diver! sity and degree of genetic variability clearly does not hold in the strictly monophagous B[ oleae[ Other poss! ible explanations for the high genetic variation observed need to be explored^ but for this purpose\ more com! prehensive information on the complete range of dis! tribution of the species is needed[

related to mating selection or that all six are linked to loci of this type[ 3# Sampling of groups of individuals belonging to di}erent populations\ the Wahlund e}ect\ is also poss! ible[ However\ the present study involves working with a species with a very speci_c ecology\ and the samples were captured in uniform _elds and in limited areas[ In addition\ it is highly unlikely that in all four analysed populations we would be sampling as a unit a group that comes from various subpopulations[ 4# Inbreeding brought about by various causes is ano! ther possible explanation[ One of these causes is bottle! necks\ which appear to occur in the B[ oleae populations during the winter\ although the almost constant avail! ability of fruit throughout the year probably prevents a drastic e}ect on the population size[ Moreover\ at the time of sampling "the months of November and December# there have been several generations of expansion "NEUENSCHWANDER et al[\ 0875^ DE ANDRES\ 0880# due to the high availability of fruits[ Secondly\ the sampling of {families| is another cause of inbreeding\ but in the case of this species this does not appear probable because each female tends to lay only one egg in each fruit "NEUENSCHWANDER et al[\ 0875^ FLETCHER\ 0878a^ DE ANDRES\ 0880#^ this sampling was carried out at a time of high availability of fruits^ and\ in addition\ several hundreds of olives were col! lected in every case[ Furthermore\ the present popu! lations show an U!shaped allele frequency distribution corresponding to that expected in large populations "NEI et al[\ 0864^ HUETTEL et al[\ 0879#[

3[1 Patterns of variability

It is true that\ as SNYDER and LINTON "0873# point out\ signi_cant departures from panmixia for some loci joined with equilibrium distribution of genotypes to some others is not compatible with some of the potential causes discussed above for heterozygous de_ciency\ such as inbreeding or the Wahlund e}ect[ With respect to the interpopulational distribution of the variability detected\ the basic characteristic observed is that for all loci "except Ao\ in QUI popu! lation# the same allele is the one which shows the maximum frequency or is _xed in all the populations[ Moreover\ the allele which shows the second highest frequency "except Ao and Mdh for QUI population# is the same[ This similarity cannot be explained by random drift as drift e}ects should be uncorrelated in the di}er! ent populations[ Notwithstanding\ the existence of gene ~ow\ which has a tendency to make populations uniform\ cannot be ignored[ Therefore\ we attempted to infer the role played by gene ~ow in relation to the genetic structure of the populations[ In the case of B[ oleae\ available knowledge about migratory capacity comes from {direct| information\ through ecological studies\ by markÐreleaseÐrecapture\ as well as by laboratory experiments[ These studies have shown that B[ oleae is a species with a capacity for dispersal over great distances\ although normal behav! iour in the presence of abundant fruit trees is of low vagility "NEUENSCHWANDER et al[\ 0875^ FLETCHER\ 0878a\ b#[ Furthermore\ obstacles such as a road or a pine forest can stop their dispersal[ If the facts that

The _rst characteristic of note in the analysis of intra! populational variability in the present data is the lack of agreement with HardyÐWeinberg expectation "table 0#[ This situation is true of all four populations studied and applies to the majority of the variable loci[ Interestingly\ the lack of concordance between the values obtained and the expected values in a situation of HardyD!Wein! berg equilibrium is due in all cases except one "Ao in QUI population# to a de_ciency of heterozygotes "posi! tive Fi values\ table 2#[ Although other authors "SNYDER and LINTON\ 0873 in Simulidae^ MARINKOVIC et al[\ 0876^ and IZQUIERDO and RUBIO\ 0878 in Drosophila# have observed this same tendency in their results with other species of Diptera\ this does not appear to be a general situation[ The heterozygous de_ciency detected in these olive ~y populations could be due to di}erent causes] 0# Technical problems\ leading to lack of detection of electromorph classes[ Yet the technique has been used on multiple species of insects along with the fact that some heterozygotes were indeed detected\ and _nally\ the excess of homozygotes for some loci is not always general for all four populations for the same loci[ 1# The presence of null alleles is another possibility[ The fact that no null homozygotes were detected\ most probably discounts this possibility[ 2# Assortative mating can also lead to de_ciency of heterozygotes[ However\ it seems unlikely that six di}erent allozyme loci would show e}ects on characters

