Sycophila pistacina ( Hymenoptera : Eurytomidae ) : A valid species

Sycophila pistacina (Rondani), which was previously synonymized with Sycophila biguttata (Swederus), is revalidated. Morphological, morphometric and molecular data confirm its status as a separate species. Diagnostic characters are provided for distinguishing it from S. biguttata. The nomenclature of the S. biguttata complex is updated. 137 * Corresponding author; e-mail: gerard.delvare@cirad.fr our observations suggest that S. pistacina is univoltine. Although Megastigmus pistaciae is bivoltine in Iran (Davatchi, 1956), it seems to be mostly univoltine in the western Mediterranean Basin (Grissell & Prinsloo, 2001; pers. observ.). We observed females of S. pistacina flying around clusters of Pistacia and ovipositing in the pods in late summer (August–September) while the host pupates at this time (Grissell & Prinsloo, 2001). Third, the biogeographical history of the associated plants is very different. The genus Pistacia belongs to the Anacardiaceae, a tropical family of plants. P. lenticus is even considered to be a relict component of the Tertiary subtropical flora present in the Mediterranean Basin (Axelrod, 1975) and known from the Oligocene–Miocene period (35–40 millions years bp). Its life-cycle and reproductive strategy retain archaic features characteristic of tropical plants (Verdú & García-Fayos, 2002). In addition, the genus Pistacia supposedly originated from Central Asia (Parfitt & Badenes, 1997). Conversely, European deciduous oaks are important components of temperate forests. They belong to the Quercus s. str. section, for which the center of diversification is in North America (Manos et al., 1999). In addition, the re-colonization of Europe soon after the last glaciation started from a couple of southern refuges and is therefore quite recent (Demesure & Musch, 2001; Kremer & Petit, 2001; Brewer et al., 2002) in contrast to the presence of Pistacia over millions of years (Axelrod, 1975). The biogeographic features of the hosts of Sycophila mimic the pattern described above for the plants. The Cynipidae are typical temperate taxa and the species associated with oaks are particularly specialized, while the Megastigmus of the pistaciae species group probably originated from Africa (Grissell & Prinsloo, 2001). The aims of this study are: (1) investigate the possibility of separating eurytomids belonging to the Sycophila biguttata complex into discrete entities; (2) provide reliable features for the recognition of these entities; (3) update the nomenclature of the biguttata complex. We looked first for morphological characters and then did a morphometric study in order to confirm the reliability of the characters selected. This data was finally compared with molecular data obtained from an analysis of a couple of genes. This study is a part of a revision of the European Eurytomidae undertaken by GD. MATERIAL AND METHODS Collecting sites and sampling Several dozens of specimens were used in the morphological study. They were collected in Corsica, France, Hungary, Iran, Spain and the Netherlands. Specimens identified as S. biguttata were reared from cynipid galls on Quercus pubescens and Q. brantii. Those identified as S. pistacina were collected from Pistacia terebinthus or P. lentiscus, together with Megastigmus pistaciae; specimens of S. iracemae were mostly collected from Q. pubescens and determined by morphological comparison with paratypes; they were not directly associated with cynipid galls. S. biguttata was identified by comparison with a series housed in the BMNH and identified as such by Claridge. Twenty five fresh specimens of S. biguttata and S. pistacina and two specimens of S. iracemae were used for the morphometric and molecular studies (Table 1). After removing fore wings and hind legs for mounting on slides and measuring, the rest of the body was preserved in 95% EtOH at –20°C and used for the genetic study. Eurytoma amygdali Enderlein, Eurytoma caninae Lotfalizadeh & Delvare (Lotfalizadeh et al., 2007) and Sycophila variegata (Curtis) were used as outgroups. Abbreviations: BMNH – Natural History Museum, London, UK; CBGP – Centre de Biologie pour la Gestion des Populations, Montferrier-sur-Lez, France; GD – Gérard Delvare personal collection, Montpellier, France; OXUM – Hope Department, Oxford University, Oxford, UK; IEE – Instituto Español de Entomología, Madrid, Spain; INRA – Institut National de la Recherche Agronomique, Montpellier, France; MZLU – Lund University, Zoological Museum, Lund, Sweden; 138 * Identification was based on the morphological characters indicated by the initial morphological study. H176H181H182 3 1 3 IRAN: Lorestan, Ghalai, ex Chilaspis israeli on Quercus brantii P17-P18-P19 3 3 10 FRANCE: Corsica, Ghisonacchia, 10 m, swept from Pistacia lentiscus Sycophila variegata (Curtis) P5-P6-P7-H179 4 4 15 FRANCE: Hérault, Cazevieille, 230 m, swept from Pistacia terebinthus Sycophila pistacina (Rondani) B1-B2 2 2 2 FRANCE: Hérault, Saint-Martin-de-Londres, 220 m, swept from Quercus pubescens Sycophila iracemae Nieves-Aldrey B6-B9-B11 2 3 20 IRAN: Lorestan, Ghalai, ex Andricus grossulariae on Quercus brantii B3-B4 2 3 3 FRANCE: Var, Sainte-Zacharie, Sainte-Beaume, swept from Quercus pubescens Sycophila biguttata (Swederus) H169 1 1 1 IRAN:Azarbayjan-e-Sharghi, Marand, ex Diplolepis fructuum on Rosa canina Eurytoma caninae Lotfalizadeh & Delvare H106 1 1 1 SYRIA: Damas, iii.2005, ex seeds of Amygdalus cominus Eurytoma amygdali Enderlein (ITS2) (COI) Specimen code Obtained sequences Sample size Collection site Species* TABLE 1. Specimens of Sycophila used in the molecular study: collecting site, sample size, obtained sequences (for COI & ITS2) and codes used.

