Phylogenetic information reveals the peculiarity of Caryedon serratus ( Coleoptera , Chrysomelidae , Bruchinae ) feeding on Cassia sieberiana DC ( Caesalpinioideae )

Studies over the past 20 years on natural populations indicated that sympatric speciation may be far more common and widespread among plants and animals than previously thought. By using molecular phylogenetics (on a combined data set of two genes), the relationships between Caryedon serratus native forms and forms feeding on groundnut was investigated. The specific objectives were to clarify the taxonomic status of C. serratus feeding on Cassia sieberiana DC. Morphological analysis was conducted, using parts of the adult specimens which were dry-mounted for the studies. Morphological study comprised extraction of the genitalia, which were cleared and mounted in Canada balsam following standard procedures. Results showed a strong differentiation of insects from different hosts’ trees, with specimens from C. sieberiana possibly representing a sibling species.

populations where there is a lot of migration.However, migration is not the only parameter to be taken into account when the genetic structuralization of a species is being studied.The demography of the populations constitutes a parameter as important as the migration.For example, at rate of equal migration, the differentiation between populations by genetic drift is made much quicker in small-sized populations than in populations with big strength.
The case of the groundnut seed-beetle, Caryedon serratus-Olivier (Coleoptera: Chrysomelidae Bruchinae) is interesting to study.The identity of C. serratus (Olivier, 1790) was established by Decelle, who published a comprehensive synonymy in 1966.It has a small number of native host plants in Africa such as Bauhinia rufescens, Bauhinia (Piliostigma) reticulata, Bauhinia (P.) thonningii, tamarind (Tamarindus indica), which all belong to the leguminous subfamily Caesalpinioideae.Less than a century ago, this beetle added groundnut (Arachis hypogaea) (Leguminoseae Papilionoideae) to its diet, thus Sembène et al. 1471 Table 1.Population sampling: origin, host plant, abbreviations and numbers of scored individuals (between parentheses).
In Asia, C. serratus is usually known as the tamarind seed-beetle (Arora, 1977), sometimes as a pest of groundnuts (Vazirani, 1975;Dick, 1987) and it is often also reported as feeding on the seeds of various species of Acacia (Mimosoideae) or Cassia (Caesalpinioideae) (Mukerji and Chatterjee, 1951;Arora, 1977;Vir and Jindal, 1996).However, Indian specimens of tamarind and/or groundnut seed beetles have often been known as Caryedon gonagra (Fabricius, 1798).In 1966, Decelle synonymized gonagra and serratus, partly on the basis of Southgate and Pope's (1957) redescription and illustration of C. gonagra.In 1894, Decaux gave a quite precise description of the morphology and biology of a tamarind seed-beetle, which he named Caryoborus tamarindi.The name tamarindi has been in use in Egypt (Shomar, 1963) and in India (Vazirani, 1975).
Reviews in the mid-1990s recognized eight confirmed host species associated with C. serratus, all of them belonging to the Caesalpiniaceae subfamily with the exception of peanut, A. hypogaea L. (Fabaceae) (Pierre and Huignard, 1990;Nilsson and Johnson, 1992;Delobel, 1995).C. serratus is widely distributed throughout Africa, from Senegal to South Africa and in southern Asia (Johnson, 1966).In the neotropics and Hawaii, it is not associated with groundnuts (Robert, 1985;Nilsson and Johnson, 1992;Delobel, 1995).C. serratus is thought to have shifted to groundnuts from trees of the Caesalpiniaceae family in West Africa in the early 1900s (Delobel, 1995).The ability of C. serratus to colonize groundnuts (introduced from South America to Africa towards the end of the 19th century) must be considered in connection with recent advances in our knowledge of the phylogeny of the genus Caryedon.Morphological and molecular data indicated that C. serratus belongs to a clade of Mimosoideae feeding species and is therefore not directly related with other Caesalpinioid feeders (Silvain and Delobel, 1998).It may be hypothesized that C. serratus has acquired a more performant secondary compound detoxification ability than its Mimosoideae feeding relatives and that this ability gave it the potential to expand its host spectrum beyond Mimosoideae (ancestral host range) and Caesalpiniodeae (extended host range).
In this study, interest is on clarifying the taxonomic status of beetles feeding on C. sieberiana.In order to solve the problem, the DNA of C. serratus specimens reared from groundnuts, B. rufescens, C. sieberiana, Piliostigma reticualtum and T. indica in Senegal and Mali was analyzed.For this, comparisons used involved both allopatric (but feeding on the same host) and sympatric (but feeding on different hosts) seed-beetle populations.

