Species of Characidotrema Paperna & Thurston, 1968 (Monogenea: Dactylogyridae) from fishes of the Alestidae (Characiformes) in Africa: new species, host-parasite associations and first insights into the phylogeny of the genus

Background African tetras (Alestidae) belonging to Brycinus Valenciennes are known to be parasitized with monogeneans attributed to two genera, Annulotrema Paperna & Thurston, 1969 and Characidotrema Paperna & Thurston, 1968 (Dactylogyridae). During a survey of monogeneans parasitizing alestids, species of Characidotrema were collected in Cameroon, D. R. Congo, Senegal, South Africa, Sudan and Zimbabwe. This paper provides new morphological data and the first molecular analysis broadening our knowledge on the diversity of these parasites. Results Seven species (four known and three new) of Characidotrema are reported from two species of Brycinus: C. auritum n. sp. and C. vespertilio n. sp. from B. imberi (Peters); and C. brevipenis Paperna, 1969, C. nursei Ergens, 1973, C. pollex n. sp., C. spinivaginus (Paperna, 1973) and C. zelotes Kritsky, Kulo & Boeger, 1987 from B. nurse (Rüppell). Species identification was based on morphological analysis of the sclerotized structures supported by nuclear ribosomal DNA (partial 18S rDNA, ITS1, and 28S rDNA) sequence data. Morphological analysis confirmed that the most apparent character distinguishing species in the genus is the morphology of the male copulatory organ and vagina. Observations on the haptoral sclerotized elements of these parasites by means of phase contrast microscopy revealed the presence of a sheath-like structure relating to the ventral anchor, a feature that supplements the generic diagnosis of Characidotrema. Maximum Likelihood and Bayesian analyses of the large subunit (28S) rDNA sequences recovered Characidotrema species isolated from the two Brycinus hosts as monophyletic, and indicated a closer relationship of this group to monogeneans parasitizing African cyprinids (Dactylogyrus spp.) and cichlids (species of Cichlidogyrus Paperna, 1960, Scutogyrus Pariselle & Euzet, 1995, and Onchobdella Paperna, 1968) than to those from catfishes (species of Quadriacanthus Paperna, 1961, Schilbetrema Paperna & Thurston, 1968 and Synodontella Dossou & Euzet, 1993). The overall agreement between the morphological diversification of the MCOs and the molecular tree observed in this study indicates that significant phylogenetic signals for clarifying relationships among species of Characidotrema are present in the characteristics of the MCO. Conclusions It seems that intra-host speciation is an important force shaping the present distribution and diversity of Characidotrema but further studies are necessary to confirm this hypothesis and assess questions related to the phylogeny of these parasites. To identify potential co-speciation events, co-phylogenetic analyses of these monogeneans and their alestid hosts are required.


Background
African tetras (Alestidae) belonging to Brycinus Valenciennes are known to be parasitized by species of Annulotrema Paperna & Thurston, 1969 (17 spp.) and Characidotrema Paperna & Thurston, 1968 (9 spp.) (see [1,2]). Species of Characidotrema are easily differentiated from those of Annulotrema by possessing a robust body with a poorly developed haptor and highly modified ventral anchor-bar complex. The presence of a ventral anchor with a diagonally truncate or scoopshaped point is an unusual feature that is unique among African dactylogyrids. However, a similarly modified ventral anchor also occurs in species of Jainus Mizelle, Kritsky & Crane, 1968, a genus parasitizing characiform fishes in South America. In fact, Kritsky et al. [3] resurrected Characidotrema, so far considered as a junior synonym of Jainus by Paperna [4]. Although both genera possess similar characteristics and parasitize fishes of the Characiformes, Kritsky et al. [3] emended the diagnosis of Characidotrema and clearly differentiated its species from those of Jainus by the comparative morphology of the haptoral sclerites and the presence of dextral vagina (vs sinistral vagina in Jainus).
During our investigation of gill parasites of alestid fishes from Cameroon, D. R. Congo, Senegal, South Africa, Sudan and Zimbabwe, four previously described and three new species of Characidotrema were collected from the gills of Brycinus imberi (Peters) and B. nurse (Rüppell). Herein, all species found are reported and described using two complementary approaches: a morphological study of the hard, sclerotized structures (i.e. those of the haptor and the distal parts of the female and male reproductive systems), and a molecular study using ribosomal DNA sequences (partial 18S-ITS1 and partial 28S rDNA sequences). In addition, to study the placement of Characidotrema among other genera of dactylogyrids parasitizing African freshwater fishes, phylogenetic analyses based on sequences of 28S ribosomal RNA gene were performed.

