sign in sign up
<<
Home , Capparimyia, Tephritidae, Trirhithrum

Heredity 76 (1996)592—602 Received14 July 1995

Allozyme divergence and phylogenetic relationships among species of tephritid

ANNA R. MALACRIDA*, CARMELA R. GUGLIELMINO, PATRIZIA D'ADAMO, CRISTINA TORTI, FULVIA MARINONI & GIULIANO GASPERI Dipartimento di Biologia Animale, Laboratorio di Zoo/ag/a, Un/versità di Pay/a, Piazza Botta 9, 1-27 100, Pa via, j-D/partimento di Genet/ca e Microbiologia, Università di Pay/a, via Abbiategrasso 207, 1-27 100, Pay/a, lstituto di Genetica Biochimica ed Evo/uzionistica del CNR, via Abbiategrasso 207, 1-27 100, Pay/a and §lstituto di Zoo/ag/a, Università di Sassari, Via Muroni 25, 1-07 100 Sassari, Italy

Multilocusenzyme electrophoresis data from 24 orthologous loci (212 alleles) were used to infer the genetic similarities between 11 pest species from the , Trirhi- thrum, , , and genera. Within some of the consid- ered species, different degrees of genetic variability were demonstrated, which appear to be related to zoogeography and to the biological traits peculiar to each species. Nei (1978) and Cavalli-Sforza & Edwards (1967) genetic distances were used to express the genetic divergence and to infer phylogenetic relationships among the species. The UPGMA clustering algorithm and the optimality criteria of Fitch & Margoliash (1967), with (KIT5cH) and without (FITCH) the tree constrained to have contemporary tips, were used. All the methods indicate the same clusters of species. One cluster is composed of , coffeae and Capparimyia savastanoi, another is composed of , and Bactrocera oleae. A further loose cluster is comprised of and Anastrepha spp. The congruence between electrophoretic phylogeny and morphological classification is discussed. Our analysis also elucidated cases, within the Ceratitis and Bactrocera genera, of interest from the evolutionary point of view, where allozyme dendrograms do not correlate well with morphological taxonomic relationships.

Keywords: clusteranalysis, genetic distances,genetic variability, multilocus enzyme electro- phoresis, Tephritidae.