C. serratus samples
Beetles used in this study were reared from eggs, larvae or pupae on/inside pods collected on the different host trees between 2004 and 2007 (Table 1).The pods of the five hosts of C. serratus (B. rufescens, C. sieberiana, P. reticulatum, T. indica) were nondehiscent.Females laid their eggs on the surface of ripe pods and newly hatched larvae bore through the husk and into the seed, where larval development took place.Five to seven weeks after oviposition, mature larvae exit the pods and fell to the ground where they pupate after spinning a whitish silk cocoon.Emergence of the adults took place about two weeks after pupation.The period during which ripe pods of P. reticulatum and C. sieberiana remain on the tree does not exceed, in a given area, 5 or 6 months, which is usually between November and May.After pods have fallen to the ground, they are often eaten by cattle or burnt and only a few seeds remain accessible to C. serratus females.Towards the end of the dry season and during the rainy season, seeds become progresssively less available.When the first new P. reticulatum or C. sieberiana pods reach maturity after the rainy season, in November, C. serratus levels were at their lowest.Infestation rates observed on P. reticulatum at that time were not higher than 2 to 4 eggs per 1000 pods (Sembène, 1997).On hosts which bear pods all year round (B.rufescens and to a lesser extent, T. indica), generations follow one another in the pods attached to the tree and infestation rates do not exhibit the wide fluctuations observed on Piliostigma.Populations, however, experience a definite decline during the rainy season, probably due to increased adult mortality (Pierre and Huignard, 1990;Ndiaye, 1991).Pods of the derived host groundnut were removed from the soil shortly after the end of the rainy season, in October.Oviposition took place on drying pods within a week after harvest.Infestation rates at that time were very low, in the order of 1 egg per 10000 seeds.After the harvest has been taken to the stores, infestation continued and population levels rapidly increased.
Pods were taken from the trees as soon as they reached maturity, at the start of new infestations (between November and January).Groundnut samples were collected from the field during drying.Adult beetles were genetically analysed almost immediately after they emerged from pods.Firstly, however, the genital parts were examined following Prevett (1965) in order to avoid confusion with Caryedon crampeli Pic (C.cassiae in Prevett), a species which also feeds on B. rufescens, C. sieberiana and P. reticulatum.Voucher specimens were kept in the I.F.A.N.Cheikh Anta Diop (Dakar) collections.In all, a total of 25 different populations (individuals of both sexes) were genetically scored in the study.

DNA extraction
Total genomic DNA was extracted from prothorax of individual seed-beetle as described by Vogler et al. (1993).The abdomen, elytra and antennae were kept apart to avoid contamination by fungi and nematodes and to permit subsequent morphological observations.

mtDNA amplification and PCR-amplified
A partial CytB end region was PCR-amplified.The CytB end region consisted of primer CB1 (5' -TATGTACTACCATGAGGACAAATATC-3') and CB2 (5'-ATTACACCTCCTAATTTATTAGGAAT-3').The 25 µl PCR reaction mixtures contained 2.5 µl enzymes buffer supplied by manufacturer, 2.5 mM MgCl2, 0.6 units of Taq polymerase, 17.5 pM of each primer, 25 nM of each DNTP and 1 µl of DNA extract.After an initial denaturation step at 92°C for 3 min, reaction was subjected to 35 cycle of 1 min at 92°C; 1 min 30 s at 48°C and 1 min at 72°C.A pool of two PCR was concentred and loaded on a 1.3 del agarose gel.The PCR gel band was cut and then purified with Quiaquick PCRgel purification kit (Qiagen) and directly sequenced on an Abi 373 automated sequencer using TaqFS and Dye-labelled terminators (Perkin-Elmer).