Sycophila biguttata (Swederus, 1795) is a common species, widely distributed in the Palearctic region (Noyes, 2002).Its biology is typical of the other European species: Mayr (1905) reared it from galls on oaks caused by 45 cynipids; Nieves-Aldrey (1983) and Pujade-Villar (1993) reared it respectively from 24 and 14 different cynipid galls.The host genera are Andricus, Cynips, Biorhiza and Neuroterus, which generally form galls on the buds of deciduous oaks (Quercus faginea, Q. robur, Q. pubescens and Q. pyrenaica); S. biguttata is rarely reared from cynipids galls collected on evergreen oaks, such as Q. suber in the Mediterranean Region.Askew (1984) records this species as an endoparasitoid of cecidogenic cynipid larvae.Available data (Claridge, 1959;Nieves-Aldrey, 1983) clearly demonstrate that this species is bivoltine, completing one generation in autumn and winter at the expense of the agamous generation of its cynipid host and another generation in spring and early summer at the expense of the sexual generation.Bou ek (1974), in his study of the chalcidoids described by Rondani (1872), deals with a species of eurytomid described as Tineomyza pistacina.Rondani quoted Palumbina guerinii (Stainton) (as P. terebinthella, a junior synonym) -a gelechiid moth infesting the pods of Pistacia terebinthus and P. vera (Mourikis et al., 1998) -as the host of this wasp.Megastigmus pistaciae Walker (Chalcidoidea: Torymidae) is another pest of Pistacia spp., which also develops within the seeds (Davatchi, 1956;Traveset, 1993).It is widely distributed in the Mediterranean Basin and Central Asia and is quoted as a host either of Sycophila biguttata (Nikol'skaya, 1935), or of a species close to it (Davatchi, 1958).Another parasitoid of M. pistaciae is described as "Decatoma trogocarpi n. sp.?" (De Stefani, 1908).Our own observations and collection from the South of France and Corsica indicate that M. pistaciae and S. pistacina are present at the same sites in the same time, either on P. terebinthus or P. lentiscus.
Further, Nieves-Aldrey (1983) described S. iracemae from specimens reared from Andricus spp. in the Salamanca Region of Spain and subsequently reared from the same host in the Madrid Province (Gómez et al., 2006).The adults all emerged from the agamous generation of the cynipids.Nevertheless it was recently reared in France from galls of Pediaspis aceris (Gmelin) (Askew et al., 2006), which is confirmed by a recent rearing of this parasitoid by one of us (GD) from the sexual form of this cynipid wasp.
Bou ek (1974) synonymized T. pistacina with S. biguttata on the basis of intraspecific variation in body color, a feature already quoted by Mayr (1905) and Claridge (1959) for populations reared from cynipid galls on oaks.A number of biological and biogeographic facts led us to challenge this synonymy.First, the endoparasitic habit is considered to be a specialized feature resulting from a long coevolution between a parasitoid and its host.As the hosts belong to different superfamilies (Cynipoidea versus Chalcidoidea) it is hardly conceivable that the same species can overcome the different immune defenses.Second, S. biguttata is clearly bivoltine, while our observations suggest that S. pistacina is univoltine.Although Megastigmus pistaciae is bivoltine in Iran (Davatchi, 1956), it seems to be mostly univoltine in the western Mediterranean Basin (Grissell & Prinsloo, 2001;pers. observ.).We observed females of S. pistacina flying around clusters of Pistacia and ovipositing in the pods in late summer (August-September) while the host pupates at this time (Grissell & Prinsloo, 2001).Third, the biogeographical history of the associated plants is very different.The genus Pistacia belongs to the Anacardiaceae, a tropical family of plants.P. lenticus is even considered to be a relict component of the Tertiary subtropical flora present in the Mediterranean Basin (Axelrod, 1975) and known from the Oligocene-Miocene period (35-40 millions years bp).Its life-cycle and reproductive strategy retain archaic features characteristic of tropical plants (Verdú & García-Fayos, 2002).In addition, the genus Pistacia supposedly originated from Central Asia (Parfitt & Badenes, 1997).Conversely, European deciduous oaks are important components of temperate forests.They belong to the Quercus s. str.section, for which the center of diversification is in North America (Manos et al., 1999).In addition, the re-colonization of Europe soon after the last glaciation started from a couple of southern refuges and is therefore quite recent (Demesure & Musch, 2001;Kremer & Petit, 2001;Brewer et al., 2002) in contrast to the presence of Pistacia over millions of years (Axelrod, 1975).The biogeographic features of the hosts of Sycophila mimic the pattern described above for the plants.The Cynipidae are typical temperate taxa and the species associated with oaks are particularly specialized, while the Megastigmus of the pistaciae species group probably originated from Africa (Grissell & Prinsloo, 2001).
The aims of this study are: (1) investigate the possibility of separating eurytomids belonging to the Sycophila biguttata complex into discrete entities; (2) provide reli-able features for the recognition of these entities; (3) update the nomenclature of the biguttata complex.We looked first for morphological characters and then did a morphometric study in order to confirm the reliability of the characters selected.This data was finally compared with molecular data obtained from an analysis of a couple of genes.This study is a part of a revision of the European Eurytomidae undertaken by GD.

