Morphometric Analysis of Mexican and South American Populations of The

Bulletin of Entomological Research (2004) 94, 487–499 DOI: 10.1079/BER2004325

Morphometric analysis of Mexican and South American populations of the Anastrepha fraterculus complex (Diptera: Tephritidae) and recognition of a distinct Mexican morphotype

V. Hernández-Ortiz1*, J.A. Gómez-Anaya1, A. Sánchez1, B.A. McPheron2 and M. Aluja1 1 Instituto de Ecología A.C., Km 2.5 carretera antigua a Coatepec No. 351, Congregación El Haya, Apartado Postal 63, C.P. 91000, Xalapa, Veracruz, México: 2 Department of Entomology, Pennsylvania State University, 501 ASI Building, University Park, PA 16802, USA

Abstract

Discriminant function and cluster analyses were performed on 19 morphometric variables of the aculeus, wing and mesonotum to determine whether populations of Anastrepha fraterculus (Wiedemann) from different parts of Mexico could be distinguished from populations from South America. Samples were collected from seven localities across Mexico, two from Brazil, and one each from Colombia and Argentina. Results showed there were statistically significant differences between Mexican and South American populations with respect to the aculeus (tip length, length of serrated section, mean number of teeth) and wing (width of S-band and connection between S- and V-bands). The degree of morphological variation observed among Mexican populations was extremely low, and as a consequence, the Mexican populations were identified as a single morphotype by discriminant analysis. The ‘Andean morphotype’ consisting of the Colombian population, and the ‘Brazilian morphotype’ consisting of the two Brazilian populations plus the single Argentinian population were also distinguished. It was concluded that the macro-geographical morphotypes from Mexico, Colombia, and Brazil plus Argentina correspond to three distinct taxonomic entities. Comparisons of results with those obtained from behavioural, karyotypic, isozyme and DNA studies suggest that sufficient evidence now exists to name a new Mexican species from within the A. fraterculus complex. This will be done in a separate publication. A provisional key to the morphotypes of A. fraterculus studied is provided.

Introduction major pests that reproduce in cultivated fruits (Aluja, 1994; Hernández-Ortiz, 1996). Although the genus is composed of Based on current knowledge, Anastrepha Loew is the about 200 known species distributed throughout the most diverse Neotropical genus of the Tephritidae (Diptera). Neotropical Region, the taxonomic status of some species It includes around seven native species usually considered complexes has never been adequately resolved (Hernández- Ortiz & Aluja, 1993; Norrbom et al., 1999). *Fax: +52 228 8187809 The fact that Anastrepha fraterculus (Wiedemann) E-mail: [email protected] (Diptera: Tephritidae) or ‘South American fruit fly’ is a 488 V. Hernández-Ortiz et al. cryptic species complex has been recognized since the Annonaceae, Combretaceae, Fabaceae, Flacourtiaceae, beginning of the past century. Steck (1999) summarized Juglandaceae, Moraceae, Myrtaceae, Oxalidaceae, published information providing evidence that multiple Punicaceae, Rosaceae, Rubiaceae, Rutaceae and Vitaceae cryptic species have been confused under this single (Norrbom, 2004). With respect to its pest status, even though currently valid species name. it sporadically attacks peaches Prunus persica (L.) Batsch Morphological variability is such that as many as nine (Rosaceae) in Mexico (Aluja et al., 2000), it can only be names have been assigned to the species (Zucchi, 1981). This considered a true pest of apples in Brazil (Sugayama et al., issue was first addressed by Lutz & Lima (1918) and Lima 1997), of grapefruit in Argentina (Nasca et al., 1981), and of (1934), and later by Stone (1942), who in his revision of the blackberries Rubus glaucus Benth. (Rosaceae) in Venezuela genus Anastrepha wrote: ‘ … this species (A. fraterculus) is (Briceño, 1975). The case of citrus fruit is particularly both an abundant and variable one. As treated here, it relevant because only some of the South American forms, extends from the Rio Grande Valley in Texas south to and not the Mexican form (sensu Baker et al., 1944), are Argentina, and it is possible that it will eventually be found proven pests of this type of fruit (see Aluja et al., 2000, 2003; to represent a complex of species rather than a single one’. for further details). Two years later, Baker et al. (1944) argued that A. fraterculus Multivariate morphometric techniques have been useful was a South American species, and considered the ‘Mexican in detecting morphological differences among populations in form’ distinct from it. Since then, a series of karyotypic a variety of organisms (Willig et al., 1986). Such methods are (Mendes, 1958; Bush, 1962; Solferini & Morgante, 1987), currently widely employed in taxonomic investigations to morphological (Hernández-Ortiz, 1992; Norrbom et al., 1999), distinguish closely related species, justify synonymies, isozyme (Morgante et al., 1980; Malavasi & Morgante, 1982; demonstrate morphological variation along altitudinal or Steck, 1991) and mitochondrial DNA (Steck & Sheppard, geographical gradients, and propose new species (Reyment 1993; Santos & Mattioli, 1996) studies have provided et al., 1984; McNamee & Dytham, 1993). In the case of fruit evidence that Alan Stone was correct. For example, Morgante flies, the use of morphometric methods has been tested and et al. (1980) suggested that at least four taxa might be found successfully employed for the following taxonomic purposes: in samples collected from different regions in Brazil, one of (i) to determine differences among species of the Bactrocera which turned out to be Anastrepha sororcula Zucchi (Diptera: dorsalis (Hendel) complex, using wing characters Tephritidae). Mendes (1958) and Bush (1962), working in (Adsavakulchai et al., 1998); (ii) to analyse the variability Brazil and Mexico, respectively, discovered karyotypic among A. fraterculus specimens collected from different host differences between samples from both countries. plants in Tucumán, Argentina (Perero et al., 1984); and (iii) to Furthermore, Solferini & Morgante (1987) described four separate A. fraterculus populations from Brazil into two karyotypes from Brazilian populations of A. fraterculus. More groups using aculeus, wing, and head (frontal plate) recently, Steck (1991) found that isozyme patterns of A. characters (Selivon, 1996). More recently, Araujo et al. (1998) fraterculus individuals from northeastern Brazil (Bahia), employed discriminant function analysis on aculeus coastal Venezuela, Costa Rica and Mexico were similar, but measurements in A. fraterculus, Anastrepha obliqua (Macquart), that individuals from populations in southern Brazil, Andean A. sororcula and Anastrepha zenildae Zucchi, all members of Venezuela, and Peru were genetically distinct. Shortly the fraterculus species group (sensu Norrbom et al., 1999) and thereafter, Steck & Sheppard (1993) found that specimens concluded that the method was a useful taxonomic tool for collected from coastal Venezuela and from the Bahia region distinguishing between closely related or sibling species. of Brazil were highly differentiated based on mitochondrial The aim here was to determine whether A. fraterculus DNA samples, even though they had originally appeared individuals collected in different parts of Mexico could be similar in isozyme analyses. distinguished from individuals from South American In sum, there is now sufficient genetic evidence to clearly populations using morphometric methods. Determining distinguish Andean populations of A. fraterculus from other whether Mexican populations of the A. fraterculus complex regions of South America and to differentiate various cryptic represent a species that is distinct from their South American species in Brazil. However, an important question still counterparts is particularly relevant, given the quarantine remains to be addressed: do differences between Mexican restrictions imposed upon Mexico by some countries and South American populations merit the recognition of a importing fresh Mexican citrus (e.g. USA, Japan) and the Mexican species from within this cryptic species complex? possibility of using the sterile insect technique to manage or Differences between A. fraterculus populations have also eradicate A. fraterculus populations from particular regions. been observed in terms of host use patterns and pest status. Considering that it has become common practice to produce For example, Bush (1962) noted that individuals in Brazilian sterile flies in one country and release them in another, the populations had a wide host range and attacked citrus, question of species identity becomes critical. For example, while individuals in Mexican populations infested rose Mexican A. fraterculus that were mass reared and sterilized in apple Syzygium jambos (L.)Alston (Myrtaceae) and guava Mexico may be of little use in, for example, Peru or vice versa. Psidium guajava L. (Myrtaceae) and were of no economic Therefore the problem is addressed not only from a taxonomic importance. Since then evidence has accumulated that A. viewpoint, but also embraces corroborative behavioural (Aluja et al., 2003) and genetic techniques (McPheron et al., fraterculus in Mexico has a much wider host range than unpublished) to increase the resolution of the analyses. previously thought, even though it prefers to infest fruit in the family Myrtaceae (Baker, 1945; Aluja et al., 1987, 2000; Hernández-Ortiz, 1992). At a macro-geographical scale Materials and Methods (from Mexico to Argentina), the A. fraterculus complex is Fruit fly samples clearly polyphagous, infesting approximately 67 host plant species in Brazil alone (Zucchi, 1999). It is able to infest fruit Populations of the A. fraterculus complex were collected from such wide-ranging families as Anacardiaceae, from seven localities in Mexico, two in Brazil, one in Morphometric analysis in the Anastrepha fraterculus complex 489