Nuclear DNA amplification
ITS1 region was PCR-amplified, using the primer CIL (5' GCGT TCGAARTGCGATGATCAA 3') and CIU (5'GTAGGTGAACCTGCA GAAGG3') developed by Vogler and DeSalle (1994).PCR conditions were similar to those used for mitochondrial domain, except that the elongation step was increased to 1 min 30 s.

Data analysis
Sequence alignment was performed using ClustalW (Thompson et al., 1994) as implemented in BioEdit.Aligned sequences were entered in McClade 3.06 (Maddison and Maddison, 1992) for subsequent treatments.The resulting data were subsequently used to calculate Kimura 2-parameter genetic distances between the haplotypes.Gtests (log-likelihood ratio test) were performed to test genetic distances homo-geneity among hosts at the same site and among sites for the same host (Sokal and Rohlf, 1981).
Phylogenetic relationships were reconstructed with PAUP 4*B8 (Swofford, 2001) using the maximum parsimony method (MP).The MP analysis was done with the heuristic search option of 50 random stepwise taxon addition replicates, using the branch swapping tree bisection-reconnection (TBR) option.A bootstrap procedure (1000 iterations with the same option of heuristic search) was used to establish the score of each node (Felsenstein, 1985) by retaining group compatible with the 50% majority rule consensus.A strict consensus tree was computed whenever multiple equally parsimonious trees were obtained.
The molecular clock hypothesis was tested following Posada and Crandall (2001), by comparing the log-likehood scores of the best trees (Shimodaira and Hasegawa, 1999), using reel Boostrap with 1000 replications used to test for a significant difference between the score of the different trees obtained.In addition, analyses were conducted using the distance-matrix method with the Neighbour-Joining (NJ) algorithm (Saitou and Nei, 1987) on Kimura 2parameter distances with PAUP 4*B8 (Swofford, 2001).The robustness of inferences was assessed through bootstrap resampling (1000 replicates).
The two analyses were completed separately with Cyt.B and ITS1.The partition homogeneity test (Farris, 1994), as implemented in PAUP was used to determine the appropriateness of combining both partial Cyt.B and ITS1 genes into a single analysis with the same option.The Wilcoxon's signed-Rank test (Templeton, 1983) was applied to compare the statistical significance of the best tree produced by each tree reconstruction method to one another.
Two outgroup were used in this study: Bruchidius atrolineatus which belongs to the family Bruchidae and C. gonagra (F.) obtained from Bauhinia variegate L. from Egypt (BVC).C. gonagra is a species which may be synonymous with C. serratus (Delobel et al., 2003).

Morphological analysis
Parts of the adult specimens were dry-mounted for morphological studies.Morphological study comprised extraction of the genitalia, which were cleared and mounted in Canada balsam following standard procedures.As often as feasible, slide preparations of male genitalia were made with internal sac everted.This gave a clearer view of the different sclerites and spines that are attached to the sac wall and of their relative position.It thus avoided possible misinterpretations due to variable positions of sclerites in variously retracted internal sacs.Slides were studied through a Leica DMLS light microscope and photographs were taken with a Nikon Coolpix 990 digital camera.Drawings were made with Adobe Illustrator 9.0, using photographs as models.

DNA sequences
populations was obtained.The alignment was straightforward and involved no insertions/deletions.The sequences could be unambiguously aligned and showed 20 different haplotypes due to 51 polymorphic sites.Of these sites, 96% were parsimony informative.The number of nucleotide differences in pair wise comparisons of C. serratus populations ranged from 0 to 17.2%.Between C. serratus sampled on C. sieberiana (called "Sieberiana") and the other (called "Serratus"), it ranged from 14.4 to 15.8%.In "Serratus" this number ranged from 0 to 0.09%.Within the same host species, the number ranged from 0 to 0.02%.
The PCR products obtained for the ITS1 domain were 874 pb in "Serratus" and 950 pb in "Sieberiana" while in both cases, 6 pb in 18S and 49 pb in 5.8S.The sequences could be unambiguously aligned.Of these sites, 35.4% were variable and were parsimony informative.The number of nucleotide differences in pairwise comparisons of C. serratus populations ranged from 0 to 24.6%.Between "Sieberianae" and "Serratus", this number ranged from 24.9 to 35.7%.In "Serratus" this number ranged from 0 to 0.1%.Within the same host species, the number ranged from 0 to 0. 02%.In "Sieberiana" the number of nucleotide differences ranged from 0 to 0.003%.
All sequences have been deposited in GenBank under accession numbers, AY449464.The two sets of sequences could be unambiguously aligned together and were used in the analysis described below.