Collecting sites and sampling
Several dozens of specimens were used in the morphological study.They were collected in Corsica, France, Hungary, Iran, Spain and the Netherlands.Specimens identified as S. biguttata were reared from cynipid galls on Quercus pubescens and Q. brantii.Those identified as S. pistacina were collected from Pistacia terebinthus or P. lentiscus, together with Megastigmus pistaciae; specimens of S. iracemae were mostly collected from Q. pubescens and determined by morphological comparison with paratypes; they were not directly associated with cynipid galls.S. biguttata was identified by comparison with a series housed in the BMNH and identified as such by Claridge.
Twenty five fresh specimens of S. biguttata and S. pistacina and two specimens of S. iracemae were used for the morphometric and molecular studies (Table 1).After removing fore wings and hind legs for mounting on slides and measuring, the rest of the body was preserved in 95% EtOH at -20°C and used for the genetic study.Eurytoma amygdali Enderlein, Eurytoma caninae Lotfalizadeh & Delvare (Lotfalizadeh et al., 2007)  MZUF -Natural History Museum of the University of Firenze -Zoological Section "La Specola".

Morphometric study
Initially eight characters (Table 2) of the two hypothesized groups, respectively identified as S. biguttata and S. pistacina, were analyzed quantitatively.Width and length measurements indicate maximum dimensions unless stated otherwise.These dimensions were measured between homologous points on all specimens regardless of treatment (Figs 1-2).The base of the marginal vein could not be identified because the parastigma continues to the marginal vein and there is no evident angle between the two veins.The end of the colorless region of the parastigma was considered to be the distal limit of the parastigma (Fig. 2).All measurements were made with a Leica DMRD stereo-microscope and the LEICA QWin® software.
Measurements were log-transformed to allow for allometry and to normalize distributions.They were not size-standardized because size is an important taxonomic criterion, related with ecological correlates such as habitat selection and food regime.Size must be considered together with shape in morphometric studies.
After discarding the variance-covariance PCA (Principal Component Analysis) owing to differences in variance, standard PCAs using correlation matrices based on a 50 × 5 matrix were performed.Differences in shape between species were analyzed in different ways: (i) by looking to the eigenvectors following the first ones in the preceding PCA; (ii) by performing a PCA on three log-transformed ratios calculated from the primary variables (as ratios can be useful in morphometry despite allometry and autocorrelation problems); (iii) by performing a correspondence analysis (CA) on the 50 × 5 matrix.In CA, the standardisation of the variable values by their sum for each individual simply and efficiently removes most of the size effects and the results are similar to a doubly centred PCA.
Furthermore, MANOVA was used to determine the morphological separation of two species.It allows separation of the two species on the basis of a linear combination of dependant variables.
Statistical analysis of the data was carried out using Statistica (StatSoft).The mean values of all characters were compared using a Student's t-test with = 0.05.
Our sample of fresh S. iracemae specimens was too small for a combined morphometric/molecular study.A special quantitative study, which only considered the relative proportion of the head (width and height) provided additional information.For this, dried specimens of the three above species were used.They were segregated on the basis of characters validated in the morphological study.