Fish collection
Fish hosts were collected by beach seine net or hookand-line or purchased at local fish markets in six African countries: Cameroon, D. R. Congo, Senegal, South Africa, Sudan and Zimbabwe (for localities and coordinates, see Table 2) during the period 2005-2017. Host names recorded here are those provided in FishBase [12]; the names used in the original descriptions are retained in parentheses as synonyms. Live fishes (only the ones captured by net or hook-and-line) were kept in aerated holding tanks until processed for parasitological examination; fishes were sacrificed by severing the spinal cord.

Parasite collection and identification
Monogeneans were collected from the gills of freshly killed fishes using fine needles and processed as in Francová et al. [13]. Parasite specimens used for morphological study of the sclerotized structures (the haptoral and reproductive hard parts) were completely flattened using coverslip pressure and fixed with a mixture of glycerine and ammonium picrate (GAP). Specimens used for DNA analyses were bisected using fine needles under a dissecting microscope, and subsequently, the posterior half of the body was placed in a 1.5 ml Eppendorf tube with 96% ethanol for genomic DNA extraction. The anterior body part containing the male copulatory organ was completely flattened under coverslip pressure and fixed with GAP for species identification.
The mounted specimens (or their parts) were studied using an Olympus BX 61 microscope equipped with phase contrast optics. Drawings were made with the aid of a drawing attachment and edited with a graphic tablet compatible with Adobe Illustrator and Adobe Photoshop. Measurements, all in micrometres, were taken using digital image analysis (Stream Motion, version 1.9.2) and are given as the range followed by the mean and number (n) of specimens measured in parentheses. Schemes of measurements of the hard structures are provided in Fig. 1. Numbering of hook pairs (in Roman numerals I-VII) is that recommended by Mizelle [14]. Male copulatory organ is henceforth abbreviated to MCO.
After morphometric analysis, specimens (or their parts) fixed with GAP were remounted in Canada balsam according to Ergens [15] and deposited as typematerial, morphological vouchers and hologenophores (see [16] for terminology). For comparative purposes, type-specimens of three previously described species of Characidotrema (syn. Jainus) deposited in the Royal Museum for Central Africa, Belgium (RMCA/MRAC), and one species deposited in the Institute of Parasitology, Czech Academy of Science, České Budějovice, Czech Republic (IPCAS), were examined. Type-and voucher specimens of monogeneans collected for the present study were deposited in the helminthological collection of the IPCAS (see below for the details).