The question of why only a few species have Introduction become major pests has been approached by studies on zoogeography (Maddison & Bartlett, 1989) and Thefamily Tephritidae, the true fruit flies, is one of on analysis of the life history strategies that each the most economically important dipteran families. species has evolved (Fletcher, 1989). The degree of Most of the species are pests of soft fruits, including phenotypic plasticity that each species has retained many commercial fruits. Because of their economi- has been related to the unpredictability of its cal importance, detailed research has been carried habitat, in terms of resource availability in time and out on their physiology, ecology, genetics and evolu- space. The majority of these flies belong to the tion (for comprehensive reviews see Robinson & r-strategists; however, most species are at the low Hooper, 1989; White & Elson-Harris, 1992). Fruit end of the spectrum. Polyphagous multivoltine trop- flies are represented in all world regions, but the ical and subtropical species such as Ceratitis capitata, major pest genera, Ceratitis, Bactrocera, Rhagoletis Bactrocera dorsalis and have and Anastrepha, each have limited natural distribu- typical r-characteristics, whereas oligophagous and tions. However, mankind has played an important stenophagous/monophagous species such as Bactro- part in altering the distribution of some of the more cera oleae, A nast rep ha fraterculus and Anast rep ha polyphagous species, as well as certain oligophagous suspensa have life history characteristics that exhibit species, by extending the range of their plant hosts. a mixture of r- and K-traits. Studies on distribution and host relationships have *Correspondence evidenced a prolific speciation resulting in a large 592 1996 The Genetical Society of Great Britain. PHYLOGENETIC RELATIONSHIPS AMONG TEPHRITID FLIES 593 number of species (>4000) and in a morphological and Bactrocera oleae, are represented by samples overlap among the higher taxa (White & Elson- collected from the wild. Samples of the species C. Harris, 1992). In addition, a large number of sibling capitata, C. rosa and T coffeae are from sympatric and cryptic species are known in Rhagoletis (Bush, populations and were collected together on the same 1966, 1969), and the Anastrepha fraterculus and host (coffee berries) in their putative original area Bactrocera dorsalis complexes (White & Elson- (Kenya; White & Elson-Harris, 1992). Wild samples Harris, 1992). In this context, Rhagoletis flies (R. from C. rosa were also collected in Reunion Island. pomonella species complex) have been at the centre The remaining six species listed in Table 1 were of a discussion on sympatric speciation (Bush, 1966, from laboratory colonies. 1969; Feder et a!., 1988, 1990). Electrophoretic studies of genetic differentiation have provided a powerful tool for the analysis of sympatric speciation Electrophoreticprocedures and phylogeny in Rhagoletis (Berlocher & Bush, Weperformed electrophoretic analysis using cellu- 1982; Berlocher et a!., 1993), for resolving species lose acetate gels (Cellogel) adopting the procedures complexes in the A. fraterculus group (Malavasi & described in Gasperi et a!. (1991). Morgante, 1983; Steck, 1991) and in population The following 24 orthologous enzyme loci (212 genetic analysis in C. capitata (Gasperi et a!., 1991; alleles) were analysed in each single of the 23 Malacrida et al., 1992; Baruffi et aL, 1995). All of samples: Pgm, Hk1, Hk2, Pgi, Pgk, Zw, Pg,n, xGpdh, these studies suggest the great potential of genetic Acon1, A con2, Idh, Fh, Mdh1, Mdh2, Ak1, Ak2, Adh2, approaches for biogeographical and phylogenetic Aox, Got1, Got2, Gpt, Had, Me, Mpi. All the consid- analysis of Tephritidae flies. eredloci produced consistently interpretable Despite years of taxonomic work, no satisfactory banding patterns in all species studied and the deter- classification and phylogeny exist for these flies mination of isozyme locus homologies was (White, 1989). The most common problems are unambigous. synonymy, homonymy and the establishment of For each of these biochemical loci, the electro- supra- specific groups based on questionable charac- phoretic banding patterns of C. capitata were used ters. This situation may be the consequence of the as a standard because electrophoretic variation in large size of the group, the regional nature of most this species is well documented (Gasperi et al., 1991; taxonomic work and the fact that it has been hard to Malacrida et a!., 1992; Baruffi eta!., 1995). find sound taxonomic characters. An immunological (Kitto, 1983) and, more recently, a molecular Data approach (Han & McPheron, 1994) have been analysis proposed to help the creation of a phylogenetically Foreach species sample we calculated standard based classification for the Tephritidae. measurements of polymorphism and heterogeneity: In this paper we have attempted to elucidate the P (proportion of polymorphic loci), A (average relationships among species of the related genera number of alleles per locus) and H (average propor- Ceratitis, Thirhithrum,Capparimyia,Bactrocera, tion of heterozygous individuals). Anastrepha and Rhagoletis using the multilocus The PHYLIP computer package (Felsenstein, 1993) enzyme electrophoretic approach (MLEE). The was used for the cluster analysis of the 23 samples. outcome of variability estimates and genetic similar- We calculated the genetic distance between each ities among the species are discussed in relation to pair of samples using both Nei (1978) and Cavalli- biological characteristics, zoogeography and to the Sforza & Edwards (1967) genetic distances. Two current taxonomic position of each of the considered different methods of tree construction were species. employed. The first utilized the unweighted pair group method using an arithmetic average (UPGMA) clustering algorithm; the other used the optimality Materials and methods criteria of Fitch and Margoliash (1967), first with (MTscH), and then without (FITcH) the tree Species samples constrained to have contemporary tips (Felsenstein, A total of 23 samples from 11 tephritid species were 1984). analysed for electrophoretic variation (Table 1). We also applied the bootstrap test (Efron, 1982) Each sample was composed of 30—50 flies. for assessing the robustness of each node in the tree Of the 11 species, five, i.e. Ceratitis capitata, Cera- topology. For this purpose we analysed 100 repli- titis rosa, Trirhithrum coffeae, Capparimyia savastanoi cates of bootstrap resampling of the original data

The Genetical Society of Great Britain, Heredity, 76, 592—602. 594 A. R. MALACRIDAETAL. matrixand constructed a consensus tree from the T coffeae, C.savastanoiand B.oleae species 100 bootstrapped trees obtained, considering all 24 loci. Among the three sympatric samples from the Results nativerange of C.capitata, C. rosa andTcoffeae we observed different levels of variability. Ceratitis capi- Parametersof genetic variability tatais the most polymorphic (H= 0.138)whereas T In Table 2 we show the levels of genetic variability coffeaeappearsto be the least variable (H= 0.060). estimated in the wild samples of C. capitata, C. rosa, Comparable low levels of variability were found for

Table 1 Origin and date of collection of the samples from the considered Tephritidae species

Species Origin Date of collection

Ceratitis capitata 1 sample* Kenya (Kabete, Machacos, Ruiru) 1984—92 Ceratitis rosa 4 samples Kenya (Kabete, Machacos, Ruiru) 1984—88 2 samples Reunion Island 1989 Trirhithrum coffeae 2 samples Kenya (Kabete, Machacos, Ruiru) 1988 Capparimyia savastanoi 2 samples Italy (Pantelleria Island) 1987—88 Bactrocera oleae 2 samples Italy (Liguria, Apulia) 1990—93 Bactrocera cucurbitae 2 samples Lab. colonies from: Col. Agric., Okinawa 1991—92 (Japan), and USDA, Honolulu (Hawaii) Bactrocera dorsalis 1 sample Lab. colony from USDA, Honolulu (Hawaii) 1992 4 samples Lab. colony from Dept. Agr. Gainesville (Florida) 1990—93 Anastrepha ludens 1 sample Lab. colony from Tapachula (Mexico) 1990 Anastrepha serpentina 1 sample Lab. colony from Tapachula (Mexico) 1990 Rhagoletis cerasi 1 sample Lab. colony from Sissac (Switzerland) 1993 *Gene frequencies in this sample result from the weighted average of 11 samples from the Kenya population. We decided to group them because they appear very similar to each other when considered singly.