Colombia, and one in Argentina. Details for each locality aculeus and wing structures of all specimens were mounted included short site description, a unique collection number on permanent slides prior to observations. Female in brackets, an acronym (in square brackets), elevation of terminalia were cleaned in a boiling solution of 10% sodium collection site (metres above sea level), collecting date and hydroxide. Measurements were made with an ocular method (i.e. adult reared from infested fruit or captured in micrometer in the compound scope and drawings were trap), and name of collector (in brackets). MEXICO: (1) produced with a camera lucida. Measurement data were Veracruz, Apazapan, Apazapan [Mex-Apaz], 250 m, 5 May entered in a data matrix and converted to millimetres. 1992, McPhail trap (G. Quintero and L. Quiroz); (2) Veracruz, Emiliano Zapata, La Jicayana [Mex-Jica], 400 m, February Aculeus 1996, McPhail trap (P. Juárez); (3) Veracruz, Teocelo, Tejería Aculeus (fig. 1, aculeus tip in ventral view): A1, ratio [Mex-Teoc], 980 m, 1 October 1996, host: Psidium guineense A2/A3; A2, basal aculeus tip length (= from the margin of Sw. (J. Piñero); (4) Veracruz, Coatepec [Mex-Coat], 1200 m, sclerotized area on ventral side to beginning of serrated 22 July 1996, host: Syzygium jambos (J.Piñero); (5) Veracruz, section); A3, apical aculeus tip length (= length of serrated Los Tuxtlas Biological Station [Mex-Tuxt], 160 m, 9 section); A4, width at end of margin of sclerotized area on September 1997, host: Psidium guajava (J. Piñero); (6) ventral side; A5, width at beginning of serrated section; A6, Chiapas, Lagunas de Montebello, San Vicente [Mex-Chis], length from the apex to the lateral base of the aculeus tip 1400 m, August 1997, host: Psidium guajava (C. Estrada); (7) (measured along its left side); A7, total length of the aculeus Quintana Roo, Chunhuhub [Mex-QRoo], 30 m, 5–12 July (not illustrated); A8, mean number of teeth on each side; A9, 1997, McPhail trap (P. Xool Cetz). BRAZIL: (8) Sao Paulo, ratio of the aculeus tip length (A2+A3) and total length of laboratory colony [Bra-SPau], (R.A. Zucchi); (9) Santa the aculeus (A7) = (A2+A3)/A7. Catarina, Caçador [Bra-SCat], 6 February 1992, McPhail trap (I. Nora). COLOMBIA: (10) Cundinamarca, La Mesa [Col- Mesonotum Cund], 1998 (exact date unknown), McPhail trap (J. Varón). ARGENTINA: (11) Tucumán [Arg-Tucu], February 1994, Mesonotum (fig. 2, dorsal view): M1, mesonotal length; host Psidium guajava (S. Ovruski). M2, mesonotal width at the level of the postsutural supra- The morphology of A. fraterculus is known to vary greatly alar seta; M3, length from the apex of the scutellum to the across a wide geographical range, although this variation postsutural supra-alar seta. cannot be correlated with ecological conditions and host preferences throughout its range (Stone, 1942). To analyse Wings this variability, the studied samples were chosen on the basis Wings (fig. 3, dorsal view): W1, wing length from extreme of the following criteria: (i) specimens were sampled during base of costa to apex; W2, wing width; W3, width of S-band a single collection event to assure their common origin; (ii) from the juncture of S-band and vein R4+5 perpendicular to specimens were obtained from localities that broadly costa, (the apical part of the S-band is homologous with the represented the ecological and biogeographical conditions anterior apical band); W4, width at base of proximal arm of V- associated with the occurrence of these flies, including band; W5, connection of S and V-bands between veins R2+3 altitude, dominant vegetation, and biogeographical and R4+5 (1 = present, 2 = absent); W6, connection of province; (iii) specimens were collected from a variety of proximal and distal arms of V-band between veins R4+5 and host plants (P. guajava, P. guineense and S. jambos). M (1 = connected, 2 = unconnected); W7, ratio of width and Ten fully-coloured adult females per population were length of wing = W2/W1. randomly selected from each collection and subjected to morphometric analyses. Additional specimens from the same samples were used for analysis by other methods such Statistical analysis as DNA sequencing. In order to improve the morphometric Descriptive statistics analysis resolution, the number of variables analysed was increased from 8 in a recent study (Araujo et al., 1998) to 19 Means and standard deviations were calculated for the in the present investigation. The variability in Mexican linear measurements, ratios, and values of qualitative populations was consistently low, suggesting that the characters associated with 19 variables. Tests for normality relatively low sample size did not bias their interpretation. and homogeneity indicated that all variables were normally distributed, except for W5 and W6 which were qualitative variables with assigned values. The results of these analyses Morphological variables were then used to construct diagram boxes to explore Morphological studies were carried out to select a suite of variation in each morphological character among samples. relevant features commonly used to identify most Anastrepha species (Zucchi, 1978; Norrbom, 1985; Hernández-Ortiz, 1992; Cluster analysis Norrbom et al., 1999). Particular attention was focused on the A cluster analysis was run for the entire data set. Means female aculeus, mesonotum and wing pattern. On the basis of for each variable were entered for all 11 sample groups, and these structures, 19 morphometric variables were selected for then the single linkage (nearest neighbour) amalgamation measurements (see figs 1–3 and description below). These method, expressed in City-block (Manhattan) distances, was variables were expressed as linear distances between two used to analyse the data. points, except in the cases of A1, A9, W7, which corresponded to ratios between two measurements, and W5 and W6, Multivariate analysis expressed as the presence/absence of qualitative characters. Mesonotal variables were observed directly with a Zeiss Discriminant function analysis was applied to the full stereo microscope and ocular micrometer, whereas the data set of 19 variables in order to determine the possibility 490 V. Hernández-Ortiz et al.