Genetic distances and phylogenetic reconstruction
Alignments of 18S, ITS1 and 28S were generated using MAFFT v.7 [18] and manually adjusted in BioEdit [19]. Interspecific genetic distances were determined using distance matrices (uncorrected p-distances) in MEGA 7 [20] separately for each genetic marker. To determine the position of Characidotrema spp. among other representatives of African dactylogyrid genera, phylogenetic analyses were conducted using Maximum Likelihood (ML) and Bayesian Inference (BI) methods. Phylogenetic reconstruction including 34 selected African dactylogyridean species (Table 4) was based on sequences of partial 28S rDNA only. Three species belonging to the Anoplodiscidae and Diplectanidae were used as the outgroup. The 28S sequences were aligned using MAFFT v. 7 [18] with the Q-INS-i algorithm [21] and other parameters set to default. Gaps and ambiguously aligned regions were removed from the alignment using GBlocks v. 0.91b [22] applying all options for a less stringent selection. ML analysis was conducted using the IQ-TREE [23] on the W-IQ-TREE web server [24]. The default "Auto" setting was selected to determine the best-fit substitution model. The TIM3+F+I+G4 model was selected as the optimal model of molecular evolution based on the Bayesian information criterion (BIC). Branch support was estimated using ultrafast bootstrap approximation [25] with 10,000 replicates. BI analysis was performed in MrBayes 3.2.1 [26]. However, since the selected substitution model is not implemented in MrBayes, the GTR+I+G was selected as the closest matching alternative. Four simultaneous chains (one cold and three heated) of the Markov Chain Monte Carlo (MCMC) algorithm were run twice for 10 7 generations. Tree topologies were sampled every 100 generations, whereby the first 25% of trees from each run were discarded as "burn-in". The remaining trees were used to construct majority-rule consensus trees and determine the Bayesian posterior probability (BPP) for each clade. The trees were visualized and edited in FigTree ver. 1.4.3. [27].

Molecular characterisation
The sequence of the 18S-ITS1 region of C. brevipenis was 978 bp long, of which 483 bp corresponded to the partial 18S rDNA and 495 bp corresponded to the ITS1 region. The nucleotide sequence of the partial 28S region was 761 bp long. Four specimens of C. brevipenis from Sudan (White and Blue Nile) were sequenced; no intraspecific sequence variation was found.

Differential diagnosis
On the basis of the comparative morphology of the sclerotized structures, our specimens are conspecific with the type-specimens of C. brevipenis (MRAC M. T.

A) collected from Brycinus nurse in Ghana by
Paperna [7]. Measurements of the haptoral and copulatory sclerites of the present specimens fall within the ranges provided by Kritsky et al. [3] in their redescription of C. brevipenis from B. cf. nurse in Togo. Characidotrema brevipenis most closely resembles C. undifera Kritsky, Kulo & Boeger, 1987 and C. zelotes Kritsky, Kulo & Boeger, 1987 in the general morphology of the haptoral sclerites and the MCO [3]. It differs from these two species by having a longer (39 vs 34 µm in C. undifera; 39 vs 28-29 µm in C. zelotes) and more smoothly curved copulatory tube (a tube with a tighter curve and recurved termination is present in C. undifera and C. zelotes). Kritsky et al. [3] recognized also the morphological similarity between C. brevipenis and C. nzoiae (Paperna, 1979), stating that C. brevipenis differs from the latter species by having a more elongate base of the copulatory tube. Nevertheless, the most apparent feature that distinguishes C. brevipenis from all the above-mentioned congeners is the goblet-shaped vagina.

Molecular characterisation
The combined 18S-ITS1 sequence of C. nursei was 979 bp long, of which 483 bp corresponded to the 18S rDNA and 496 bp corresponded to the ITS1 region. The sequence of partial 28S rDNA was 836 bp long. Ten specimens from Sudan (Blue Nile) were sequenced and no intraspecific variability was found.

Differential diagnosis
This species was originally described from the gills of Brycinus nurse in Egypt by Ergens [9]. Later, but in the same year, Paperna [4] proposed Jainus longipenis from the same host in Uganda. Under the revision of Characidotrema, Kritsky et al. [3] considered Jainus longipenis and Jainus cf. longipenis of Paperna [8] [10] suggest that these authors were dealing with a different species. In their paper, the MCO is depicted as J-shaped with a comparatively shorter copulatory tube than that in C. nursei (40-

Molecular characterisation
The combined 18S-ITS1 sequence of C. spinivaginus was 942 bp long, of which 483 bp corresponded to the 18S rDNA and 459 bp corresponded to the ITS1 region. The sequence of partial 28S rDNA was 762 bp long. Two specimens were sequenced from Sudan (White Nile: Kosti), and no intraspecific variability was observed.