Genetic population structure in olive fruit ~y

excess dryness is a limiting element for this stenotrophic species\ and that the analysed populations are separated by hundreds of kilometres of dry climate "except SER Ð FEL#\ are considered then we would _rst expect the gene ~ow to be relatively scarce among the populations[ In fact\ it has been demonstrated that dispersal of indi! viduals and hence gene ~ow is over much shorter dis! tances than individuals are capable of moving "EHRLICH and RAVEN\ 0858^ SLATKIN\ 0874a\ 0876#[ We have quanti_ed gene ~ow between the popu! lations using Wright|s method "WRIGHT\ 0832#[ Sur! prisingly enough\ the resulting value can be considered high\ Nm  7[8[ This data seem to indicate that a high gene ~ow exists among the olive ~y populations studied\ in spite of what was suspected from previous ecological knowledge of the species and knowing that {potentially| it is capable of migrating long distances[ In spite of using common methods for calculation of the gene ~ow\ such as Slatkin|s {private alleles| "SLATKIN\ 0874b\ 0876# and that derived from Wright|s {_xation indices| "WRIGHT\ 0840#\ as an initial approach\ it is apparent that they unfortunately represent only a rough approxi! mation to reality "SLATKIN\ 0870\ 0874b^ DALY\ 0878#[ Therefore\ even when these limitations are kept in mind\ if the gene ~ow is e}ectively high\ we would expect a clear geographic uniformity of the variability[ The QUI population shows two evident exceptions "loci Ao and Mdh\ table 0# and in general\ among the present populations there are clear quantitative di}erences in the variables loci "table 0#[ If the gene ~ow is\ in e}ect\ common\ all the loci should experience the same stan! dardizing e}ect\ without exception[ According to SLATKIN "0876#\ species for which direct methods indicate little current gene ~ow but indirect methods indicate much higher levels\ probably have undergone large!scale demographic changes rela! tively frequently[ Gene!frequency di}erences may exist even if gene ~ow is common in species in which popu! lation extinctions and recolonizations are occurring[ Bactrocera oleae seems to go through a certain bottle! neck in the winter\ but it is not known to what extent\ and\ besides\ this probably does not imply an exterior recolonization of existing populations\ since all of them su}er the bottleneck more or less simultaneously[ Perhaps\ and this seems most probable to us and is in accordance with the other data already discussed "deviation from the HardyÐWeinberg equilibrium in intrapopulation variability discussion#\ gene ~ow could occur at a relatively high rate\ at least among three of the populations "ATA\ SER\ FEL#\ but local variation in selection intensity "perhaps caused by di}erences in insecticide use among farmers# might be strong enough to maintain di}erentiation among populations[ On the other hand\ the mean value of the FST _xation index is 9[9163 "table 2#\ which can be considered low[ This could mean that the interpopulational di}er! entiation is low\ on average\ and so the gene ~ow e}ect is strong[ However\ if the values for each locus are analysed\ great di}erences\ in some cases over 09\ are seen^ and furthermore\ for three of the six variable loci analysed\ the di}erentiation is statistically highly sig! ni_cant[ This high degree of heterogeneity among loci