Figs 1–3. Drawings of the structures and reference points employed in the morphometric analyses of Anastrepha fraterculus: 1, aculeus, ventral view; 2, mesonotum, dorsal view; 3, wing, dorsal view. Variables are described in the methods section.

of an a priori segregation of the 11 already established was derived using Mahalanobis distances as measured from groups. The only criterion used for grouping individuals group centroids. All statistical analyses were executed with was their site of origin. A discriminant model was the Statistica (1999) software program. constructed using the method of forward stepwise analysis that reviews all variables and evaluates which ones will contribute most to the discrimination between groups. A Results Wilks’ lambda statistic was applied as a measure of Morphological variation significance of the discriminatory power of the model and for each variable. Means and standard deviations for 19 selected variables A canonical correlation analysis to determine the from 11 population samples of the A. fraterculus complex canonical variables (or roots) was conducted, and their were obtained for comparison. Even though sample sizes for significance in the model were estimated using Chi-square each population were relatively small, standard deviations tests with successive roots removed. The individual factor were consistently small, suggesting the data were scores for the first two discriminant functions were plotted reasonably representative of the natural variation occurring to visualize the results. Pair-wise comparisons to test for the in the populations sampled (table 1). equality of groups in the canonical variates space were Certain measurements, such as the length of the aculeus based on Mahalanobis distances, as a measure of distance tip (A2+A3), were greater in the Mexican samples between groups based on means, variances and covariances (0.266–0.292 mm) than in the samples from South America (Zar, 1999). The subsequent correct classification of cases (0.233–0.263 mm). Most of this difference occurred in the Morphometric analysis in the Anastrepha fraterculus complex 491 from from Mexico, Brazil, , A7 expressed A7 as expressed mean of , teeth, Anastrepha Anastrepha fraterculus Mex-QRoo Mex-Jica Mex-Teoc Mex-Apaz Mex-Chis Mex-Tuxt Mex-Coat Bra-SPau Bra-SCat Col-Cund Arg-Tucu whereas W5 and W6 are qualitative characters (see text). W5 and W6 are whereas Table 1.Table Means and standard deviations for 19 morphological variables measured among population 11 samples of Argentina. Colombia and N=A9 and ratios W7 A1, of which two in represent except All expressed or millimetres, morphometric for measurements more measures 10 10 10 10 10 10 10 10 10 10 10 A1A2A3A4A5 0.76±0.08A6 0.12±0.01 0.81±0.07A7 0.15±0.01 0.12±0.01A8 0.72±0.06 0.12±0.01 0.15±0.01A9 0.12±0.01 0.08±0.01 0.81±0.11 0.12±0.01W1 0.17±0.01 0.28±0.02 0.12±0.01 0.08±0.01W2 0.75±0.09 0.13±0.01 1.70±0.06 0.15±0.02 0.27±0.03W3 0.12±0.01 0.09±0.01 11.3±1.1 0.78±0.11 0.12±0.01 1.72±0.12W4 0.17±0.01 0.29±0.01 0.16±0.00 0.11±0.01 0.08±0.01 0.69±0.09W5 0.13±0.01 10.3±1.1 6.11±0.42 1.78±0.07 0.15±0.01 0.28±0.02 0.15±0.01 0.11±0.01W6 0.09±0.01 2.56±0.20 0.78±0.06 0.11±0.01 5.64±0.54 1.79±0.10 0.16±0.01 12.0±0.6W7 0.30±0.02 0.44±0.03 0.16±0.01 0.11±0.01 0.08±0.01 2.37±0.23 0.89±0.11 0.12±0.01M1 1.82±0.07 5.99±0.34 1.48±0.12 0.14±0.01 0.27±0.02 0.41±0.05 0.15±0.01 0.12±0.01 10.3±1.4 0.09±0.00M2 2.60±0.19 0.94±0.07 1.20±0.42 0.12±0.00 1.59±0.08 6.13±0.44 1.31±0.09 0.14±0.01 0.28±0.01M3 0.16±0.01 0.43±0.05 0.12±0.01 1.00±0.00 0.08±0.01 12.7±1.1 2.56±0.26 1.00±0.00 0.82±0.08 0.12±0.01 1.74±0.04 6.42±0.41 1.34±0.08 0.13±0.01 0.42±0.02 0.26±0.01 0.16±0.01 0.44±0.03 1.00±0.00 0.11±0.01 0.08±0.00 2.82±0.18 1.00±0.00 0.11±0.01 2.82±0.24 1.61±0.04 6.00±0.59 11.4±2.1 1.38±0.10 0.42±0.01 0.13±0.01 0.27±0.02 0.16±0.01 0.44±0.03 1.00±0.00 0.07±0.00 1.89±0.14 2.60±0.20 1.00±0.00 2.64±0.27 0.11±0.01 5.89±0.24 1.66±0.07 1.39±0.10 0.43±0.01 13.0±1.2 0.25±0.01 1.73±0.15 0.16±0.01 0.41±0.04 1.00±0.00 1.75±0.17 0.07±0.01 2.46±0.12 1.10±0.32 2.80±0.18 1.73±0.09 6.32±0.50 1.27±0.08 0.42±0.02 1.66±0.20 0.24±0.01 0.41±0.02 0.16±0.01 1.00±0.00 1.88±0.13 2.66±0.20 9.4±1.1 1.00±0.00 2.77±0.28 1.52±0.06 6.67±0.52 1.34±0.04 0.44±0.01 1.81±0.14 0.14±0.01 0.38±0.04 1.00±0.00 1.85±0.22 2.83±0.24 1.00±0.00 2.95±0.24 6.46±0.36 1.37±0.09 0.43±0.02 9.2±1.1 1.74±0.21 0.15±0.01 0.38±0.03 1.00±0.00 2.04±0.15 2.65±0.18 1.90±0.32 2.83±0.26 6.55±0.32 1.60±0.18 0.42±0.01 1.86±0.13 0.28±0.02 10.7±1.91.30±0.48 1.89±0.15 2.69±0.11 2.00±0.00 2.70±0.14 1.52±0.09 0.42±0.01 1.85±0.14 0.35±0.03 1.10±0.32 1.80±0.06 2.00±0.00 2.84±0.10 9.4±1.1 1.49±0.10 0.42±0.01 1.71±0.08 1.90±0.32 1.90±0.12 2.00±0.00 2.89±0.26 0.41±0.02 1.78±0.08 2.00±0.00 1.91±0.16 2.78±0.25 0.41±0.01 1.79±0.15 1.79±0.19 2.71±0.18 1.73±0.16 1.82±0.07 1.69±0.11 492 V. Hernández-Ortiz et al. length of the serrated section (A3), which was longer in grouped into two large clusters (City-block distance = 2.9), Mexican specimens (0.147–0.167 mm) than in those from one comprising the Mexican samples and the other the South America (0.126–0.144 mm), but there was no South American populations. However, at a slightly shorter significant difference between the Mexican and South distance (2.1) three clusters formed: one cluster grouped all American specimens with respect to the distance from the the Mexican samples, a second cluster included the two margin of the sclerotized area on the ventral side to the Brazilian samples plus the one Argentinian sample, and the initiation of the teeth (A2). This result implied that the third cluster comprised the Colombian sample alone. serrated section of the aculeus tip was shorter in South American individuals than in Mexican individuals, but that Discriminant function analysis the non-serrated section of this structure was similar in length among all the populations studied. The discriminant function analysis yielded a model Another taxonomic character of significance, was the including 18 variables, only the length of mesonotum (M1) overall length of the aculeus (A7) which ranged from 1.70 to was removed from the model. The stepwise analysis 1.82 mm in Mexican individuals (except in the Mexico- summary showed a significant variability among the 11