Differential diagnosis
This species was incompletely described (without the drawings of haptoral structures) by Paperna [4,8] as Jainus spinivaginus from B. nurse in Ghana and Uganda. Kritsky et al. [3] subsequently transferred the species to Characidotrema and supplemented the original description with drawings of both the ventral and dorsal anchorbar complexes based on a re-examination of the holotype. However, they did not provide drawings of the hooks, probably because of the poor condition of the holotype. Morphological comparisons of the present specimens with the holotype of C. spinivaginus (MRAC M. T. 35.942) confirmed that all are conspecific. Also, the measurements of the sclerotized structures of the present specimens do not appear to vary significantly from those originally reported by Paperna [4,8] since only two specimens were originally measured. Thus, the present findings of C. spinivaginus on the gills of B. nurse (Senegal, Sudan) represent new host and geographical records for this species. In addition, the taxonomically important sclerotized structures are illustrated here in one figure for the first time and the following information from the present specimens supplements the original description: (i) the accessory piece guiding the distal portion of the copulatory tube is present (not mentioned/depicted in the original description); (ii) the fingerlike basal process, while clearly illustrated in the original drawing of the MCO, was mistakenly considered as an accessory piece by Paperna [8]; and (iii) delicate hooks are characterized by an undilated shank and poorly developed thumb (not described by Paperna [8]).

Differential diagnosis
The sclerotized structures of the present specimens correspond both in size and morphology to those in the original description of C. zelotes by Kritsky et al. [3].  Etymology: The specific epithet (a noun) is from the Latin (pollex = thumb) and refers to the thumb-like process arising from the base of the copulatory tube.

Molecular characterisation
The combined 18S-ITS1 sequence of C. pollex was 882 bp long, of which 483 bp corresponded to the 18S rDNA and 399 bp corresponded to the ITS1 region. The sequence of partial 28S rDNA was 829 bp long. Three specimens were sequenced from Sudan (White and Blue Nile) and no intraspecies variability for 28S rDNA 18S-ITS1 was observed.

Differential diagnosis
Characidotrema pollex n. sp. belongs to the group of congeners having a copulatory tube with a finger-like basal process that is not articulated to the accessory piece: C. brevipenis Paperna, 1969; C. nursei Ergens, 1973; C. nzoiae (Paperna, 1979); C. regia Birgi, 1988 (doubtful due to poor original drawing); C. ruahae (Paperna, 1979); C. spinivaginus (Paperna, 1973  between these species. Characidotrema pollex n. sp. differs from all the above congeners by having J-shaped copulatory tube with shortened base and relatively small finger-like process, which rises from the base at an angle of about 40°. It most resembles C. ruahae, from which it differs by the latter species possessing a comparatively longer finger-like process, rising from the base at an angle of about 90°. Characidotrema pollex n. sp. is morphologically similar to Jainus cf. longipenis of Paperna (1979) reported on Brycinus leuciscus (syn. Alestes leuciscus) in Ghana. Although the voucher specimen of this form was not examined, illustrations of the haptoral structures and MCO were provided by Kritsky et al. [3] (see also remarks to C. nursei). On the basis of their drawings, both species have a copulatory tube with a short fingerlike basal process rising from a small base at an angle of about 40°. The new species differs from J. cf. longipenis by having a comparatively shorter (31-39 vs 50 µm) and J-shaped (vs coiled) shaft of the tube. However, these morphological differences are relatively small, and in absence of comparative material, new specimens of J. cf. longipenis will be needed to make a redescription and to obtain molecular data. Then a decision on synonymy of C. polex n. sp. with J. cf. longipenis may be made.

