Considerations on the taxonomy and morphology of Microcotyle spp.: redescription of M. erythrini van Beneden & Hesse, 1863 (sensu stricto) (Monogenea: Microcotylidae) and the description of a new species from Dentex dentex (L.) (Teleostei: Sparidae)

Background Microcotyle erythrini van Beneden & Hesse, 1863 (Platyhelminthes: Monogenea) and other closely related species of the genus are often considered as cryptic. Records in hosts other than the type-host with no species confirmation by molecular analyses have contributed to this situation. Methods Gill parasites of five sparid fishes, Boops boops (L.), Pagellus erythrinus (L.), P. acarne (Risso), Dentex dentex (L.) and Pagrus pagrus (L.), from the Western Mediterranean off Spain were collected. Specimens of Microcotyle spp. were characterised both molecularly and morphologically. Partial fragments (domains D1-D3) of the 28S rRNA gene and the cytochrome c oxidase subunit 1 (cox1) gene were amplified and used for molecular identification and phylogenetic reconstruction. Principal components analysis was used to look for patterns of morphological separation. Results Molecular analyses confirmed the identity of three species: M. erythrini ex P. erythrinus and Pa. pagrus; M. isyebi Bouguerche, Gey, Justine & Tazerouti, 2019 ex B. boops; and a species new to science described herein, M. whittingtoni n. sp. ex D. dentex. The specific morphological traits and confirmed hosts (P. erythrinus and Pa. pagrus) are delimited here in order to avoid misidentifications of M. erythrini (sensu stricto). Microcotyle erythrini (s.s.) is mostly differentiated by the shape of its haptor, which is also longer than in the other congeners. New morphological and molecular data are provided for M. isyebi from the Spanish Mediterranean enlarging the data on its geographical range. Microcotyle whittingtoni n. sp. is described from D. dentex and distinguished from the remaining currently recognised species of the genus by the number and robustness of the clamps. Conclusions New diagnostic morphological traits useful to differentiate Microcotyle spp. are suggested: (i) haptor dimensions including lobes; (ii) the thickness of the clamps; (iii) the size and shape of spines of the genital atrium; (iv) the extension of the posterior extremities of vitelline fields; and (v) the shape of egg filaments. The use of new morphological approaches may allow considering these species of Microcotyle as being pseudocryptic. The use of representative undamaged specimens that have been genetically confirmed as conspecific is considered crucial to avoid abnormally wide morphological ranges that prevent species differentiation.


Background
Microcotyle erythrini van Beneden & Hesse, 1863 (Monogenea: Microcotylidae) was originally described from Pagellus erythrinus (L.) (Teleostei: Sparidae) off the coast of Brest (France, North-East Atlantic) and to date it has been listed and considered a valid species [1][2][3]. Like many of the earliest descriptions of species of Microcotyle, M. erythrini was described briefly, and only differentiated by the number of clamps and testes, and the traits of the genital atrium [4]. Since the original description, many authors have recorded and described new specimens identified as M. erythrini in different sparid species, mostly in the Mediterranean Sea (see Table 1 in Bouguerche et al. [3], for details on the records of M. erythrini). These publications sometimes offered morphological ranges based on a combination of measurements of specimens from different host species (e.g. [5,6]). Along this process, the morphological ranges of M. erythrini have been enlarged abnormally, which has made it difficult to define a clear and distinguishing morphology. Recently, with the help of molecular tools (cox1 partial fragment), M. erythrini has been split into two species, each in a different sparid host off the Algerian coast: M. erythrini ex P. erythrinus and M. isyebi Bouguerche, Gey, Justine & Tazerouti, 2019 ex Boops boops (L.) [3]. These authors also included the most recent morphometric information on M. erythrini from the type-host P. erythrinus. Bouguerche et al. [2,3] suggested that morphological and molecular characterization of M. erythrini-like specimens infecting different sparid hosts would reveal higher parasite diversity.
The aim of the present study is a revision of the taxonomy of Microcotyle spp. in sparids from the Western Mediterranean off Spain. The specific objectives of the study are: (i) to describe a new species of Microcotyle parasitic in Dentex dentex (L.); (ii) to redescribe M. erythrini with the support of molecular evidence, define the actual morphological boundaries of the species and indicate the valid historical records; and (iii) to provide new morphological and molecular data useful for the taxonomy of Microcotyle spp. New morphological approaches and classification tools for species discrimination are proposed for these monogeneans which are notoriously difficult to differentiate.