Tuxtlas sample with a mean of 1.59 mm), whereas the range population samples (Wilks’ lambda = 0.00029; F (180, 741) = for South American specimens varied between 1.52 and 6.28; P < 0.0001). Independent contributions to the 1.73 mm (table 1). The mean number of teeth on the serrated prediction for each variable in the model were highly section of the aculeus (A8) ranged from 10.3 to 13.0 in significant (P < 0.05) for some variables of the wing (W3, W6, Mexican specimens, but was fewer in South American W5 and W4), and for the aculeus (A8, A3, A7, A1, A9, A2 and specimens (9.2 to 10.7). With respect to this character, A4), whereas for the mesonotum only M3 indicated individuals from Colombia were more similar to the significant differences (table 2). Mexican specimens than they were to the specimens from The results yielded ten discriminant functions, and the Brazil or Argentina. Wing length (W1) was found to be Chi-square tests indicated that only the first six roots longer in South American (6.32 to 6.67 mm) than in Mexican (canonical variables) were of significance. On the basis of material (5.64 to 6.42 mm), and the width of the S-band near standardized coefficients for the 19 morphological features, the apex (W3) was greater in Mexican samples it was found that first canonical variable (CV-1) accounted (0.412–0.444 mm) when compared to South American for 71.1% of the discriminatory power of the model, the material (0.276–0.384 mm). second variable (CV-2) accounted for 11.8%, and together The qualitative character W5, representing the presence CV-3 to CV-6 accounted for only 14.9%. Key variables for or absence of a connection between the S and V-bands on the group separation along CV-1 were the length of the serrated wing, was very consistent (table 1). In all South American section of the aculeus tip (A3), the total length of the aculeus samples this connection was absent, whereas in the Mexican (A7), wing length (W1), and wing width (W2). The means of samples, a high proportion of individuals (almost 98%) population groups indicated that all Mexican samples were showed this connection. In the same way, the upper mainly discriminated by CV-1, which had positive connection of both arms of the V-band (W6) was always coefficients with respect to all South American samples, the present in the Mexican and Brazilian samples, but was sample from Cundinamarca, Colombia being highly absent in those samples collected from Colombia and differentiated (table 3). Argentina. Mahalanobis distances used to compare morphometric divergence among population group centroids indicated a low degree of segregation and little inter-population Cluster analysis variability among Mexican samples. The closest distance Amalgamation of the 11 samples based on means for all was observed between the samples from Apazapan and variables resulted in strong morphometric dissimilarities Jicayana (MD = 3.3), whereas the greatest distance occurred among the clusters (fig. 4). The sample populations were between the Chiapas and Tuxtlas samples (MD = 30.7). Nevertheless, comparisons between Mexican and South American samples produced the overall greatest Mahalanobis distance values, suggesting a large morphometric divergence between populations from the two continental areas. In this case, the smallest distance occurred between the Mexico-Tuxtlas and Brazil-Sâo Paulo samples (MD = 43.1), whereas the greatest distance was found between Mexico-Teocelo and Colombia- Cundinamarca samples (MD = 186.5). Comparisons of Mahalanobis distance values among South American samples showed that there was much greater divergence than among Mexican samples. The two Brazilian samples were closely related to each other (MD = 15.1) and these, in turn, were fairly similar to the Argentina- Tucuman sample (with Brazil-Sâo Paulo = 29.9 and Brazil-Santa Catarina = 36.6, respectively). The greatest Fig. 4. Dendrogram derived from the cluster analysis using distances were found when comparing the Colombian means of 19 variables from 11 populations of Anastrepha sample with all the other samples from Brazil, Argentina fraterculus. Joining rule: single linkage; distance metric: City- and Mexico (see table 4). block (Manhattan) distances. The percentage of correct classifications for all Morphometric analysis in the Anastrepha fraterculus complex 493