Molecular characterisation
The combined 18S-ITS1 sequence of C. auritum n. sp. was 973 bp long, of which 483 bp corresponded to the 18S rDNA and 490 bp corresponded to the ITS1 region. The sequence of partial 28S rDNA was 772 bp long. Five specimens were sequenced from South Africa and Zimbabwe and no intraspecies variability for the 28S and 18S-ITS1 regions was observed.

Differential diagnosis
Characidotrema auritum n. sp. is a sister species to C. vespertilio n. sp., according to the similarities of both morphology and DNA sequences. These two new species are probably close to C. elongata described from Brycinus jacksonii (syn. Alestes jacksoni) in Uganda by Paperna & Thurston [5] and later reported on B. leuciscus (syn. Alestes leuciscus) in Ghana [7] (see also [3]). All three species possess a copulatory tube with a base having a proximal (terminal) flange and a robust basal process to which an accessory piece is attached. However, C. auritum n. sp. clearly differs from these two species by possessing a coiled (vs sigmoid in C. vespertilio and C. elongata) and markedly longer copulatory tube (109-126 vs 45-51 µm in C. vespertilio and 30-40 µm in C. elongata). In addition, the vagina of C. auritum n. sp. is markedly longer than that of C. vespertilio n. sp. (29-51 vs 12-19 µm). (Peters, 1852).

Molecular characterization
The combined 18S-ITS1 sequence of C. vespertilio n. sp. was 972 bp long, of which 483 bp corresponded to the 18S rDNA and 489 bp corresponded to the ITS1 region. The sequence of partial 28S rDNA was 761 bp long. A total of 10 specimens were sequenced from D.R Congo and Cameroon and no intraspecies genetic variability was observed in any marker.

Differential diagnosis
This new species is similar to C. auritum n. sp. and C. elongata Paperna & Thurston, 1968 in having an accessory piece attached to a robust basal process of the copulatory tube. Features distinguishing C. vespertilio n. sp. from C. auritum n. sp. are given in the differential diagnosis for the latter species. Characidotrema vespertilio n. sp. differs from C. elongata by possessing a more robust accessory piece with the distal end formed as three

Genetic relationships among Characidotrema spp.
No intraspecific sequence variation was observed between specimens of the same species of Characidotrema. Values of pairwise genetic comparisons are shown in Table 5. The uncorrected p-distances among six Characidotrema spp. ranged between 0-0.019 for 18S, 0.024-0.218 for ITS1 and 0.004-0.059 for 28S, while the average interspecific p-distances were 0.011, 0.174 and 0.046 for 18S, ITS1 and 28S, respectively. The smallest interspecific distances were observed between C. auritum n. sp. and C. vespertilio n. sp. for each marker, both infecting B. imberi. Characidotrema spinivaginus was revealed as the most genetically distant species to C. auritum n. sp. or C. vespertilio n. sp. (0.019) for 18S, to C. brevipenis (0.218) for ITS1, and to C. pollex n. sp. (0.059) for 28S.