070

is not explainable solely by the e}ect of such a high gene ~ow as that indicated by the statistical analysis[ The e}ects of gene ~ow should necessarily be con! sistent among all loci because gene ~ow\ as well as genetic drift\ both a}ect all loci and alleles in the same way[ Therefore\ neither of these processes can indi! vidually explain the present variability[ Thus\ if there is signi_cant heterogeneity among loci\ which is true in the present case\ this heterogeneity may be taken as evidence for selection "LEWONTIN and KRAKAUER\ 0862#[ The FST _xation index measures the di}erentiation of loci among multiple populations considered as a whole[ However\ it presents two disadvantages when glo! balizing the data] very di}erent FST can be averaged for di}erent loci "as has been shown#\ and it considers equally the relationship among all the populations\ even when the patterns of variability could not be uniformly distributed over the entire area of distribution of the species[ This can cloud the relationships of each popu! lation with each of the others[ For this reason\ genetic distances and identities which take into account the relative weight of the di}erent loci in the di}erentiation\ as well as the possible distinct degree of di}erentiation between di}erent populations compared two by two\ have also been used in the interpopulational compari! sons[ Even so\ genetic distances have the disadvantage of normally high standard errors[ Thus\ genetic distances were calculated in order to evaluate the variability di}erentiation among the popu! lations[ The values obtained range from 9[9049 to 9[9419 "table 3#\ which can be considered normal for the geographic population di}erentiation in insects in general\ and in Tephritidae speci_cally "BERLOCHER and BUSH\ 0871^ MILANI et al[\ 0878^ GASPERI et al[\ 0880#[ With respect to Bactrocera\ the only available data are between two Spanish populations with a dis! tance of 9[9830 "OCHANDO et al[\ 0883#\ which is higher than those obtained in the present study[ In any case\ the information available up to this time is too limited to allow for general conclusions as to the degree of geographic di}erentiation among local populations of this species[ The phylogenetic tree resulting from these distances\ is extraordinarily illustrative "_gure#[ Three clear groups stand out] one represented by the QUI population\ ano! ther by the ATA population\ and a third\ consisting of SER and FEL populations[ These groups are totally coherent with the climatology and geographical

Figure[ Dendrogram of four Spanish natural populations of Bactrocera oleae constructed from the distance matrix shown in table 3

071

location of the sampled areas "see Material and methods#[ Consequently\ it appears to add arguments for the already!mentioned action of natural selection as being responsible for the uncovered variability[ Summarizing\ three important results emerge from the present study[ First\ the amount of variability detected in the populations of Bactrocera oleae\ is higher than that observed in other Tephritidae species\ which disagrees with the expected relationship between amount of variability and degree of environmental diversity[ Second\ gene ~ow could be high as the {gen! eral| similarity in the most frequent alleles stand up[ But\ and third\ the distribution pattern of the variability shows a clear inhomogeneity among loci[ As gene ~ow "like genetic drift#\ has the same average e}ect on all nuclear genes\ it would be reasonable to conclude that natural selection\ probably due\ at least in part\ to agri! cultural practices\ in combination with gene ~ow\ are responsible for the geographic patterns of allozyme vari! ation observed in B[ oleae[ Thus\ a more e.cient _ght against this pest must take into account the high level of gene ~ow among di}erent geographic areas and a higher coordination of agricultural practices from di}erent administrations[ Acknowledgements The authors gratefully acknowledge E[ ZOUROS and P[ CAS! TAN ERA for important criticisms and comments on the manu! script[ We also thank F[ BARRANQUERO and J[ SANCHEZ for their help in obtaining part of the biological material[ This study has been carried out due to the _nancial assist! ance of the DGIGYT "PB78Ð9003#\ CAM "C127Ð80#\ and EC!FAIR2!CT85Ð0861[

References DE ANDRES\ F[\ 0880] Enfermedades y Plagas del Olivo[ Mad!

rid] Riquelme y Vargas Edics[ AYALA\ F[ J[^ POWELL\ J[ R[^ TRACEY\ M[ L[^ MOURAO\ C[ A[^ PEREZ!SALAS\ S[\ 0861] Enzyme variability in the Drosophila willistoni group[ IV[ Genetic variation in natu! ral populations of Drosophila willistoni[ Genetics 69\ 002Ð 028[ BERLOCHER\ S[ H[^ BUSH\ G[ L[\ 0871] An electrophoretic analysis of Rhagoletis "Diptera] Tephritidae# phylogeny[ Syst[ Zool[ 20\ 025Ð044[ BUSH\ G[ L[^ KITTO\ G[ B[\ 0868] Research on the genetic structure of wild and laboratory strains of the olive ~y[ In] Development of Pest Management Systems for Olive Culture Program[ FAO report[ Rome] FAO[ GRE58:414\ 09Ð19[ DALY\ J[ C[\ 0878] The use of electrophoretic data in a study of gene ~ow in the pest species Heliotis armigera "Hubner# and H[ punctigera Wallenger "Lepidoptera] Noctuidae#[ In] Electrophoretic Studies on Agricultural Pests Ed[ by LOXDALE\ H[ D[^ DEN HOLLANDER\ J[ Systematic Associ! ation Oxford] Clarendon Press\ 004Ð030[ EHRLICH\ P[ R[^ RAVEN\ P[ H[\ 0858] Di}erentiation of popu! lations[ Science 054\ 0117Ð0121[ FLETCHER\ B[ S[\ 0878a] Life history of Tephritid fruit ~ies[ In] Fruit Flies] Their Biology\ Natural Enemies and Control[ Ed[ by ROBINSON\ A[ S[^ HOOPER\ G[ Vol[ 2B[ Amsterdam] Elsevier Sc[ Publishers\ 084Ð197[ FLETCHER\ B[ S[\ 0878b] Movements of Tephritid fruit ~ies[