Table 2. Discriminant function analysis summary. Independent contributions to the predictions in the model for each variable. Wilks’ Partial F-remove 1-Tolerance Variable lambda lambda (10,82) P-level Tolerance (R-sqr.) W5 0.000535 0.545 6.849 0.000* 0.913 0.087 W6 0.000556 0.524 7.448 0.000* 0.874 0.126 A7 0.000375 0.778 2.349 0.017* 0.005 0.995 W3 0.000623 0.468 9.322 0.000* 0.398 0.602 W2 0.000337 0.865 1.284 0.253 0.002 0.998 A3 0.000375 0.778 2.346 0.017* 0.005 0.995 W4 0.000430 0.678 3.893 0.000* 0.501 0.499 A8 0.000439 0.665 4.140 0.000* 0.588 0.412 W1 0.000342 0.853 1.415 0.188 0.002 0.998 A4 0.000367 0.794 2.126 0.031* 0.305 0.695 M3 0.000374 0.779 2.327 0.018* 0.309 0.691 M2 0.000345 0.846 1.490 0.158 0.291 0.709 W7 0.000336 0.868 1.252 0.271 0.009 0.991 A5 0.000339 0.861 1.327 0.230 0.269 0.731 A9 0.000375 0.778 2.334 0.017* 0.004 0.996 A1 0.000380 0.768 2.471 0.012* 0.011 0.989 A2 0.000364 0.801 2.043 0.038* 0.004 0.996 A6 0.000339 0.859 1.343 0.221 0.028 0.972

*Statistically signiﬁcant.

Table 3. Means of canonical variables for all groups. Groups CV-1 CV-2 CV-3 CV-4 CV-5 CV-6 Mex-QRoo 2.288 0.869 0.169 1.387 0.257 0.976 Mex-Jica 3.145 –0.110 0.314 0.820 1.354 –0.894 Mex-Teoc4.297 –0.729 –0.455 –0.318 –1.403 –0.397 Mex-Apaz 3.085 –0.742 0.740 0.743 1.307 –0.632 Mex-Chis 3.884 –0.765 0.301 –0.400 –1.511 0.324 Mex-Tuxt 1.545 1.355 –1.276 –2.525 1.546 0.439 Mex-Coat 3.787 –1.136 –0.715 0.345 –0.503 0.557 Bra-SPau –3.144 2.292 0.261 –0.753 –1.194 –1.099 Bra-SCat –4.185 1.420 2.752 –0.328 0.031 0.633 Col-Cund –8.006 –4.329 –0.108 –0.479 0.238 –0.004 Arg-Tucu –6.697 1.875 –1.983 1.509 –0.122 0.097

individuals was 79.1%. Specimens from Colombia and the Colombian sample alone. CV-1 discriminated the Mexican Argentina were classified correctly 100% of the time, and South American populations, whereas CV-2 segregated whereas the Brazilian specimens from Sâo Paulo and Santa the two morphotypes from South America (fig. 5). A second Catarina were classified correctly 70% and 90% of the time, scatterplot of canonical scores regrouping samples by respectively. However, the percent correct classification for morphotypes also suggested a high degree of segregation the Mexican specimens varied from 60% (Mexico-Chiapas, (Wilks’ lambda = 0.02521; F(20,196) = 51.91; P < 0.0001). The Mexico-Coatepec) to 90% (Mexico-Tuxtlas). Several confidence boundaries (95%) of the Mexican and South individuals from Mexico were misclassified with other American ellipses did not overlap, whereas those of the Mexican localities, but never with Colombian or Argentinian Brazilian and Andean morphotypes overlapped slightly (fig. samples, or with the Brazilian sample from Santa Catarina. 6). In this case, the percentage of correct classifications of However, one Mexican specimen from Quintana Roo was individuals by morphotype was high: the Mexican grouped with the Brazil-Sâo Paulo sample, and one morphotype grouped 98.6% of individuals originating from specimen from this Brazilian locality was classified within Mexico; the Brazilian morphotype grouped 96.7% of the Mexico-Tuxtlas sample. individuals belonging to Brazil and Argentina; and the The graph of individual factor scores obtained by Andean morphotype included 100% of individuals from contrasting the first two functions (CV-1 and CV-2) Colombia. Pair-wise comparisons among the morphotype distinguished at least three groups defined here as centroids based on Mahalanobis distances, showed that the morphotypes. The ‘Mexican morphotype’ included all divergences between the Mexican/Brazilian and the specimens from the seven Mexican localities; the ‘Brazilian Andean/Brazilian types were similar in both cases (MD = 44.3 morphotype’ grouped the two Brazilian plus the Argentinian and 37.5 respectively), whereas the divergence between the samples; and the ‘Andean morphotype’ was represented by Mexican/Andean types was extremely large (MD = 105.9). 494 V. Hernández-Ortiz et al.

Key to separate the morphotypes of the Anastrepha fraterculus complex characterized in the present study 1. Mean of canonical variable CV-1 always > 0; length of aculeus tip (A2+A3) ranged from 0.266–0.292 mm; serrated section of the aculeus (A3) 0.147–0.167 mm (fig. 7a–d); width of anterior apical band (W3) 0.412–0.444 mm; S and V-bands usually well connected (fig. 8a–d)...... Mexican type – Mean of canonical variable CV-1 always < 0; length of aculeus tip (A2+A3) 0.233–0.263 mm; serrated section of the acuelus (A3) 0.126–0.144 mm (fig. 8e–h); width of anterior apical band (W3) 0.276–0.384 mm; S and V-bands usually well separated (fig. 7e–h)...... 2 2. Mean of canonical variable CV-2 < 0; width of anterior apical band (W3) 0.28 ± 0.02 mm; both arms of V-band separated anteriorly (fig. 7 h)...... Andean type – Mean of canonical CV-2 > 0; width of anterior apical band (W3) 0.35–0.38±0.04 mm; both arms of V-band usually connected anteriorly (fig. 7e–g)...... Brazilian type