Distance matrix and phylogenetic trees
Among the C. serratus, the genetic distances measured on the total alignment of CytB + ITS1 (1418 pb) ranged between 0 and 0.204, but clearly fall in two classes: "Serratus" from 0 to 0.037 and "Sieberiana" from 0.186 to 0.195, separating C. serratus into two major groups.One group ("Serratus") comprises haplotypes sampled in A. hypogaea, B. rufescens, P. reticulatum and T. indica.The second group (Sieberiana") gathers the haplotypes raised from C. sieberiana.The genetic distance between the two exceeds genetic distances typical of geographical populations or subspecies in invertebrates (Klein and Seitz, 1994;Emelianov et al., 1995;Kerdelhué et al., 2002).
Genetic distance between hosts was highly significant (0.295) and due, for a large part, to sample obtained from C. sieberiana, which differed from all other samples.In another study, after considering that C. serratus samples raised from C. sieberiana might represent a separate species, they were excluded from the original data set and genetic distances within and between host-plants and within and between localities were reanalysed.Genetic distance values fell but remained significant (P < 0.001) in both cases (Sembène, 2006 andSembène et al., 2008 for commentary).
The MP analysis on Cyt B nucleotide data yielded 29 equally parsimonious trees that were 845 steps long.The Sembène et al. 1473 same methods on the ITS1 data set yielded seven equally parsimonious trees, 245 steps long (CI = 0.963; RI = 0.981).All analyses of combined data set yielded clades, one for all "Serratus" specimens, whatever their host and a second for all "Sieberiana" specimens.Bootstrap values were all over 50%.Separation according to host plant was clear.C. sieberiana samples showed a similar trend.Similar patterns of relationships were obtained with NJ analysis.Samples typically clustered according to host plant.C. sieberiana samples were clearly separated from all other samples and showed high bootstrap values.These confirmed results given by Sembène and Delobel (1998), Sembène (2006) and Sembène et al., (2008) for Senegalese specimens and in particular the relative genetic isolation of populations on B. rufescens and the close relationship existing between P. reticulatum and A. hypogaea populations.C. sieberiana samples showed a similar trend.These beetles were phylogenetically closer to C. gonagra than that of "Serratus".
The genetic distance between C. serratus and C. gonagra was computed from a portion of the Cyt.B gene that was successfully used to distinguish species in the rodent genus, Praomys (Lecompte et al., 2002), or aphids in the genus, Aphis (Coeur-d'Acier, unpublished).The present data indicated that Senegalese and Egyptian specimens of Caryedon belong to two distinct clades.
For each data set, the topology obtained with MP and NJ methods was compared using the Shimodaira-Hasewaga test and the Wilcoxon's signed-rank test.These tests did not support a significant difference between the trees.The strict consensus trees of MP and NJ trees are presented in Figure 1 for Cyt.B; in Figure 2 for ITS1 and in Figure 3 for the pooled data.