M[ D[ Ochando and A[ Reyes In] Fruit Flies] Their Biology\ Natural Enemies and Control[\ Vol[ 2B Ed[ by ROBINSON\ A[ S[^ HOOPER\ G[ Vol[ 2B[ Amsterdam] Elsevier Sc[ Publishers\ 198Ð108[ GASPERI\ G[^ MALACRIDA\ A[^ TOSETTI\ M[^ MILANI\ R[\ 0875] Enzyme variability] a tool for investigating the gen! ome organization and the population structure of Cer! atitis capitata[ In] Fruit Flies of Economic Importance 73[ Ed[ by CAVALLORO\ R[ Rotterdam] A[A[ Balkema\ 042Ð050[ GASPERI\ G[^ GUGLIELMINO\ C[ R[^ MALACRIDA\ A[ R[^ MIL! ANI\ R[\ 0880] Genetic variability and gene ~ow in geo! graphical populations of Ceratitis capitata "Wied[# "med~y#[ Heredity 56\ 236Ð245[ GILLESPIE\ J[ H[^ LANGLEY\ C[ H[\ 0863] A general model to account for enzyme variation in natural populations[ Genetics 65\ 726Ð737[ HUETTEL\ M[ D[^ FUERST\ P[ A[^ MARUYAMA\ T[^ CHAK! RABORTY\ R[\ 0879] Genetic e}ects of multiple population bottlenecks in the Mediterranean fruit ~y "Ceratitis cap! itata#[ Genetics 83\ s36[ IZQUIERDO\ J[ I[^ RUBIO\ J[\ 0878] Allozyme polymorphism at the aGpdh and Adh loci and _tness in Drosophila mel! anogaster[ Heredity 52\ 232Ð241[ LEVINTON\ J[\ 0862] Genetic variations in a gradient of environmental variability] marine Bivalvia "Mollusca#[ Science 079\ 64Ð65[ LEWONTIN\ R[ C[^ KRAKAUER\ J[\ 0862] Distribution of gene frequency as a test of the theory of the selective neutrality of polymorphisms[ Genetics 63\ 064Ð084[ LOUKAS\ M[\ 0878] Population genetics studies of fruit ~ies of economic importance\ especially med~y and olive fruit ~y\ using electrophoretic methods[ In] Electrophoretic Studies on Agricultural Pests\ Ed[ by LOXDALE\ H[ D[^ DEN HOLLANDER\ J[ Vol[ 28[ Systematic Association Oxford] Clarendon Press\ 58Ð091[ LOUKAS\ M[^ ECONOMOPOULOS\ A[ P[^ ZOUROS\ E[^ VERGINI\ Y[\ 0874] Genetic changes in arti_cially reared colonies of the olive fruit ~y "Diptera] Tephritidae#[ Ann[ Entomol[ Soc[ Am[ 67\ 048Ð054[ LOXDALE\ H[ D[^ DEN HOLLANDER\ J[\ "eds# 0878] Elec! trophoretic Studies on Agricultural Pests[ Systematic Association Oxford] Clarendon Press[ MALACRIDA\ A[ R[^ GUGLIELMINO\ C[ R[^ GASPERI\ G[^ BARUFFI\ L[^ MILANI\ R[\ 0881] Spatial and temporal di}erentiation in colonizing populations of Ceratitis cap! itata[ Heredity 58\ 090Ð000[ MALAVASI\ A[^ MORGANTE\ J[ S[\ 0872] Population genetics of Anastrepha fraterculus "Diptera\ Tephritidae# in di}erent hosts] genetic di}erentiation and heterozygosity[ Genetica 59\ 196Ð100[ MARINKOVIC\ D[^ TUCIC\ N[^ MOYA\ A[^ AYALA\ F[ J[\ 0876] Genetic diversity and linkage disequilibrium in Drosophila melanogaster with di}erent rates of development[ Gen! etics 006\ 402Ð419[ MCPHERON\ B[ A[\ 0889] Genetic structure of apple maggot ~y "Diptera] Tephritidae# populations[ Ann[ Entomol[ Soc[ Am[ 72\ 457Ð466[ MCPHERON\ B[ A[^ JORGENSEN\ C[ D[^ BERLOCHER\ S[ H[\ 0877] Low genetic variability in a Utah cherry!infesting population of the apple maggot\ Rhagoletis pomonella[ Entomol[ Exp[ Appl[ 35\ 044Ð059[ MCPHERON\ B[ A[^ STECK\ G[ J[\ "eds# 0885] Fruit Fly Pests[ A World of Assessment of Their Biology and Management[ Derzy Beach^ St[ Lucie Press[ MENKEN\ S[ B[ J[^ ULENBERG\ S[ A[\ 0876] Biochemical characters in agricultural entomology[ Agricultural Zoological Revue[ Vol[ 1\ Ch[ 7[ 294Ð259[ MILANI\ R[^ GASPERI\ G[^ MALACRIDA\ A[\ 0878] Biochemical genetics[ In] Fruit Flies] Their Biology\ Natural Enemies