Discussion
Species of Characidotrema are known only from the gills of African tetras (Alestidae), although the majority of them occur on species of Brycinus, a speciose Pan-African genus including commercially important species. Of the 13 valid species in Characidotrema (including the new species described herein), ten have been recorded only from Brycinus spp., two species from hosts representing both Brycinus and Alestes, and only one species  Table 1). All members of Characidotrema including the new species described herein possess a relatively uniform configuration and morphology of the haptoral sclerites [3]. The dorsal anchors, characterized by a submedially swollen shaft, are morphologically (in shape) indistinguishable among species of Characidotrema; therefore, they are not suitable for species determination. Additionally, the shape of the highly modified ventral anchors is also very similar in all species of the genus. However, the shape of their roots slightly varies among specimens of individual species of Characidotrema and even between the ventral anchors of one and the same specimen (see Figs. 2,3,4,5,6,7,8), although the length of the roots is relatively constant (i.e. species-specific). The diagonally truncate or scoop-shaped point of the ventral anchor represents an unusual feature that is unique among African dactylogyrids [3]. Our observations on the ventral anchors of Characidotrema spp. by means of phase contrast microscopy revealed the presence of a sheath-like structure associated with the shaft and point of the ventral anchors (Fig. 10). Molnár & Mossalam [10] were the only authors who described and illustrated such a structure before the present study. However, it should be mentioned that, since these structures are not (or only weakly) sclerotized, they were not always visible in all specimens of the species of Characidotrema in the present collection. A similar feature has previously been reported for species of Triacanthinella Bychowsky & Nagibina, 1968 (Dactylogyridae) from triacanthid teleosts of Peninsular Malaysia [29]. However, in Triacanthinella spp., a structure associated with the lower portion of the anchors is formed as a sheath-like sclerite closely enveloping the shaft and point of the anchors, whereas in the present species this structure appears to be more independent, with a distal part resembling a caudal fin. Unfortunately, it is impossible, without studying the haptoral armament using more sophisticated techniques (e.g. laser scanning confocal fluorescence microscopy), to suggest the function of these structures. The ventral bar in Characidotrema spp. appears to be flexible, resulting in variability in its shape in fixed specimens. The hooks basically exhibit the typical "ancyrocephaline" distribution described by Mizelle [14] for dactylogyrids. However, in all specimens studied here, pairs I and V are placed in close proximity (at the level) of the ventral bar (pair I lies more anteriorly, i.e. between the arms of the bar), whereas in other dactylogyrids these pairs are usually situated further anteriorly (pair I) and posteriorly (pair V) from the ventral bar, i.e. further apart from each other. The same arrangement of hooks is also observed in the original drawings of the haptoral configuration given for the following previously described species of Characidotrema: C. brevipenis [3], C. regia [11], C. spiropenis Birgi, 1988 [11] and C. undifera [3]. It seems highly probable, although not mentioned or figured in their original descriptions, that this hook distribution occurs in all previously described species of Characidotrema.
As indicated above, the most apparent character distinguishing species in the genus is the morphology of the MCO and vagina. In the case of the MCO, two morphological types dividing Characidotrema spp. in the present collection into morphological groups may be defined.
The first group includes five species (C. brevipenis, C. nursei, C. pollex n. sp., C. spinivaginus and C. zelotes) with an MCO composed of a copulatory tube with a finger-like basal process rising (at different angles) from the distal part of the base, and an accessory piece mostly not articulated to the base (if articulated, then by a poorly sclerotized part as in C. brevipenis; Figs. 2, 11). The second group, including two new species (C. auritum n. sp. and C. vespertilio n. sp.), is characterized by a copulatory tube with a crest-shaped terminal flange and a robust basal process to which an accessory piece is articulated.
The division into two morphological groups was also supported by the results of phylogenetic analyses based on 28S rDNA sequences (Fig. 11). The ML phylogenetic reconstruction showed that Characidotrema auritum n. sp. from B. imberi clusters together with C. vespertilio n. sp. from the same host. This clade was well separated from the clade of Characidotrema spp. found on Brycinus nurse. Thus, the basic structure of the MCO (along with host preferences) appears to be the best morphological tool for resolving relationships among species of Characidotrema. This result supports the MCO type concept [30] suggesting that the morphology of the MCO carries much information of systematic/phylogenetic importance and that species with a particular MCO type often parasitize a particular major host taxon [13,31,32]. However, C. ruahae (Paperna, 1979) from