Morphological analysis
A careful study of the external and genital morphology showed that variation occurred between samples of different geographical origins.Colour appeared as rather variable within populations.The density of dark markings ranged from almost absent, with entirely reddish elytral integument, to very dense black mottling on elytra that is almost entirely black underside with the legs being variously dotted with black.However, a constant trend existed in the colour of legs, with Egyptian and Asian specimens showing more contrasted femora and tibiae.Interesting diagnostic characters were the shape of the hind femur, which was more or less oblong and the number of teeth of its ventral pecten.Morphology of male genitalia provided several diagnostic characters, in particular the fine morphology of sclerites in the internal sac and the size and density of spines in its apical part were large and numerous in African specimens but smaller and much less numerous in Asian specimens.Female genitalia, particularly the relative size of dorsal and ventral vaginal sclerites, also showed constant differences.In West African specimens, these sclerites were almost of equal size, whereas Egyptian specimens had a much smaller dorsal sclerite.Size and density of spines in the bursa copulatrix were also a good diagnostic character.

Di ag n o si s "Serratus"
It is a large greyish-brown species, with small black spots, and pale antennae and legs.It has about 10 -12 teeth at the ventral side of the hind femur.Hind femur is 2.2 times longer than its width.

a) Red esc ri p ti o n
"Serratus" length was 3.7 -6.2 mm; and width, 1.9 -3.2 mm (excluding head).Fresh insects had greyish brown colour.The integument was reddish with small, scattered black or dark brown spots, merging here and there into larger, irregular spots.Antennae and legs (except hind femora and tibiae) were pale.Antennomeres 1 and 5 -11 were often darker on disk.Their vestiture was dense, but not quite covering the integument, and it was recumbent except for a few erect setae, especially on pygidium.Setae were pale greyish over the red parts of the integument and blackish or dark brown over dark spots.A few areas of the setae had denser dark spots.The third apex of the elytra, especially along the suture, basal half of pygidium, ventro-apical part of hind femora; first The male species had a short head, constricted back eyes, bulging eyes, maximum head width about 1.7 times width for back eyes; ocular sinus well defined; eye width about 2.5 times minimum distance between eyes, sharp median longitudinal carina on frons, vertex without interocular tubercle.Antenna reached the fifth base of the elytra; antennal segments 1-4 were cylindrical.Segments 5-10 were serrate, segment 11 was oblong; segment 1 about twice as long as segment 2; segment 3 was 1.6 times longer than 2, segment 4 hardly shorter than 2, segments 5 -10 as long as antennomere 1 (9 -10 slightly longer), but widened at the apex; segment 11 is 1.4 times longer than 1 and about 2.8 times as long as the width.
Pronotum was about 1.6 to 1.7 times as wide and as long, with the greatest width at the base.Sides were almost parallel at base, straight or slightly concave, then abruptly converging at about two third of the length, disc feebly convex; punctures on disc irregularly spaced, setous; distances between punctures varied from 1 to 3 diameters; cuticle between coarse puncturation with fine punctures.
Elytra was about 1.5 times as long as their combined width; sides regularly convex; disc convex; striae on disc thin and deep, punctured; punctures elongated, with setae, distances between punctures less or about equal to their diameter; interstriae convex, with strong micropunctation and a very small number (varying from 0 to 2 per interstria) of coarser punctures.
Legs were without sexual dimorphism; hind femora strongly incrassate, at their widest (at base of first spine) 2.2 times longer than width; mesoventral margin with a series of blunt teeth not hidden by hair, followed by a pecten of 10 -12 sharp teeth, with first tooth 1.5 to 2 times longer than second; hind tibiae was acute, with 5 carinae complete; apex of hind tibia with mucro was as long as tibial width at apex (lateral view).
Abdomen was without modified setae; sternite 5 emerges to about one third of its length; pygidium was 1.2 times as long as width.Females were similar to male; fifth sternite did not emerge.
One specimen has been seen from Réunion Island: St Paul, Boucan Canot, 5m, 6.v.1952 (J. Hamon), identified as C. serratus Ol. by J. Decelle in 1971 [coll. IRD, Paris]; however the specimen, a female, has no genitalia and cannot therefore be assigned to one or the other species with certainty.

d ) Affi n i ti es
C. serratus is part of a small group of species with similar Sembène et al. 1477 genitalia and elytral pattern.Typical of this species are the shape of the larger hooks, the presence of two pairs of thin, and strongly curved sclerites at apex of internal sac.It is most closely related with C. acaciae sensu Decelle, from which it may be distinguished by the absence of modified setae on the first two abdominal sternites.It is also closely related with C. longispinosus sensu Decelle (nomen nudum) (Silvain and Delobel, 1998;Delobel et al., 2000), which is characterized by a particularly long first tooth of pecten at hind femur, but has a similar genital morphology.