Table 2 Parameters of genetic variability in the wild samples of Ceratitis capitata, Ceratitis rosa, Trirhithrum coffeae, Capparimyia savastanoi and Bactroceraoleae

Species Origin A±SD* P±SD* H±SD*

C. capitata Kenya 3.300 0.417 0.138 C. rosa Kenya 1.625 0.396±0.024 0.107±0.034 Reunion 1.450 0.071 0.312 0.029 0.117 0.078 T coffeae Kenya 1.300±0.000 0.250±0.000 0.060±0.047 C. savastanoi Italy 1.300±0.000 0.125±0.059 0.059±0.011 B. oleae Italy 1.550±0.071 0.291 0.089±0.002 *SD Standard deviation.

The Genetical Society of Great Britain, Heredity, 76, 592—602. PHYLOGENETIC RELATIONSHIPS AMONG TEPHRITID FLIES 595 the other two species, C. savastanoi (H= 0.059)and ' N —N N B. oleae (H =0.089). * +1+1 +1 +1 +1 +1 +1 +1 +1 +1 General low levels of variability (not reported in 00 C N * Table 2) were found in the considered laboratory l0qNoqNc•.-—— N — — — — N N strains of B. cucurbitae (H =0.061), B. dorsalis i- c — —,- (H =0.049),A. suspensa (H =0.053)and A. serpen- * +1 +1 +1 +1 +1 +1 +1 +1 +1 tina (H =0.089);in the colonies of A. ludens and R. CN trQ00 m * cerasihigher levels of variability were detected k 00 N — m 00 C (H =0.121and H =0.132, respectively). — 00 (1) * +1 +1 +1 +1 +1 +1 +1 +1 : Geneticdistances N— 'C C') N C') '—4* - -4'NN — Thematrix of the interspecific Nei genetic distances (Nei, 1978) is shown in Table 3. The lowest genetic * distance (D =0.78)is observed between C. capitata +1 +1 +1 +1 +1 +1 +1 : and T coffeae. The genetic distance observed trtr N 'C N * between the two congeneric species C. capitata and '-4N ,-m'N C) C. rosa is D =1.02,and those observed between C. 0 rosa and the Anastrepha species are of the same C.) — * N +1+1 +1 +1 +1 +1 order of magnitude, i.e. D =1.05with A. suspensa, C) —'C * D =1.31with A. ludens and D =0.89with A. serpen- N'Coqi— N — tina. Within Anastrepha we observed a distance value —N — N'-4 of 0.81 between the two species (A. suspensa and A. C) +1+1 +1 +1 +1 ludens) from the fraterculus group (Norrbom & Kim, 'C) NN 'C ":t (fl * C) 00 -N— 1988). Slightly higher distance values separate these C) last two species from A. serpentina (serpentina 'C) 0.83and D = group), being D = 0.93,respectively. 0C) +1+1 + +I Within the Bactrocera we observed the highest C) NN N 00 * interspecific distance values: B. cucurbitae vs. B. N — N -4 dorsalis is 1.53, and vs. B. oleae is 2.15. C N 'C In Table 4 we summarize the average Nei genetic * distance values between the considered genera. The +1+1 +1 lowest value is the distance between the Ceratitis and 00mrn 00 * Anastrepha genera (D =1.103)followed by the distance between Ceratitis and Trirhithrum N * (D =1.242).The largest distances separate the +1+1 : 00N * genus Bactrocera from the Trirhithrum (D =2.679), Capparmyia (D =2.534)and Anastrepha (D =2.285) N genera. * N+1: * Clusteranalysis '-4 * Theresults of the cluster analysis are shown in Figs * * 00 1—3. UPGMA trees computed using Nei genetic * N distances and Cavalli-Sforza chord measures (Cavalli-Sforza & Edwards, 1967) are shown in Fig. C) zC) 1(a,b). They represent the consensus trees of 100 C) 'I, bootstrap resamples of the original data set. Both C.)0 trees show the same topology, confirmed also by c) . . .. 'I, relatively high bootstrap values especially at the C.) ,...... C.'- c 'C) terminal nodes of the trees. In both trees the first .. .,.- C) CO splits separate two of the Bactrocera species (B. C) C) .D dorsalis and B. oleae) and Rhagoletis cerasi. The two C) —r N 00 C — next splits separate two groups of species: one 00 ,- — *

The Genetical Society of Great Britain, Heredity, 76, 592—602. 596 A. R. MALACR IDA ETAL.