Discussion The results showed that there were statistically significant differences between Mexican and South American populations of the A. fraterculus complex in respect of aculeus shape (length of tip, length of serrated section, and mean number of teeth) and wing morphology (width of S-band, connection between S and V-bands, and the connection of the arms of the V-band). In contrast to what has been reported for South American populations (Stone, 1942; Baker et al., 1944; Zucchi, 1981), the degree of morphological variation among Mexican populations sampled in this study was extremely low. As a consequence, the Mexican populations were identified as a single morphotype by discriminant function analysis. A second group, consisting of the Colombian population, and a third group consisting of the Brazilian populations plus the single Argentinian population, were also clearly distinguishable. These findings lead to the conclusion that the Mexican, Andean and Brazilian morphotypes characterized here will eventually be recognized as three distinct taxonomic entities. A comparison of these results with related behavioural, karyotypic, isozyme and DNA studies (i.e. Mendes, 1958; Bush, 1962; Morgante et al., 1980; Steck, 1991; Steck & Sheppard, 1993; McPheron et al., 1999; Aluja et al., 2003), suggest that sufficient evidence has now accrued to justify naming a new Mexican species from within the A. fraterculus complex.

Mexican populations Discriminant analysis showed that the Mexican samples were morphologically homogeneous, with Mahalanobis distances between group centroids ranging from 3.3 to 30.7. Generally, larger distances were found when comparing the Tuxtlas sample with the other Mexican populations (23.1–30.7). Thus, it is clear that this sample showed the greatest morphometric divergence at the micro-geographical scale in Mexico. Morphological relationships among the Mexican samples were possibly inﬂuenced by local

MEX-QRoo MEX-Jica MEX-Teoc MEX-Apaz MEX-Chis MEX-Tuxt MEX-Coat BRA-SPau BRA-SCat COL-Cundecological ARG-Tucu and biogeographical features. Samples from Tuxtlas and Quintana Roo in Mexico turned out to be mor-

Pair-wise comparisons among all samples expressed in Mahalanobis distances. comparisons among all samples expressed Pair-wise phologically closer to the sample from Sâo Paulo, Brazil than . to any other Mexican populations possibly due to similarities in the surrounding tropical rain forest vegetation

Table 4 Table Groups Mex-QRooMex-JicaMex-Teoc17.1Mex-Apaz 0 Mex-ChisMex-Tuxt 9.0Mex-Coat 10.7 14.1Bra-SPau 16.9Bra-SCat 23.2Col-Cund 12.8Arg-Tucu 0 47.8 3.3 0 15.6 59.3 153.2 23.1 99.1 11.4 61.6 14.5 162.9 75.5 4.9 30.5 122.9 5.6(Sousa 74.4 186.5 100.6 0 13.9 151.5 26.0& Cabrera, 11.1 156.2 64.8 74.3 126.3 0 30.7 1983; 177.0 8.4 71.1 89.1 145.0Hernández-Ortiz 0 145.6 28.5 43.1 62.7 97.9 169.7 & 73.4 0 138.4 93.4Pérez-Alonso, 79.1 0 29.9 15.1 63.0 36.6 0 0 53.8 0 Morphometric analysis in the Anastrepha fraterculus complex 495

Fig. 5. Graph contrasting discriminant functions CV1 and CV2 among populations of Anastrepha fraterculus from Mexico (, Quintana Roo; , La Jicayana; , Teocelo; , Apazapan; , Chiapas; –, Tuxtlas; , Coatepec), Brazil (, Sao Paulo; , Santa Catarina), Colombia ( , Cundinamarca) and Argentina (+, Tucumán) using 19 morphometric variables in the full data set.

Fig. 6. Scatterplot factor scores of discriminant function analysis for the Mexican (), Brazilian () and Andean () morphotypes of the Anastrepha fraterculus complex. Centroids with 95% conﬁdence ellipses. 496 V. Hernández-Ortiz et al.

Fig. 7. Typical wing pattern of individuals of the Anastrepha fraterculus complex from: a, Mexico-Apazapan; b, Mexico-Tuxtlas; c, Mexico-Quintana Roo; d, Mexico-Chiapas; e, Brazil-Sâo Paulo; f, Brazil-Santa Catarina; g, Argentina-Tucuman; and h, Colombia- Cundinamarca.

1993; Souza Filho et al., 2000). Recent studies have shown All populations that the Mesoamerican rain forests from Mexico possess At least three discrete morphotypes can be distinguished close relationships with the Brazilian humid forests at the macro-geographical scale from Mexico to Argentina. (Rzedowski, 1988). The ‘Mexican morphotype’ represented by all the samples In spite of the fact that Mexican populations were from Mexico is characterized by a longer aculeus tip associated with different ecological and biogeographical (A2+A3) with a longer serrated section (A3) and a greater conditions (i.e. different altitude, climate, host species, and number of teeth (A8), a wider anterior apical wing band vegetational community), they showed great morphological (W3), and by a connection between the S and V-bands on the similarity and are considered to represent a single wings (W5). The ‘Brazilian morphotype’ can be morphotype distinct from the South American samples. distinguished by the short length of the aculeus (mean Thus, the Mexican samples cannot be considered to be a 1.598 mm), and by the reduced number of teeth along the cryptic species complex, as occurs in South America. This serrated section (9.3 teeth per side); whereas the ‘Andean conclusion is also supported by the results of McPheron et al. morphotype’ appears to have the shortest serrated section (1999), who analysed mitochondrial DNA sequences of (0.126 mm), and the narrowest anterior apical band to the samples of the A. fraterculus complex from nine different wing (0.276 mm). locations from Mexico to Brazil. These authors concluded The results of the morphometric analysis suggest that that Mexican and Central American samples were identical, South American samples of A. fraterculus show not only suggesting that populations from both regions represented a great morphological divergence among populations, but single species. also represent a complex that includes several morphotypes. Morphometric analysis in the Anastrepha fraterculus complex 497

Fig. 8. Aculeus tip shape of individuals of the Anastrepha fraterculus complex from: a, Mexico-Apazapan; b, Mexico-Tuxtlas; c, Mexico- Quintana Roo; d, Mexico-Chiapas; e, Brazil-Sâo Paulo; f, Brazil-Santa Catarina; g, Argentina-Tucuman; and h, Colombia-Cundinamarca.