B. imberi is morphologically more similar to C. brevipenis and C. nursei from B. nurse [3] than to the two new species from B. imberi. Although we cannot verify the accuracy of the host identification, it could be interesting to resolve the phylogenetic relationships between these Characidotrema spp. possessing similar morphological characters of the MCO but parasitizing different host species. Conversely, in terms of the haptor, all thirteen species of Characidotrema considered valid exhibit a relatively uniform morphology of the haptoral sclerites. Considering that similar attachment organs will lead to the choice of the same microhabitats on the host, as postulated by Rohde [33], then it can be assumed that all currently known species of Characidotrema would occupy the same niche on the gills of their hosts. In addition, as indicated in Table 1, communities of Characidotrema spp. on respective hosts can display a relatively high species richness, which ranges from one (e.g. on B. kingsleyae) to six (reported on B. nurse) species per host species. Our study revealed that up to three species, C. nursei, C. pollex n. sp. and C. spinivaginus or C. brevipenis, C. spinivaginus and C. zelotes, may simultaneously parasitize B. nurse. The coexistence of several species of Characidotrema on the same host, together with the narrow interspecific variability in the haptoral sclerites and the comparatively wider interspecific variability in the MCOs, supports the idea that parasite species occupying the same microhabitats are reproductively isolated by morphological differentiation of their copulatory organs that probably avoids the hybridization [34][35][36]. Our results also suggest that the morphology of the MCO is not the only factor affecting reproductive isolation among Characidotrema spp. The morphological differences of the vagina, associated with the degree of sclerotization, probably also play an important role in the reproductive isolation of these species. On the other hand, due to the high variability of this organ even among closely related species, it appears that the morphology of the vagina can make only a minimal contribution with respect to clarifying relationships among species of this genus.
Phylogenetic analyses based on 28S rDNA sequences supported the monophyly of Characidotrema spp., and indicated the closer relationship of this genus to Arrows indicate one of the principal characters dividing Characidotrema spp. into two morphological groups monogeneans parasitizing African cyprinids (Dactylogyrus spp.) and cichlids (species of Cichlidogyrus, Scutogyrus and Onchobdella) than to those parasitizing catfishes (species of Quadriacanthus, Schilbetrema and Synodontella). Our results indicate that Characidotrema spp. from the same host are more closely related and suggest possible intra-host speciation.
The majority of Characidotrema spp. are restricted to species of Brycinus, which suggests a long association between these parasites and the host genus. Recent molecular studies on Brycinus species have repeatedly indicated that this taxon is polyphyletic [37][38][39][40] and support the monophyly of the "macrolepidotus" species group of Paugy [41] comprising eight large-bodied species (including B. macrolepidotus Valenciennes, the type-species of the genus) mainly distributed in Central Africa. Phylogenetic analyses presented by Arroyave & Stiassny [40] placed B. imberi and B. nurse together in a clade of Brycinus (sensu lato) recovered as sister to a group of species belonging to Bryconaethiops Günther. Parasites, especially highly hostspecific monogeneans, are usually considered to be good biological markers of their host diversity and evolution [42,43]. However, current knowledge on the diversity and distribution of Characidotrema spp. is insufficient for either supporting or rejecting the above hypotheses on Brycinus phylogeny. A total of 13 species of Characidotrema and 19 species of Annulotrema are known from five and seven species of Brycinus, respectively. Considering that 36 species of Brycinus are currently recognized as valid (Fish-Base [12]), the diversity of monogeneans parasitizing these fishes appears to be poorly understood.

Conclusions
The present study provides first insights into the molecular phylogeny of monogeneans parasitizing African tetras. Morphological and molecular data suggest Characidotrema to have a monophyletic nature and indicate the closer relationship of this genus to monogeneans parasitizing African cyprinids and cichlids than to those parasitizing catfishes. The overall agreement between the morphological diversification of the MCOs and the molecular tree generated in this study indicates that significant phylogenetic signal for clarifying relationships among species of Characidotrema is associated with certain characteristics of the MCO. The present study suggests that intra-host speciation is probably one of the important ecological processes involved in species diversification in Characidotrema. Nevertheless, further morphological and molecular samplings of monogeneans of both genera, Characidotrema and Annulotrema, are needed to elucidate questions related to the ecological similarity, coexistence, and phylogeny of these parasites. To identify potential co-speciation events, co-phylogenetic analyses of these monogeneans and their alestid hosts are required.