Ceratitis rosa (a) 8 9Ceratitis rosa rosa gCeratitis rosa Ceratitis rosa 8 Ceratitis5 Ceratitis rosa 23 80Anastrepha suspensa Anastrepha suspensa 36 Anastrepha suspensa 20 Anastrepha suspensa Anastrepha ludens Anastrepha serpentina 56Bactrocera cucurbitae L Bactrocera cucurbitae Ceratitis capitata 87 Trirhithrum coffeae Trirhithrum coffeae savastanoi 262CapparimyiaCapparimyia savastanoi 73 Bactrocera oleae 651 [Bactrocera oleae 2 5 Bactrocera dorsa/is Rhagoletis cerasi I I I 1.98 1.76 1.54 1.32 1.10 .88 .66 .44 .22 0.0

Nei Distance

(b) Ceratitis rosa 92 Ceratitis rosa Ceratitis rosa Ceratitis1 00 rosaCeratitis rosa Ceratitis rosa 28 — 86 Anastrepha suspensa 100 rHAnastrepha suspensa 30 6 4L__Anastrepha suspensa Anastrepha suspensa Anastrepha ludens 6 Anastrepha serpent/na 27 1 00 -Bactrocera cucurb/tae Bactrocera cucurbitae 6 3 Ceratitis capitata 27 I 1 00 Trirthithrum coffeae L E'ITrirthithrum coffeae 42 1 00 Capparimy/a savastanoi Capparimyia savastanoi Rhagoletis cerasi Bactrocera oleae 1 ooj 94 Bactrocera oleae Bactrocera dorsalis

I I I .8 .7 .6 .5 .4 .3 .2 .1 .0 CavaIIi-Sforza and Edwards Distance

Fig.I Tephritid relationships inferred from UPGMA dendrograms obtained from 100 bootstrap resamplings of: (a) Nei unbiased genetic distance (1978) and (b) Cavalli Sforza and Edwards (1967) chord distance. Numbers at the nodes represent the percentage of a group's occurrence in 100 bootstrap replicates.

The Genetical Society of Great Britain, Heredity, 76, 592—602. PHYLOGENETIC RELATIONSHIPS AMONG TEPHRITID FLIES 597

Table 4 Matrix of average genetic distances* (±sd), between the genera Ceratitis, Trirhirithrum, Capparimyia, Bactrocera, Anastrepha, Rhagoletis

1 2 3 4 5 6

1.Ceratitis * 1.242±0.207 1.735±0.258 1.725±0.580 1.103±0.177 1.772±0.040 2.Trirhithrum * 1.375 0.059 2.679 0.740 1.973 0.292 1.983 0.067 * 3.Capparimyia 2.534 1.073 1.942 0.282 2.153 0.005 4.Bactrocera * 2.285 0.679 1.7520.050 5.Anastrepha * 1.810±0.200 * 6.Rhagoletis *Unbiased genetic distance; Nei (1978).

0 43 Bactrocera oleae 0.54 repeatability, especially in the Cavalli-Sforza & Edwards tree. lBactroceraBactrocera oleae dorsalis 0.97 The ultrametric tree derived from Fitch—- Rhagoletis cerasi hash optimality criteria (IuTscH) is shown in Fig. 2.

0.84 Capparimyia savastanol This tree confirms the grouping pattern shown in the [Capparimyla savastanol previous trees which were based on the same 0 39 Trirhithrum coffeas assumption of equal rates of evolution among the ..22...... J1Trirhithrum coffeae taxa. Ceratitis capitata The unrooted tree in Fig. 3, constructed without 0.14 the constraint of a constant rate of evolution Ariastrepha suspensa (FITcH), shows similar clustering of species; the clus- F[Anastrephasuspensa ters C. savastanoi—C. capitata—T coffeae and R. 0.40 0.16 [lAnastrepha suspensa cerasi—B. dorsalis—B. oleae are preserved; but we ,4astTepha suspensa observe that C. rosa and the Anastrepha species are 0.09 0.40 Anastrepha ludens 0.42 not accommodated in a single cluster although they Anastrepha serpentina separate in subsequent lineages. Ceratitis rosa F[Ceratitis rosa Discussion 0.50 —eratitis rosa 0 Weused genetic distance to express the total genetic [-Ceratitisrosa 0.05 divergence and to infer the phylogenetic relation- Ceratitisrosa ships among some economically important species of Lceratjtjs rosa Tephritidae, the majority of which are subject to alternative classifications on morphological bases 0.60 rBactrocera cucurbitae (Fig. 4). LBactrocera cucurbitae We used allozyme data from 24 orthologous loci. Fig. 2 Tephritid relationships inferred from KITSCH (ultra- On the basis that these loci are widely distributed metric) tree derived from Fitch—Margoliash optimality over the genetic maps of some Rhagoletis species criteria with the assumption of equal rates of evolution (Feder, 1989) and C. capitata (Malacrida et a!., (Felsenstein, 1984). The numbers represent the length of 1990), it is reasonable to assume that these loci may the branches (only those greater than 0.04). be considered a random, although small, sample of the genomes of the considered Tephritidae species. In addition, these loci code for a variety of enzymes includes C. rosa and the three species of Anastrepha; with both singular and multiple physiological the other groups T coffeae, C. capitata and C. savas- substrates, and for both monomeric and multimeric tanoi. The hypothetical relationship between species enzymes, which contribute differentially to overall within these two clusters is suggested in general by heterozygosity (Zouros, 1976). less than a bootstrap proportion of 50 per cent; In a preliniinaiy attempt to compare the genetic however, the affinity between T coffeae, C. capitata variation among the wild samples of the considered and C. savastanoi is suggested by a higher value of species, we found different levels of intraspecific