The Colombian population constituted a discrete taxon addition to the molecular techniques that were currently taken from the Andes Mountains that is probably more being used (Steck & Sheppard, 1993). closely related to the Andean populations from Venezuela Given that in this study the ‘Mexican morphotype’ of A. and Peru examined by Steck (1991). Steck demonstrated fraterculus was clearly characterized by morphological strong genetic differentiation among these samples with uniformity among its populations (see also Baker et al., respect to those from Mexico, Costa Rica and the Bahia 1944), and that there is also a genetic basis (McPheron et al., region of Brazil. unpublished), this morphotype should from now on be The idea that currently known populations of A. consistently separated from A. fraterculus populations fraterculus constitute a cryptic species complex (Stone, 1942; occurring in South America. Further studies based on a Morgante et al., 1980; Malavasi & Morgante, 1982; Steck, larger number of samples from South America are required, 1991; Steck & Sheppard, 1993; Santos & Mattioli, 1996), is so that a wider range of morphometric variation can be thus reinforced by the data presented here. The methods and considered. In this way, the geographical distribution limits morphological characters utilized in this study have of each morphotype within the complex can be determined provided useful tools in distinguishing these species, in more precisely. 498 V. Hernández-Ortiz et al.

Acknowledgements (Anastrepha spp., Diptera: Tephritidae) y sus plantas hospederas en los Andes venezolanos. Revista de la Facultad The authors wish to thank Roberto A. Zucchi de Agronomía 3, 45–49. (Universidade de Sâo Paulo, Brazil), Ildelbrando Nora Bush, G.L. (1962) The cytotaxonomy of the larvae of some (Consejo Nacional de Pesquisas, Experimental Station of Mexican fruit flies in the genus Anastrepha. Psyche 68, 87–101. Caçador, Brazil), Sergio Ovruski (Consejo Nacional de Hernández-Ortiz, V. (1992) El género Anastrepha Schiner en Investigaciones Científicas y Técnicas, Argentina) and Jorge México (Diptera: Tephritidae): Taxonomía, distribución y Varón (Universidad Nacional de Colombia) for providing us sus plantas huéspedes. Instituto de Ecología, Publicación 33, with valuable samples from their respective countries. México, 162 pp. Similarly, we would like to recognize individuals who Hernández-Ortiz, V. (1996) Tephritidae (Diptera). pp. 603–617 in contributed material for this study from Mexico: Jaime Llorente, J.E., García Aldrete, A. & González, E. (Eds) Piñero, Carlos Estrada, Pedro Juárez, Abel Bañuelos M. and Biodiversidad, taxonomía y biogeografía de artrópodos de México: Pedro Xool Cetz. We extend a special thanks to Miguel hacia una síntesis de su conocimiento. Universidad Nacional Equihua Z. and Vinicio Sosa F. (Instituto de Ecología A.C.) Autónoma de México, México. for their useful comments and suggestions on earlier Hernández-Ortiz, V. & Aluja, M. (1993) Listado de especies del versions of the manuscript. Thanks are also due to Dan género neotropical Anastrepha (Diptera: Tephritidae) con Bennack for translating the manuscript from Spanish to notas sobre su distribución y plantas hospederas. Folia English and for many helpful suggestions. We are especially Entomológica Mexicana 88, 89–105. grateful to Jesús Reyes Flores, former director of the Hernández-Ortiz, V. & Pérez-Alonso, R. (1993) The natural host Campaña Nacional contra Moscas de la Fruta (Mexico), for plants of Anastrepha (Diptera: Tephritidae) in a tropical rain his constant support and for having commissioned this forest of Mexico. Florida Entomologist 76, 447–460. study. This research received financial support from the Lutz, A. & Lima, da Costa A.M. (1918) Contribuçao para o project ‘Morphological variability, genetics, behaviour, and estudo das Tripaneidas (moscas de frutas) brazileiras. food preferences in populations of Anastrepha fraterculus’, Memorias do Instituto Oswaldo Cruz 10, 5–15. from the Direccion General de Sanidad Vegetal (SAGARPA) Lima, da Costa A.M. (1934) Moscas das frutas do genero and the Instituto de Ecología A.C. (INECOL) under Anastrepha Schiner, 1868 (Diptera: Trypetidae). Memorias do agreement DGSV-SAGARPA – IICA – INECOL (904–74). Instituto Oswaldo Cruz 28, 487–575. Malavasi, A. & Morgante. J.S. (1982) Genetic variation in References natural populations of Anastrepha (Diptera: Tephritidae). Revista Brasileira de Genetica 5, 253–278. Adsavakulchai, S., Baimai, V., Prachyabrued, W., Grote, P.J. & McNamee, S. & Dytham, C. (1993) Morphometric Lertlum, S. (1998) Morphometric study using wing image discrimination of the species Drosophila melanogaster analysis for identification of the Bactrocera dorsalis complex (Meigen) and D. simulans (Sturtevant) (Diptera: (Diptera: Tephritidae). Journal of Biology 3, 1–6. Drosophilidae). Systematic Entomology 18, 231–236. Aluja, M. (1994) Bionomics and management of Anastrepha. McPheron, B.A., Han, H.Y., Silva, J.G. & Norrbom, A.L. (1999) Annual Review of Entomology 39, 155–178. Phylogeny of the genera Anastrepha and Toxotrypana Aluja, M., Guillén, J., De la Rosa, G., Cabrera, M., Celedonio, H., (Trypetinae: Toxotrypanini) based upon 16S rRNA Liedo, P. & Hendrichs. J. (1987) Natural host plant survey of mitochondrial DNA sequences. pp. 343–361 in Aluja, M. & the economically important fruit flies (Diptera: Tephritidae) of Chiapas, Mexico. Florida Entomologist 70, 329–338. Norrbom, A.L. (Eds) Fruit flies (Tephritidae): phylogeny and Aluja, M., Piñero, J. López, M., Ruíz, C., Zúñiga, A., Piedra, E., evolution of behavior. Boca Raton, Florida, CRC Press. Díaz-Fleischer, F. & Sivinski, J. (2000) New host plant and Mendes, L.O.T. (1958) Observaçoes citologicas em moscas das distribution records in Mexico for Anastrepha spp., frutas. Bragantia 17, 29–39. Toxotrypana curvicauda Gerstaecker, Rhagoletis zoqui Bush, Morgante, J.S., Malavasi, A. & Bush, G.L. (1980) Biochemical Rhagoletis sp., and Hexachaeta sp. (Diptera: Tephritidae). systematics and evolutionary relationships of neotropical Proceedings of the Entomological Society of Washington 102, Anastrepha. Annals of the Entomological Society of America 73, 802–815. 622–630. Aluja, M., Pérez-Staples, D., Macías-Ordoñez, R., Piñero, J., Nasca, A.J., Terán, A.L., Fernández, R.V. & Pasqualini. A.J. McPheron, B.A. & Hernández-Ortiz, V. (2003) Nonhost (1981) Animales perjudicales y benéficos a los cítricos en el status of Citrus sinensis cultivar Valencia and C. paradisi Noroeste Argentino. 62 pp. Centro de Investigaciones para la cultivar Ruby Red to Mexican Anastrepha fraterculus Regulación de Organismos Nocivos, Argentina. (Diptera: Tephritidae). Journal of Economic Entomology 96, Norrbom, A.L. (1985) Phylogenetic analysis and taxonomy of the 1693–1703. cryptostrepha, daciformis, robusta and schausi species Araujo, E.L., Nascimento, F.M. & Zucchi. R.A. (1998) Utilizaçâo groups of Anastrepha Schiner (Diptera: Tephritidae). 355 da análise discriminante em estudos taxonômicos de pp. PhD dissertation (unpublished), Pennsylvania State moscas-das-frutas do gênero Anastrepha Schiner, 1868 University, USA. (Diptera: Tephritidae). Scientia Agricola 55, 105–110. Norrbom, A.L. (2004) Host plant database for Anastrepha and Baker, A.C. (1945) Studies on the Mexican fruit fly known as Toxotrypana (Diptera: Tephritidae: Toxotrypanini). Diptera Anastrepha fraterculus. Journal of Economic Entomology 38, Data Dissemination Disk (CD-ROM) 2: 5984 records. 95–100. Norrbom, A.L., Zucchi, R.A. & Hernández-Ortiz, V. (1999) Baker, A.C., Stone, W.E., Plummer, C.C. & McPhail, M.A. Phylogeny of the genera Anastrepha and Toxotrypana (1944) A review of studies on the Mexican fruit fly and (Trypetinae: Toxotrypanini) based on morphology. related Mexican species. United States Department of pp. 299–341 in Aluja, M. & Norrbom, A.L. (Eds) Fruit flies Agriculture Miscellaneous Publication 531, 1–155. (Tephritidae): phylogeny and evolution of behavior. Boca Raton, Briceño, A. (1975) Distribución de las moscas de las frutas Florida, CRC Press. Morphometric analysis in the Anastrepha fraterculus complex 499