Genetic population structure in olive fruit ~y and Control[ Ed[ by ROBINSON\ A[ S[^ HOOPER\ G[ Vol[ 2B[ Amsterdam] Elsevier Sc[\ Publications\ 22Ð45[ MORGANTE\ J[ S[^ DE SOUZA\ H[ M[ L[^ DE CONTI\ E[^ CYTRY! NOWICK\ M[\ 0870] Allozymic variability in an introduced fruit ~y pest Ceratitis capitata "Wiedeman#\ 0713\ "Dip! tera!Tephritidae#[ Rev[ Brasil[ Gen[ 1\ 072Ð080[ NEI\ M[\ 0861] Genetic distance between populations[ Am[ Nat[ 095\ 172Ð181[ NEI\ M[\ 0866] F!statistics and analysis of gene diversity in subdivided populations[ Ann[ Human Gen[ 30\ 114Ð122[ NEI\ M[\ 0867] Estimation of average heterozygosity and gen! etic distance from a small number of individuals[ Genetics 78\ 472Ð489[ NEI\ M[\ 0876] Molecular Evolutionary Genetics[ New York] Columbia University Press[ NEI\ M[^ CHESSER\ R[ K[\ 0872] Estimation of _xation indices and gene diversities[ Ann[ Human Gen[ 36\ 142Ð148[ NEI\ M[^ MARUYAMA\ T[^ CHAKRABORTY\ R[\ 0864] The bottleneck e}ect and genetic variability in populations[ Evolution 18\ 0Ð09[ NEUENSCHWANDER\ P[^ MICHELAKIS\ S[^ KAPATOS\ E[\ 0875] Tephritidae[ In] Entomologie Oleicole[ Ed[ by Aram! bourg\ Y[ Madrid] Conseil Oleicole International[ OCHANDO\ M[ D[^ CALLEJAS\ C[^ FERNANDEZ[ O[^ REYES\ A[\ 0883] Variabilidad genetica aloenzimatica en Dacus oleae "Gmelin#[ "Diptera] Tephritidae#[ I[ Analisis de dos poblaciones del sureste espan½ol[ Bol[ San[ Veg[ 19\ 24Ð33[ POULIK\ M[ D[\ 0846] Starch gel electrophoresis in a dis! continuous system of bu}ers[ Nature 079\ 0366Ð0368[ POWELL\ J[ P[\ 0860] Genetic polymorphisms in varied environments[ Science 063\ 0924Ð0925[ REYES\ A[^ OCHANDO\ M[ D[\ 0883] A study of gene!enzyme variability in three Spanish populations of Ceratitis capi! tata[ IOBC:Wprs Bull[ 06\ 040Ð059[ ROBINSON\ A[ S[^ HOOPER\ G[\ "eds# 0878] Fruit Flies[ Their Natural Enemies and Control[ Vol[ 2B[ Amsterdam] Else! vier[ SINGH\ R[ S[^ RHOMBERG\ L[ R[\ 0876] A comprehensive study of genetic variation in natural populations of Drosophila melanogaster[ II[ Estimates of heterozygosity and patterns of geographic di}erentiation[ Genetics 006\ 144Ð160[ SLATKIN\ M[\ 0870] Estimating levels of gene ~ow in natural populations[ Genetics 88\ 212Ð224[ SLATKIN\ M[\ 0874a] Gene ~ow in natural populations[ Ann[ Rev[ Ecol[ Syst[ 05\ 282Ð329[ SLATKIN\ M[\ 0874b] Rare alleles as indicators of gene ~ow[ Evolution 28\ 42Ð54[