The Genetical Society of Great Britain, Heredity, 76,592—602. 598 A. R. MALACRIDAETAL.

rCeratitis rosa 0.34 '-Ceratitis rosa Anastrepha serpentina 0 23 0 Anastrepha ludens Anastrepha suspensa Anastrepha suspensa Anastrepha suspensa 0.07 Anastrepha suspensa 0.94 — Rhagoletis cerasi 0.22 Bactrocera dorsalis Bactrocera oleae Bactroceraoleae 0.25 - oo _j-Capparimyiasavastanol L Capparimyia savastanol 0.21 Ceratitiscapitata .crTniitum coffeae '-Trirhithrum coffeae

0.81 rBactrocera cucurbitae LBactrocera cucurbitae

Ceratitis rosa Lceratitis rosa Fig. 3 FITCH unrooted tree (Felsen- stein, 1984) derived from Fitch— Margoliash optimality criteria without the assumption of equal rates of evolution for all tip species of Tephri- rosa tidae considered. The numbers repre- LCeratitis rosa sent the length of the major branches.

variability.In particular, marked differences in the lowest level of genetic variability. The intermedi- genetic variability were found among sympatric ate level of genetic variability shown by Ceratitis rosa samples of the three species C.capitata,C.rosaand corresponds to the medium level of geographical T coffeae collected together in one of their putative diffusion of this polyphagous species. During the original areas (Kenya; White & Elson-Harris, 1992). dispersion processes C. capitata loses the greatest As these samples are sympatric we can exclude the part of its variability (Baruffi et a!., 1995); its peri- possibility that the observed intraspecific variability pheral Mediterranean populations show levels of is affected by geographical and/or climatic factors. polymorphism comparable to the ones detected in This level of genetic variability may reflect the Kenyan samples of T coffeae. No information is specific genetic plasticity and parallels the differen- available on the genetic structure or dispersion of tial dispersion capacity of these three species. In geographical populations of C. rosa and T coffeae. fact, C. capitata which is polyphagous and has a The low level of variability detected in the wild cosmopolitan geographical distribution is the most samples of the other two species, C. savastanoi and polymorphic, whereas T coffeae which is monopha- B. oleae, can be related to their narrow host special- gous and is considered an endemic species of ization (Nevo et a!., 1984). For B. oleae high genetic Western Africa (White & Elson-Harris, 1992) has similarity has been found among distant geographi-

The Genetical Society of Great Britain, Heredity, 76, 592—602. PHYLOGENETIC RELATIONSHIPS AMONG TEPHRITID FLIES 599

(a) IDaculus oleae (Musca oleae) cucurbitae ( cucurbitae)

Teph riti dae Ceratitis capitata CeratitinaePterandrus rosa (Cerafitis rosa) Trirhithrum coffeae (Ceratitis nigra) Capparimyfa savastanol

Ceratitis (Ceratitis) capitata (b) Ceratitis (Pterandrus) rosa Ceratitiffi Trirhithrum coffeae Fig. 4 Morphologically based Tephri- Capparimyia savastano! tidae subfamily relationships Ceratitinae proposed by various authors for the lB. oleae (Musca oleae) species considered in this study. The Bactrocera B. dorsalls (Dacus dorsalis) relationships proposed by Cogan & Dacini lB. cucurbitae (Dacus cucurbitae) Munro (1980) and Kugler & Freid- Tephritidae berg (1975) are summarized in (a), Dacus and those proposed by Hancock (1984, 1987) and White & Elson- IA. fraterculus Harris (1992) are summarized in (b). roxotrypanini {Ariastrepha .A. susperisa The original designations for the Trypet:nae IA. serpentina species which were subject to alter- nate classification are given in Trypetini (Rhagoletis RhagoIetis cerasi brackets.

cal populations (Zouros & Loukas, 1989) and this composedof T coffeae, C. capitata and C. savasta- finding has been correlated with the 's total noi, and another is composed of R. cerasi, B. dorsalis dependence on the fruit. and B. oleae. A further loose cluster includes C. rosa The specific life history characteristics of these and the Anastrepha species; affinity between them is species may also be related to their degree of varia- evident in all the trees; however, only trees based on bilty. Species such as C. capitata and C. rosa, which the assumption of constant evolutionary rate accom- have the attributes of r-strategist species (Fletcher, modate them under a common hypothetical 1989) are also highly polymorphic, whereas species ancestor. such as C. savastanoi and B. oleae, which are consid- ered r-K strategists, are also less polymorphic. Congruencebetween electrophoretic phylogeny and the conventional classification Congruencebetween electrophoretic trees Figure4 shows the Tephritidae subfamily relation- Inthis study we produced different electrophoretic ships based on morphological traits, proposed by estimates of the phylogeny of Tephritidae flies, one various authors in recent years for the species assuming a molecular clock (UPGMA and KITCH considered in this study. Clearly, there is no gener- trees) and one making no rate assumptions (FITCH ally accepted classification of the Tephritidae. tree). The first conclusion which can be drawn from Nevertheless, there are areas of agreement the tree analysis is that all the topologies obtained between the electophoretic trees and some of the are similar, suggesting that we cannot exclude the proposed classifications. The primary agreement possibility that rates of enzyme evolution in different concerns the closely conserved electrophoretic clades are similar. All methods indicate the presence cluster: C. capitata, T coffeae and C. savastanoi. of the same clusters of species. One cluster is That is, the demonstrated genetic affinity parallels