Sample collection
A total of 150 fishes of four sparid species were examined for microcotylid infections: 40 bogues (Boops boops), 40 common pandoras (Pagellus erythrinus), 40 common dentexes (Dentex dentex) and 30 red porgies (Pagrus pagrus). Additionally, 40 axillary seabreams (P. acarne (Risso)) were also examined. Fishes were caught by commercial bottom trawling vessels during July of 2012 and 2013, off Guardamar del Segura, Alicante, Spain (38°05′N, 0°39′W; Western Mediterranean Sea, FAO fishing subarea 37.1). Fishes were transported on ice to the laboratory, where they were weighed, measured (weight provided in g and standard length in cm, expressed as the range with the mean and standard deviation (SD) in parentheses; only provided for infected hosts in the taxonomic summary) and then dissected for gill examination. Each pair of gills was dissected and inspected for parasites under a stereomicroscope. All parasites were collected and washed in 0.9% saline solution. For Microcotyle spp. specimens, two different protocols were used. Adult and completely mature specimens in optimal conditions (not broken, contracted, stretched, wrinkled or folded) were selected for morphological analyses; these were fixed in 4% formaldehyde solution and preserved for four days, then the specimens were transferred into 70% ethanol. For molecular analyses, fresh specimens were selected; the testes and clamps were counted and photographed and then the specimens were divided into three pieces, storing the anterior and posterior parts as molecular vouchers. The middle pieces were fixed and preserved in molecular-grade ethanol. Prevalence, expressed as a percentage (infected fish and total number of analysed fish in parentheses), and mean intensity, expressed as the mean with standard deviation, in each host, were calculated according to Bush et al. [7].

Alignment and data analyses
Contiguous sequences were assembled in MEGA v.6 [13] and alignments with currently available sequences for Microcotyle spp. in the GenBank database (retrieved on 25th July 2019) were constructed using MAFFT v.7 [14] under default gap parameters on EMBL-EBL bioinformatics web platform (http://www.ebi.ac.uk/Tools /msa/mafft ). The outgroup choice was based on previous phylogenies of the group [15][16][17]. The cox1 alignment (381 nt) comprised a total of 12 newly generated sequences and 20 sequences for 10 species available on GenBank. Bivagina pagrosomi ex Sparus aurata (L.) (GenBank: Z83003) was used as the outgroup. The 28S alignment (823 nt) comprised 4 newly generated sequences and 10 sequences available on GenBank. Bivagina pagrosomi ex S. aurata (GenBank: Z83002) was used as the outgroup. Distance matrices (using the uncorrected p-distance model) were calculated in MEGA v. 6. Neighbour-joining analyses based on Kimura 2-parameter distances were also performed in MEGA v.6 with nodal support estimated using 1000 bootstrap resamplings. Model-based Bayesian inference (BI) and maximum likelihood (ML) analyses were carried out using MrBayes v.3.2.6 on XSEDE at the CIPRES Science Gateway v. 3.3 [18] and PhyML v.3.0 [19] as an online execution on the ATGC bioinformatics platform (http://www. atgc-montp elier .fr/) with a non-parametric bootstrap validation of 1000 pseudoreplicates, respectively. The MCMC chains were run for 10,000,000 generations with trees sampled every 1000 generation. Posterior probability and mean marginal likelihood values were calculated. The first 25% of the sampled trees were discarded as ‛burn-inʼ. Prior to analyses, jModelTest v.2.1.4 [20,21] was used to select the best-fitting models of nucleotide substitution under the Akaikeʼs information criterion. These were the general time-reversible model with gamma distributed among-site rate variation and estimates of invariant sites (GTR+Г+I) for the cox1 dataset and the Hasegawa-Kishino-Yano model (HKY) for the 28S dataset. Consensus topologies and nodal supports were visualized in FigTree v.1.4.3 [22], posterior probabilities (pp) and bootstrap support (bs) values are summarised on the BI trees (as pp/bs).

Morphological analyses
Parasites selected for morphological analyses were stained with iron acetocarmine, dehydrated through an ethanol series, cleared in dimethyl phthalate and prepared as permanent mounts in Canada balsam. After mounting, there was a second selection of specimens suitable for morphological studies, i.e. only specimens in optimal condition (not broken, contracted, stretched, wrinkled or folded). Parasites were examined using a light microscope Nikon Optiphot-2 (Nikon Instruments, Tokyo, Japan) with differential interference contrast at magnifications of 400-1000×. A total of 86 specimens of Microcotyle spp. were selected and drawn (n = 22 ex B. boops; n = 21 ex D. dentex; n = 23 ex P. erythrinus; n = 20 ex Pa. pagrus). Drawings were made with the aid of a drawing tube attached to a light microscope Nikon Optiphot-2. Measurements were taken from digitalized illustrations using ImageJ v.1.48 software [23] and expressed in micrometers as the range followed by the mean in parentheses unless otherwise stated. When characters were visible, a total of 52 morphometric measurements were taken from each specimen. Clamp thickness was estimated as both the maximum width of the distal end of the antero-lateral sclerite ('c' , see Fig. 1a) and its relation to the clamp length. The type-specimens were deposited in the Collection of the Natural History Museum (NHMUK), London, UK.
To look for patterns of separation between Microcotyle spp. specimens from different host species, a principal components analysis (PCA) was applied to a dataset of 86 specimens using morphometrical variables associated with body shape. Prior to the analysis, the data were divided by total body length to account for the effect of body size while visualising possible morphometric differences between species. The specimens were identified as M. erythrini (n = 23 ex P. erythrinus; n = 20 ex Pa. pagrus), M. isyebi (n = 22 ex B. boops) and Microcotyle whittingtoni n. sp. ex D. dentex (n = 21).