Molecular study DNA extraction
Sources of DNA included 50 specimens of adults preserved in 95% ethanol and stored in a freezer at -20°C.The mitochondrial Cytochrome Oxidase I (COI) and nuclear Internal Transcribed Spacer 2 (ITS2) were used as genetic markers.
Genomic DNA was isolated by using the following protocol: first, single individuals were removed from ethanol and ground in an ependorf tube containing 50 µl CTAB buffer (2% CTAB, 1.4 M NaCI, 100 mM Tris-HCl (pH 8), 20 mM EDTA and 0.2% 2-mercaptoethanol); after which 150 µl of CTAB was added to each tube.Specimens in CTAB buffer were incubated at 65°C for 50 min.Proteins were removed by adding chloroform isoamyl alcohol and centrifugation for 10 min at 13,000 rpm.The supernatant was transferred to a clean tube and 150 µl cold isopropanol added and vortex-mixed.After precipitation for 60 min at -20°C, extracts were re-centrifuged for 30 min at 4°C and 13,000 rpm.The proteins were cleaned with 500 µl absolute ethanol then centrifuged for 10 min (4°C and 13000 rpm).Finally, DNA was resuspended in 30 µl of distilled water. 1 -ML = Length of marginal vein (Fig. 2) 2 -MWB = Width of marginal vein at base (Fig. 2) 3 -MWA = Width of marginal vein at apex (Fig. 2) 4 -SP3L = Length of penultimate spiniform seta on hind tibia (Fig. 1) 5 -TB3W = Width of hind tibia at apex (Fig. 1) I -Simple characters (linear measurements): TABLE 2. Characters used in the morphometric study.
The same conditions were used for the PCR amplification of ITS2, with following changes: denaturation for 2 min, 30 amplification cycles, 1 min denaturing at 94°C and 1 min annealing at 45°C.PCR products were separated on a 1.5% agarose gel.Purified DNA fragments were directly sequenced from both directions using an automated sequencer at CBGP.Primers used for amplification also served as sequencing primers.
All of the taxa used in this study are represented by secondary voucher specimens deposited in the INRA collection at Montpellier, France.

Sequence alignment
The alignment of the COI and ITS2 sequences was done with CLUSTAL (Thompson et al., 1997).The computer package BioEdit 7.0.5 (Hall, 2005) was used for manual inspection and correction of sequences according to the secondary structure of Drosophila yakuba.Aligned sequences were reformatted appropriately and entered into MEGA 3 (Kumar et al., 2004) for calculation of sequence statistics.Sequences were deposited in GenBank (accession numbers EU017316 to EU017329) and the aligned data matrix is available from the authors upon request.

Phylogenetic analysis
Kimura's two-parameter model of base substitution was used to calculate genetic distances with MEGA 3.1 software (Kumar et al., 2004).The mitochondrial haplotypes and ITS2 sequences were analyzed separately with maximum parsimony (MP) and neighbour-joining (NJ) analysis using PAUP*4.0b10(Swofford, 2001).MP was carried out through branch and bound search.Support for monophyly of clades was assessed by bootstrap values (Felsenstein 1985) with 1000 replicates (full heuristic search) using PAUP*.NJ analyses involved the K2P distance model (Kimura, 1980), but the results did not change with use of other models.