The Genetical Society of Great Britain, Heredüy, 76, 592—602. 600 A. R. MALACRIDA ETAL. the classical in that the three species are range of those expected between different genera grouped under a separate subfamily (Ceratitinae) (Thorpe, 1982). As reported in White & Elson- according to Cogan & Munro (1980) and even Harris (1992) B. cucurbitae, like other Zeugodacus under a separate tribe (Ceratitini) according to species, has a pattern of host relationships, which White & Elson-Harris (1992). Trirhithrum coffeae differentiate this species from other Bactrocera. This was previously included in the Ceratitis genus by the species attacks the flowers rather than the fruit of original designation of Ceratitis nigra Graham the Cucurbitaceae, a trait which is more typical of (Cogan & Munro, 1980). The low genetic distance Dacus than Bactrocera. (0.78) between C. capitata and T coffeae is in agree- Regarding tribe—genus relationships, our electro- ment with this previous classification. phoretic data support the close affinities between The two primary disagreements concern all the Ceratitini species and Dacini species proposed by electrophoretic trees and involve the species—genus Hancock (1986) and Kitto (1983). White & Elson- relationships. First, the two congeneric species C. Harris (1992) placed these two tribes within a single capitata and C. rosa are separated, secondly, B. subfamily of Ceratitinae. cucurbitae appears to be genetically unrelated to its For Rhagoletis we have considered only one congeners B. dorsalis and B. oleae. In our trees C. sample from a single species: R. cerasi. In our tree capitata is closer to T coffeae than to C. rosa. This this Rhagoletis sample is clustered with B. dorsalis result may indicate how poorly external morphology and B. oleae. The current classification places Rhago- reflects genetic affinity. Furthermore, C. rosa, like T letis and Bactrocera indifferent subfamilies. coffeae, has been assigned to various taxonomic rela- However, genetic affinity between the two Bactro- tionships. It is regarded now (Hancock 1984, 1987) cera species (B. oleae and B. dorsalis) and the prim- as a member of the subgenus Pterandrus Bezzi of the itive Rhagoletis species, i.e. R. cerasi, parallels a Ceratitis genus, whereas previously Cogan & Munro result of Han & McPheron (1994) who recognized a (1980) regarded Pterandrus as a separate genus. close similarity between two Bactrocera species (B. From the morphological point of view, Ceratitis cucurbitae and B. dorsalis) and one Rhagoletis (Ceratitis) capitata and Ceratitis (Pterandrus) rosa are species (R. striatella) based on the analysis of separated on the basis of the male secondary sexual nuclear ribosomal DNA. characters, with females being inseparable at the generic level (Hancock, 1984). We can speculate that speciation between these two Ceratitis species Conclusion may have been accelerated by this specific sexual Severalconsiderations emerge from our results. differentiation (Singh, 1988). On the other hand, Individual variation is critical for the study of the some caution is necessary to exclude in our study the systematics of closely related species, unlike higher possibility that by chance we looked at enzyme loci taxonomic levels (Soto-Adames et al. 1994). As which remain unalterated in some widely separated expected, electrophoretic data offer a great degree lines, but were strongly selected in C. rosa. of reliability in ordering genetic similarities between Concerning the relationships deduced from the closely related tephritid species. They will be useful genetic distances among the three Anastrepha in a reanalysis of the morphological data for a reor- species (A. serpentina, A. ludens, A. suspensa), this is ganization of the formal taxonomic structure of an area of agreement within the infrageneric classifi- Tephritidae flies, especially at genus—species level. cation of Norrbom & Kim (1988): the two fraterculus Moreover, our analysis demonstrated cases, within group species A. suspensa and A. ludens are more the Ceratitis and Bactrocera genera, of most interest closely related to each other than to A. serpentina from the evolutionary point of view, in which allo- which belongs to its own subgroup. zyme dendrograms do not conform well with the The second open question from our electropho- morphological taxonomic relationships. If certain retic results is the unexpected separation within the loci can be pinpointed, an important clue to the Bactrocera genus. The Bactrocera species here microevolution of these species may be at hand. considered are members of the following different Finally, the different degrees of genetic variability subgenera: B. () oleae, B. (Zeugodacus) demonstrated for the different pest species seem cucurbitae and B. (Bactrocera) dorsalis (White & related to zoogeography and to biological traits Elson-Harris, 1992). The large genetic distance esti- which are peculiar to each species. This opens the mates which separate B. cucurbitae from B. oleae problem of the role of genetic variability in disper- (D =2.15)and from B. dorsalis (D =1.53)are in the sion processes of these species.