"Sieberianae"
It is a large greyish-brown species, with small black spots, and pale antennae and legs; it has about 13 -14 teeth at ventral side of hind femur.Hind femur is 1.8 times times longer than its width.

a) Redescription
Its length is 2.8 -5.2 mm; width: 1.5 -2.6 mm (excluding head).Fresh insects had a greyish brown colour.The integument was reddish with small, scattered black or dark brown spots, merging here and there into larger, irregular spots.Antennomeres 1 and 5-11 were often darker on disk.Its vestiture was dense, but not completely covering integument, it was recumbent except for a few erect setae, especially on pygidium.Setae were pale greyish over red parts of the integument and blackish or dark brown over dark spots.A few areas had denser dark spots; the third apex of elytra, especially along suture, basal half of pygidium, ventro-apical part of hind femora; first abdominal sternites are often almost entirely black.Fore and median legs with apex of femur and central two-thirds of tibia were darker.Pronotum is often devoid of black spots.Apex and median line of pygidium had whitish setae.The male species had, a short head, constricted back eyes; bulging eyes, maximum head width about 1.7 times width behind eyes; ocular sinus well defined; eye width about 4 times minimum distance between eyes; sharp median longitudinal carina on frons, vertex without interocular tubercle.Antenna reached the fifth base of the elytra; antennal segments 1 -4 cylindrical; segments 5 -10 serrate; segment 11 oblong; segment 1 about twice as long as segment 2; segment 3 1.6 times longer than 2; segment 4 equal to 2; segments 5 -10 as long as antennomere 1, but widened at apex; segment 11 about 1.2 times longer than 1 and 2.4 times as long as width.
Pronotum was about 1.6 to 1.7 times as wide and as long, with greatest width at base.The sides were almost parallel at base, straight or slightly concave, then abruptly converging at about two third of their length, disc feebly convex; punctures on disc irregularly spaced, setous.Distances between punctures varied from 0 to 2 diameters; cuticle between coarse puncturation with fine punctures.
Elytra was about 1.5 times as long as their combined width; sides regularly convex; disc convex; striae on disc thin and deep, punctured; punctures elongated, with setae, distances between punctures less or about equal to their diameter; interstriae convex, with strong micropunctation.
Legs were without sexual dimorphism; hind femora strongly incrassate, at their widest (at base of first spine) 1.8 times longer than width; mesoventral margin with a series of blunt teeth not hidden by hair, followed by a pecten of 12-14 sharp teeth, with the first tooth 1.5 to 2 times longer than the second, to which it is usually coalescent.Hind tibiae strongly acute, with 5 carinae complete; apex of tibiae with mucro about as long or slightly longer than tibial width.
Abdomen was without modified setae; sternite 5 emerge to about one third its length; pygidium 1.2 times as long as width.Females were similar to male; fifth sternite did not emerge.

DISCUSSION
Different levels of genetic differentiation between cooccurring Bauhinia, Cassia, Piliostigma and Tamarindus wild populations and groundnut populations of C. serratus have being identified There is evidence of at least four distinct biotypes in Senegal, with restricted gene flow between each other.This genetic differentiation observed in C. serratus (a phytophagous insect) is necessarily accompanied by the installation of pre or post-zygotic reproductive isolation in the absence of extrinsic barriers to gene flow.The main factors that could possibly maintain allele frequency differences among ancestral populations and between them and derived groundnut populations include the following: post-mating reproductive isolation between races, differential larval survivorship within the seed, differential host recognition by adult beetles, temporal and spatial differences in adult emergence and host seed availability.Many of these mechanisms are discussed in Sembène (2006) and Sembene et al. (2008).