Morphology (Figs 5-16)
Sycophila biguttata, S. pistacina and S. iracemae are very similar.In side-by-side comparisons 2 qualitative and 3 quantitative characters placed specimens into 2 discrete groups, with S. pistacina and S. iracemae morpho-140 front and hind margins distally diverging (Fig. 13) front and hind margins parallel (Fig. 12) Marginal vein generally longer (Fig. 16) shorter (Fig. 15) Length of spiniform setae on hind tibia relative to width of hind tibia at distal end 1 row of areoles (Fig. 11) 2 rows of areoles (Fig. 10) Propodeum: basal sculpture of median furrow carinate above and barely pilose (Fig. 11) narrow but not carinate, distinctly pilose (Figs 7, 8) Intertorular projection generally below 1.27 exceptionally over 1.3 (Fig. 6) over 1.3 (Fig. 5) Ratio head width : head height S. pistacina S. biguttata Characters TABLE 3. Morphological characters separating S. pistacina from S. biguttata.Fig. 4. Scatterplot of the head ratio "width/height" and "head width" for S. biguttata, S. iracemae and S. pistacina.logically indistinct.These characters are presented in Table 3. See also Figs 5-16.The intertorular projection in S. biguttata is slightly wider than in S. pistacina and S. iracemae.In these latter two species the intertorular projection is narrow, sharp and carinate above.It is also more distinctly pilose in S. biguttata, bearing 15-22 hairs while 6-8 hairs are visible in the other species (Figs 7-9).The median furrow of the propodeum is also different.In S. biguttata two rows of areoles are visible because a median ridge is present.In S. pistacina and S. iracemae, the ridge is absent and only one row of areoles can be seen .
However, S. pistacina can hardly be distinguished from S. iracemae.Only its marginal vein is slightly more thickened with the anterior and posterior margins diverging rather more sharply, and the vein is overall somewhat shorter .However these characters intergrade; they are very difficult to discern without experi-ence and require a reference collection.The relative proportions of the head can also be used (Fig. 4).In S. iracemae, the ratio of width/head height is 1.3-1.4,whereas in 80% of the specimens of S. pistacina it does not reach 1.25; unfortunately this ratio seems more susceptible to allometry and in a few cases it exceeds 1.30.

Morphometric study
Principal Component Analysis of the morphological variables resulted in a highly significant discrimination of the species.This analysis, using 7 characters (the length of marginal vein was rejected), provides strong statistical support for reliable species discrimination.In all but few Figs 12-16.12, 15 -Sycophila biguttata.13, 16 -Sycophila pistacina.14 -S.iracemae.12-14 -fore wing venation; 15-16hind tibia in lateral view.marginal cases the species can be directly separated using the proposed ratios.
Marginal separation was obtained between S. biguttata and S. pistacina based on the first and second components (PC1 & PC2) (Fig. 3).Although the remaining components in each analysis contained a significant proportion (8%) of the variance, they had no more diagnostic value than the first two components.The first component has high loadings of all variables (92%) and is assumed to represent size.The variable with a high loading on the first component was the length of the hind tibial spiniform seta, which is consistent with the univariate separation involving this variable.Univariate measurements, along any particular axis, are either overlapping or very marginally separate species (Table 4), exemplifying the problems associated with separating species by individual measurements.
More distinctive non-overlapping differences occur using a combination of characters.Two species (S. biguttata and S. pistacina) are distinct in all the bivariate plots.In spite of overlap in some of the studied characters, the ratio of the penultimate spiniform seta length to width of hind tibia (SP3L/TB3W) and length of marginal vein to its width at the apex (ML/MWA) is probably the best way to separate them (Table 6).There is a narrow zone of overlap between the two species in the range of the 3 ratios, which may be the result of their close relationship.

Molecular study
We obtained 18 and 19 sequences respectively for COI and ITS2 (Table 1).DNA variation between and within species of S. biguttata and S. pistacina was examined in a 814 bp segment of the mitochondrial COI and 493 bp segment of the nuclear ITS2.

COI
One sequence, 593 bp in length, included only half of this region.Including or excluding these shorter sequences did not substantially affect any of the results reported below because every species is represented by at least one long sequence.Of the 814 nucleotide sites in the dataset, 206 (25.3%) were variable and 151 (18.6%) parsimony informative.The second codon position with 89.8% of the non-variable sites is the most constant.The third codon position is highly variable with only 29.6% constant sites.The average frequency of A+T, variable sites and parsimony informative sites are the highest for the third codon position, respectively 90.2%, 66.7% and 66.5%, (Table 7).
Of the 553 bp, 451 (81.6%) characters are constant and the number of parsimony-informative characters is 54 (9.8%).The composition of each base was relatively homogeneous (Table 7).The average frequency of A+T nucleotides in ITS2 (54.1%) is lower than in the COI gene (74.5%).