Perero, J.G., Nasca, A.J. & Stilinovic, D. (1984) Introducción a fraterculus (Wied.); advances in artificial rearing, taxonomic un estudio morfológico-taxonómico de especímenes de status and biological studies. International Atomic Energy Anastrepha fraterculus Wiedemann colectados sobre Agency, IAEA-TECDOC-1064. distintos hospederos en la Provincia de Tucumán. Steck, G.J. & Sheppard, W.S. (1993) Mitochondrial DNA pp. 421–455 in Anales II Congreso Internacional de variation in Anastrepha fraterculus. pp. 9–14 in Aluja, M. & Biomatemáticas, Facultad de Ciencias Exactas y Naturales, Liedo, P. (Eds) Fruit flies: biology and management. New York, Buenos Aires, Argentina. Springer Verlag. Reyment, R.A., Blackith, R.E. & Campbell, N.A. (1984) Stone, A. (1942) The fruit flies of the genus Anastrepha. United Multivariate morphometrics. 233 pp. London, Academic States Department of Agriculture Miscellaneous Publication Press. 439, 1–112. Rzedowski, J. (1988) Vegetación de México. 432 pp. Editorial Sugayama, R.L., Branco, E.S., Malavasi, A., Kovaleski, A. & Limusa, México, Cuarta reimpresión. Nora, I. (1997) Oviposition behavior of Anastrepha Santos, P. & Matioli, S.R. (1996) Mitochondrial and ribosomal fraterculus in apple and diel pattern of activity in an apple DNA variation in Brazilian populations of three Anastrepha orchard in Brazil. Entomologia Experimentalis et Applicata 83, species (Diptera: Tephritidae). pp. 46–47 in Proceedings of 239–245. the Second Meeting on fruit flies of the Western Hemisphere, Willig, R.M., Owen, R.D. & Colbert, R.L. (1986) Assessment of Viña del Mar, Chile. morphometric variation in natural populations: the Selivon, D. (1996) Estudo sobre a diferenciaçâo populacional em inadequacy of the univariate approach. Systematic Zoology Anastrepha fraterculus (Wiedemann) (Diptera: Tephritidae). 35, 195–203. 137 pp. PhD thesis (unpublished), Instituto de Biociencias, Zar, J.H. (1999) Biostatistical analysis. 663 p. New Jersey, Prentice- Universidade de Sâo Paulo, Brazil. Hall, Inc. Solferini, V.N. & Morgante, J.S. (1987) Karyotype study of eight Zucchi, R.A. (1978) Taxonomia das especies de Anastrepha Schiner, species of Anastrepha (Diptera: Tephritidae). Carylogia 40, 1868 (Diptera: Tephritidae) assinaladas no Brasil. 105 pp. (PhD 229–241. thesis (unpublished), Universidade de Sao Paulo, Escola Sousa, M.S. & Cabrera, E.F. (1983) Listados Florísticos de México: Superior de Agricultura Luiz de Queiroz, Piracicaba, II. Flora de Quintana Roo. 100 pp. Instituto de Biología, Brazil. Universidad Nacional Autónoma de México, México. Zucchi, R.A. (1981) Anastrepha Schiner, 1868 (Diptera: Souza Filho, M.F., Raga, A. & Zucchi, R.A. (2000) Moscas-das- Tephritidae): novas sinonimias. Revista Brasileira de frutas nos estados Brasileiros: Sâo Paulo. pp. 277–283 in Entomologia 25, 289–194. Malavasi, A. & Zucchi, R.A. (Eds) Moscas-das-frutas de Zucchi, R.A. (1999) Espécies de Anastrepha, sinonimias, plantas importância econômica no Brasil: conhecimento básico e aplicado. hospedeiras e parasitóides. pp. 41–48 in Malavasi, A. & Brazil, Holos Editora. Zucchi, R.A. (Eds) Moscas-das-frutas de importância econômica Statistica (1999) Software for PC statistical analysis. Statsoft, Inc., no Brasil: conhecimento básico e aplicado. Holos Editora, Tulsa, Oklahoma. Brasil. Steck, G.J. (1991) Biochemical systematics and population genetic structure of Anastrepha fraterculus and related species (Diptera: Tephritidae). Annals of the Entomological Society of America 84, 10–28. Steck, G.J. (1999) Taxonomic status of Anastrepha fraterculus. (Accepted 8 July 2004) pp. 13–20 in The South American fruit fly, Anastrepha © CAB International, 2004