Serratus versus sieberianae
A strong differentiation clearly exists between C. serratus feeding on C. sieberiana and on the four other host plants.The average value of Kimura genetic distance between the two groups (0.198) exceeds genetic distances typical of sympatric (Kerdelhué et al., 2002) and geographical populations or subspecies in invertebrates (Klein and Seitz, 1994).The individuality of "C.sieberiana" samples may be explained by an incipient sympatric speciation or by the existence on C. sieberiana of a sibling species of C. serratus as a result of a previous diffe-rentiation of populations in allopatry, with presentday overlap of C. sieberiana and other host plants.These hypotheses cannot be tested without further knowledge of biogeographic and temporal variations in host plant distribution.However, host race formation mechanisms as temporal and spatial differences in adult emergence and host availability, certainly play a major part in maintaining homogeneity within the C. sieberiana population because C. sieberiana trees bear pods over a short period of time, from December to February.C. serratus females keep laying eggs on the pods after abscission, but populations must then face interspecific competition with congeneric C. crampeli Decelle.In mid-January 1997, a sample of C. sieberiana pods (on the tree) which yielded 1000 adults seed beetle, of which 87 were C. serratus and 13 were C. crampeli was collected in Sokone (16°22W, 13°53N).A sample of fallen pods collected in March under the same tree yielded a total of 327 beetles, of which only 56 (17%) were C. serratus and the rest (83%) were C. crampeli.It may be assumed that this drastic population reduction had negative impact on the beetle's capability to utilize host plant species.In the same order, hybridization experiments between individuals from C. sieberiana and P. reticulatum, which were genetically the most distant indicated that post-mating isolation is not likely, as laboratory F1 progeny were fertile but significant differential survivorship of larvae of the different host populations was observed.The fertility of the beetles reared on the same host plant was significantly greater than that obtained after sexual crossings (Sembène, 2004).However, reproducing in the laboratory does not indicate that the two "species" interbreed in nature as the two gene pools seem isolated.The peculiarity of C. sieberiana-associated forms raises the question of an appropriate taxonomic classification of this group.A 76 bp difference existed in the ITS1 nucleotides between "Sieberianae" and "Serratus".The genetic distance observed between the two entities was higher or about existing between species of the same genera."Sieberianae" is genetically closer to C. gonagra than C. serratus (Delobel et al., 2003) -Sieberianae" amplifies only three of the 10 microsatellites loci identified in "Serratus" (Sembène et al., 2003) -"Sieberianae" is the only one to present a duplication of locus ADH1 (Sembène et al., 1998).
There are several significant controversies associated with using molecular techniques to study systematics and evolution.These included debates over the relative importance of molecular versus morphological data, and the constancy of the molecular clock

Figure 1 .
Figure 1.Phylogenetic relationships among nucleotide sequences of partial cytochrome b gene (468 pb) of 25 populations of Careyedon serratus.This tree is between a maximum parsimony (MP) consensus tree and a Neighbour-Joining (NJ) consensus tree.The first numbers above or under branches are MP% bootstrap values (1000 replicates) and the second numbers, the NJ % bootstrap values (1000 replicates).Bruchidius atrolineatus (BrA) and Caryedon gonagra (BVC) are the outgroups.

Figure 2 .
Figure 2. Phylogenetic relationships among nucleotide sequences of the total ITS1 of the total gene (950 pb) of 25 Caryedon Serratus populations.This tree is the consensus between a maximum parsimony (MP) consensus tree and a neighbour-Joining (NJ) consensus tree.The first numbers above branches are MP% bootstrap values (1000 replicates) and the second numbers are the NJ% bootstrap values (1000 replicates).Bruchidius atrolineatus (BrA) and Caryedon gonagra (BVC) are the outgroups.

Figure 3 .
Figure 3. Phylogenetic relationships among nucleotide sequences of the pooled data (partial cytochrome b and ITS1 genes) of 25 Careyedon serratus populations.This tree is the consensus between a MP consensus tree and a Neighbour-joining (NJ) consensus tree.The first numbers above branches are MP% bootstrap values (1000 replicates) and the second values are the NJ% bootstrap values replicates.Bruchidius atrolineatus (BrA) and Caryedon gonagra (BVC) are the outgroups.