The Genetical Society of Great Britain, Heredity, 76, 592—602. PHYLOGENETIC RELATIONSHIPS AMONG TEPHRITID FLIES 601

variation between apple and hawthorn populations of Acknowlegdements Rhagoletis pomonella in western Michigan. Evolution, Weare indebted to the following people for having 44,595—608. provided the specimens studied: J.Hendrich (IAEA, FELSENSTEIN, i. 1984. Distance methods for inferring Vienna, Austria), T. Mukiama (University of evolutionary trees: a justification. Evolution, 38, 16—24. FELSENSTEIN, i. 1985. Confidence limits on phylogenies: an Nairobi, Kenya); E. Osir (ICIPE, Nairobi, Kenya); S. Quilici (IFRA, Saint-Pierre, La Reunion); D.O. approach using the bootstrap. Evolution, 39, 783—791. FELSENSTEIN, s. 1993. PFIYLIP (Phylogeny Inference Mc-Innis (USDA-ARS, Honolulu, Hawaii); P. Liedo Package) version 3.5c. Distributed by the author. (Tapachula, Messico); C.O. Calkins (USDA, Wash- Department of Genetics, University of Washington, ington, USA). We thank Drs D. Frohlich and L. Seattle. Gomuiski for reading the manuscript. This research FITCH, W. M. AND MARGOLIASH, E. 1967. Construction of was supported by the C.N.R. (National Research phylogenetic trees. Science, 155, 279—284. Council of Italy), Special Project RAISA, Subproject FLETCHER, B. S. 1989. Life history strategies of tephritid No. 2, paper no. 2471. Grants from M.U.R.S.T. fruit flies. In: Robinson, A. S. and Hooper, G. H. S. (National Ministry of the University and Scientific (eds) Fruit Flies: Their Biology, Natural Enemies and Control, vol. 3B, pp. 195—208. Elsevier, Amsterdam, Research and Technology) (funds 40 per cent) and The Netherlands. from IAEA (International Atomic Energy Agency, GASPERI, G., GUGLIELMINO, C. R., MALACRIDA, A. R. AND Vienna, Austria) also contributed to this work. MILANJ, R. 1991. Genetic variability and gene flow in geographical populations of Ceratitis capitata (medfly). References Heredity, 67, 347—356. HAN, H. AND McPHERON, B. 1994. Phylogenetic study of BARUFFI,L., DAMIANI, 0., GUGLIELMINO, C. R., BANDI, C., selected Tephritid flies (Insecta: Diptera: Tephritidae) MALACRIDA, A. R. AND GA5pERI, o. 1995. Polymorphism using partial sequence of the nuclear 18S ribosomal within and between populations of Ceratitis capitata: DNA. Biochem. Syst. Ecol., 22, 447—457. comparison between RAPD and multilocus enzyme HANCOCK, ID. L. 1984. Ceratitinae (Diptera: Tephritidae) electrophoresis data. Heredity, 74, 425—437. from Malagasy subregion. J. Entomol. Soc. South Aft, BERLOCHER, S. H. AND BUSH, G. L. 1982. An electrophoretic 47, 277—301. analysis of Rhagoletis (Diptera: Tephritidae) phylogeny. HANCOCK, D. L. 1986. Classification of Syst. Zool., 31, 136—155. (Diptera: Tephritidae) with a discussion of the Afro- BERLOCHER, S. H., MePHERON, B. A., FEDER, J. L. AND BUSH, tropical fauna. J. Entomol. Soc. South. Aft, 49, G. L. 1993. Genetic differentiation at allozyme loci in 275—305. the Rhagoletis pomonella (Diptera: Tephritidae) species HANCOCK, D. L. 1987. Notes on some African Ceratitinae complex. Ann. Entomol. Soc. Am., 86, 716—727. (Diptera: Thephritidae), with special reference to the BUSH, o. L. 1966. The taxonomy, cytology, and evolution Zimbabwean fauna. Trans. Zim. Sci. Ass., 63, 47—57. of the genus Rhagoletis in North America (Diptera, KrvrO, o. .1983.An immunological approach to the Tephritidae). Bull. Mus. Comp. Zool., 134, 431—562. phylogeny of the Tephritidae. In: Cavalloro, R. (ed.) BUSH, 0. L. 1969. Sympatric host race formation and speci- Fruit Flies of Economic Importance, pp. 203—211. ation in frugivorous flies of the genus Rhogoletis Balkema A.A., Rotterdam. (Diptera, Tephritidae). Evolution, 23, 237—251. KUGLER, J. AND FEIDBERG, A. 1975. A list of the fruitflies CAVALLI-SFORZA, L. L. AND EDWARDS, A. W. F. 1967. Phylo- (Diptera: Tephritidae) of Israel and nearby areas, their genetic analysis: models and estimation procedures. host plants and distribution. Israel J. Erit., 10, 5 1—71. Evolution, 21, 550—570. MADDISON, P. A. AND BARTLETF, B. j.1989.A contribution COGAN, B. H. AND MUNRO, H. K. 1980. Family Tephritidae. towards the zoogeography of the Tephritidae. In: In: Crosskey, R. W. (ed.) Catalogue of the Diptera of the Robinson, A. S. and Hooper, G. H. S. (eds) Fruit Flies, Afrotropical Region, pp. 518—554. British Museum Their Biology, Natural Enemies and Control, vol. 3A, pp. (Natural History), London. 27—36. Elsevier, Amsterdam, The Netherlands. EFRON, B. 1982. The jackknife, the bootstrap, and other MALACRIDA, A. R., GASPERI, 0., BARUFFI, L. AND MILANI, R. resampling plans. Conf Board Math. Sci. Soc. md. Appl. 1990. The contribution of formal genetic studies to the Math., 38, 1—92. species characterization of Ceratitis capitata Wied. In: FEDER, j.L.1989. Inheritance and linkage relationships of Genetic Sexing of the Mediterranean Fruit Fly ,pp. allozymes in the fly. J. Hered., 80, 85—89. International Atomic Energy Agency Press, 277—283. Vienna, Austria. FEDER, J. L., CHILCOTE, C. A. AND BUSH, G. L. 1988. Genetic MALACRIDA, A. R., GUGLIELMINO, C. R., GASPERI, 0., differentiation between sympatric host races of Rhagole- BARUFFI, L. AND MILANI, R. 1992. Spatial and temporal tis pomonella. Nature, 336, 6 1—64. differentiation in colonizing populations of Ceratitis FEDER, J. L., CHILCOTE, C. A. AND BUSH, G. L. 1990. capitata. Heredity, 69, 101—111. Regional, local and microgeographic allele frequency MALAVASI, A. AND MORGANTE, i. s. 1983. Population genet-