072 SLATKIN\ M[\ 0876] Gene ~ow and the geographic structure of natural populations[ Science 125\ 676Ð681[ SNYDER\ T[ P[^ LINTON\ M[ C[\ 0873] Population structure in black~ies] allozymic and morphological estimates for Prosimulium mixtum and P[ fuscum "Diptera] Simulidae#[ Evolution 27\ 831Ð845[ TSAKAS\ S[ C[^ KRIMBAS\ C[ B[\ 0864] How many genes are selected in populations of Dacus oleae[ Genetics 68\ 564Ð 568[ TSAKAS\ S[ C[^ ZOUROS\ E[\ 0879] Genetic di}erences among natural and laboratory!reared populations of the olive! fruit ~y Dacus oleae "Diptera] Tephritidae#[ Entomol[ Exp[ App[ 17\ 157Ð165[ WRIGHT\ S[\ 0832] Isolation by distance[ Genetics 7\ 003Ð027[ WRIGHT\ S[\ 0840] The genetic structure of populations[ Ann[ Eugen[ 04\ 212Ð243[ WRIGHT\ S[\ 0854] The interpretation of population structure by F!statistics with special regard to systems of mating[ Evolution 08\ 284Ð319[ WRIGHT\ S[\ 0858] Evolution and the Genetics of Populations\ Vol[ 1[ The Theory of Gene Frequencies[ Chicago] The University of Chicago Press[ WRIGHT\ S[\ 0867] Evolution and the Genetics of Populations\ Vol[ 3[ Variability Within and Among Natural Popu! lations[ Chicago[ The University of Chicago Press[ YONG\ H[ S[\ 0877] Allozyme variation in the artocarpus fruit ~y Dacus umbrosus "Insecta] Tephritidae# from Peninsular Malaysia[ Comp[ Bioch[ Phys[ 80B\ 74Ð78[ YONG\ H[ S[\ 0889] Low allozyme variability in Bactrocera albistrigata "Insecta] Diptera] Tephritidae# from Pen! insular Malaysia[ Comp[ Bioch[ Phys[ 86B\ 008Ð010[ YONG\ H[ S[\ 0881] Allozyme variation in the melon ~y Dacus cucurbitae "Insecta] Diptera] Tephritidae# from Pen! insular Malaysia[ Comp[ Bioch[ Phys[ 091B\ 256Ð269[ ZOUROS\ E[^ KRIMBAS\ C[ B[\ 0858] The genetics of Dacus oleae[ III[ Amount of variation at two esterase loci in a Greek population[ Gen[ Res[ 03\ 138Ð147[ ZOUROS\ E[^ LOUKAS\ M[\ 0878] Biochemical and colonization genetics of Dacus oleae "Gmelin#[ In] Fruit Flies] Their Biology\ Natural Enemies and Control[ Ed[ by ROBIN! SON\ A[ S[^ HOOPER\ G[ Vol[ 2B[ Amsterdam] Elsevier Sc[ Publishers\ 64Ð76[ Authors address] M[ D[ OCHANDO and A[ REYES\ Dpto[ de Genetica\ Facultad de CC[ Biologicas\ Universidad Com! plutense\ 17939!Madrid\ Spain[ E!mail] ochandoÝe! ucmax[sim[ucm[es