Phylogenetic relationships
Whatever gene is considered, the specimens initially identified as S. pistacina branch on a node different from specimens attributed to S. biguttata.The species are hence clearly distinct.S. pistacina together with S. iracemae form a monophyletic and well supported group.With COI, the pair of specimens a priori identified as S. iracemae group together, supported by a high bootstrap value (95%).The genetic divergence from S. pistacina is very slight (0.9-1.0% mixed.The fact that S. pistacina exhibits no intraspecific variation is surprising, as one population was collected in the island of Corsica, which has been isolated from the mainland for a long time.Is this species quite stable genetically?Or was it recently introduced? S. pistacina + S. iracemae is a sister group of S. variegata using COI and of S. biguttata using ITS2.The first relationship is better supported by the bootstrap (100 versus 85) but conflicts with the overall morphology, as S. pistacina and S. biguttata are relatively similar.

CONCLUSION
Sycophila biguttata and S. pistacina, although superficially similar, are clearly distinct species.They can be discriminated by several morphological characters, some of them having been confirmed by morphometry.Moreover, the genetic evidence provided by the sequencing of two genes unambiguously confirms the separation of these species.
The distinction between S. pistacina and S. iracemae, either on morphological characters or morphometric variables, is much more problematic.However, we are reluctant to synonymize both names as the species exhibit quite distinct biologies and hosts.S. pistacina is definitively with Pistacia spp.and their chalcid seed-eater, M. pistaciae; the species is monovoltine, at least in Western Europe, and adults occur from July to October with maximum activity in August-September.S. iracemae is a parasitoid of the agamous generation of a few Andricus spp.(A.kollari and A. coriarius) galling deciduous oaks and apparently of the sexual form of Pediaspis aceris on Acer spp.; nevertheless further data are required to confirm this latter assessment.The adults occur from May to October and therefore a second generation might be produced.Further molecular data, based on populations reared under the same conditions as the holotype, are required to precisely resolve the status of S. iracemae.
The phylogeny of the whole complex is presently unresolved as the topologies achieved with COI and ITS2 do not coincide.A larger sample of the biguttata species group including populations of S. submutica (Thomson) and further populations of S. variegata is necessary to resolve this problem.At present it is only possible to indicate that the genes in the respective populations of S. biguttata and S. pistacina are quite stable.8. Divergence (below the diagonal) and number of differences (above the diagonal) in the COI and ITS2 sequences based on pair-wise comparisons among populations of the Sycophila biguttata complex and outgroups (2 Eurytoma spp.).Abbreviations.Eamyg -Eurytoma amygdali from Syria; Ecani -Eurytoma caninae from Iran; Syva -S.variegata; Sycir -S.iracemae from France; SybiF -S.biguttata from France; SybiI -S.biguttata from Iran; SypiF -S.pistacina from Southern France (Hérault); SypiC -S.pistacina from Corsica.
Eurytoma trogocarpi (De Stefani): Noyes, 2002. Eudecatoma mallorcae Hedqvist, 1962: 212.Type locality: Mallorca.New synonymy.Synonymized with S. biguttata by Bou ek, 1974: 263.We did not examine the type of Decatoma trogocarpi.It is quite evident from its distribution and biology that it is the same as S. pistacina.Ferrière (1968) re-described the species and underlined some of the characters quoted in the present paper to separate it from S. biguttata.The species is erroneously included in Eurytoma by Noyes (2002).We examined the type series of E. mallorcae and agree with Bou ek (1974), who considered it as belonging to the same species as S. pistacina.

Fig. 3 .
Fig. 3. Star shape of the first two principal components (PC1-PC2) from the principal component analysis of five of the linear characters from log-transformed data set.Groups 1 and 2 define a priori taxa, S. biguttata and S. pistacina, respectively.Fig.4.Scatterplot of the head ratio "width/height" and "head width" for S. biguttata, S. iracemae and S. pistacina.

TABLE 1 .
Identification was based on the morphological characters indicated by the initial morphological study.Specimens of Sycophila used in the molecular study: collecting site, sample size, obtained sequences (for COI & ITS2) and codes used. *

TABLE 4 .
).For ITS2, the two populations are Univariate statistics, in micrometers, of the simple characters used in the morphometric analysis of Sycophila spp.

TABLE 7 .
Summary of nucleotide composition (%) and number of variables sites in the COI and ITS2 data of the Sycophila species examined.