The Genetical Society of Great Britain, Heredity, 76, 592—602. 602 A.R. MALACRIDAETAL.

ics of Anastrepha fraterculus (Diptera, Tephritidae) in of G6pdh at different taxonomic levels in hexapoda with different hosts: Genetic differentiation and heterozygo- emphasis in Diptera. Ann. Entomol. Soc. Am., 87, sity. Genetica, 60, 207—211. 723—736. NEI, M. 1978. Estimation of average heterozygosity and STECK, G. i. 1991. Biochemical systematics and population genetic distance from a small number of individuals. genetic structure of Anastrepha fraterculus and related Genetics, 89, 583—590. species (Diptera: Tephritidae). Ann. Entomol. Soc. Am., NEVO, E., BElLES, A. AND BEN-SHLOMO, R. 1984. The evolu- 84,10—28. tionary significance of genetic diversity: ecological THORPE, J. P. 1982. The molecular clock hypothesis: demographic and life history correlates. In: Mani, G. S. biochemical evolution, genetic differentiation and (ed.) Evolutionary Dynamics of Genetic Diversity. systematics. Ann. Rev. Ecol. Syst., 13, 139—168. Lecture notes in Biomathematics, vol. 53, pp. 13—213. WI-IITE, i. M. 1989. The state of fruit fly taxonomy and Springer-Verlag, Berlin. future research priorities. In: Cavalloro, R. (ed.) Fruit NORRBOM, A. L. AND KIM, K. C. 1988. A List of the Reported Flies of Economic Importance 87, pp. 543—552. Balkema Host Plants of the Species of Anastrepha (Diptera: A.A., Rotterdam. Tephritidae). United States Department of Agriculture WHITE J. M. AND ELSON-HARRIS M. (eds). 1992. Fruit Flies (APHIS 81—52), Washington. of Economic Significance: Their Identification and ROBINSON A. S. AND HOOPER G. H. S. (eds). 1989. Fruit Bionomics. C.A.B. International, Wallingford, Oxford- Flies: Their Biology, Natural Enemies and Control, vols shire. 3A and B. Elsevier, Amsterdam, The Netherlands. ZOUROS, E. 1976. Hybrid molecules and superiority of the SINGH, R. S. 1988. Genetic studies of species differences heterozygote. Nature, 262, 227—229. and their relevance to the problem of species forma- ZOUROS, E. AND LOUKAS, M. 1989. Biochemical and coloni- tion. In: Loxdale, H. D. and Hollander, J. (eds) Electro- zation genetics of Dacus oleae. In: Robinson, A. S. and phoretic Studies on Agricultural Pests, pp. 20—45. Oxford Hooper, G. H. S. (eds) Fruit Flies: Their Biology, Natural University Press, New York. Enemies and Contro4 vol. 3B, pp. 75—88. Elsevier, SOTO-ADAMES, F. N., ROBERTSON, H. M. AND BERLOCHER, S. Amsterdam, The Netherlands. H. 1994. Phylogenetic utility of partial DNA sequences

The Genetical Society of Great Britain, Heredity, 76, 592—602.