Molecular identification
A total of 12 cox1 and four 28S rDNA sequences were generated for the newly collected specimens of Microcotyle spp. from the four fish species from the Western Mediterranean off Spain. Partial cox1 (434 nt) sequences were generated for a total of 12 isolates, i.e. 4 M. isyebi ex B. boops, 6 M. erythrini (4 ex Pa. pagrus and 2 ex P. erythrinus) and 2 M. whittingtoni n. sp. ex D. dentex. Partial 28S rDNA sequences (1238-1527 nt) were generated for a representative subset of the specimens used for cox1 sequence generation; single sequences per species were used for the reconstruction of the 28S rDNA phylogeny. The newly generated sequences for the isolates recovered in the present study were analysed in two separate datasets together with all currently available sequences in the GenBank database for Microcotyle spp. (see Table 1 for details on the ingroup taxa used in the analyses). Posterior probabilities (pp) and bootstrap support (bs) values are summarised on the BI trees (as pp/bs).
The newly generated cox1 sequences were analysed together with 19 published sequences for Microcotyle spp. (Table 1). Phylogenetic analysis revealed that the newly sequenced isolates belonged to 3 species: M. erythrini ex P. erythrinus and Pa. pagrus; M. isyebi ex B. boops; and M. whittingtoni n. sp. ex D. dentex. The tree from the BI analysis is provided in Fig. 2 together with the statistical support from the ML analysis. The four isolates recovered from B. boops clustered together with an isolate of M. isyebi from the same host species reported from the southern coast of the Western Mediterranean off Algeria [3]. The sequences for the isolates recovered from Pa. pagrus and P. erythrinus clustered together with the published sequences for M. erythrini ex P. erythrinus from the Western Mediterranean off France [15]. The two sequences for M. whittingtoni n. sp. ex D. dentex clustered together in a basal clade to the remaining representatives of Microcotyle spp. All of the above clades were strongly supported in both BI and ML analyses. Overall, the cox1 phylogeny ( Fig. 2)  The novel 28S sequences for M. erythrini recovered from the two fish host species differed by a single base. Schematic drawings for measurements of microcotylid clamps and haptors. a Clamp measurements: 'a' , 'c' , 'd' and 'e' , microcotylid sclerites according to Llewellyn [47]. b-e Body outlines, number of clamps and measurements in mounted microcotylids: unmounted specimen in 3D view (b); mounted specimens in 2D view (c-e). Haptor anterior lobe lays on the body in c and e, haptor obliquely mounted in d (anterior lobe not laying on the body); e represents damaged specimen with missing pieces (arrow) and clamps (arrowheads). Abbreviations: ahl, anterior haptor lobe length; bl, body length; bl-h, body length without haptor; cl, clamp length; cw, clamp width; csw, 'c' sclerite width; hl, haptor length; phl, posterior haptor lobe length  [52-81% (66%) (n = 19)]. Haptor armed with two rows of sessile clamps, 90-124 in number for specimens from both host species, in two lateral frills, joining at anterior and posterior extremities of haptor; with slightly smaller clamps at anterior and posterior margins of haptor. Clamps of "microcotylid" type, slender, "c" sclerite maximum width 3 ± 1 for specimens from both host species and 0.084 ± 0.015 (n = 25) [0.068 ± 0.019 (n = 25)] corrected by clamp length; with trident-shaped accessory sclerite ('e' , see Fig. 1a) formed by thick central bar reaching to distal tips of antero-lateral sclerites 'c' and two thin short sclerites directly branched from basis of 'e' . Mouth subventral, within conical vestibule with pair of septate buccal suckers. Pharynx subspherical; oesophagus short; intestinal bifurcation posterior to genital atrium, sometimes at level of atrium. Caeca extend into haptor or peduncle, with inner and intricate external lateral ramifications.

Remarks
Microcotyle erythrini was described by van Beneden & Hesse [4] and mostly characterized by its specific host, P. erythrinus, as authors provided limited morphological information (mostly at the generic level) and with no supporting drawing. Parona & Perugia [5] redescribed this species; however the description is unreliable as these authors provided pooled morphological information from material ex B. boops, host of M. isyebi (see [3] and present study) and ex P. acarne, a host not confirmed for M. erythrini. Morphological data with pooled information form specimens collected from more than one host species or parasites collected in fish species different from the type-host or other confirmed hosts should not be considered as suitable (see also Additional file 3: Table S3). Several new geographical records of M. erythrini ex P. erythrinus, exclusively, have been published by other authors since 1863 (see Table 1 in [3]). Among these records, only Radujković & Euzet [31] and Bouguerche et al. [3] provided morphological and morphometric data for specimens off Montenegro and Séte, respectively (see also Additional file 3: Table S3). Here, we provide metrical data (  [3] and the present study [from here onwards M. erythrini (sensu stricto)], the diagnostic characters of M. erythrini (s.s.) agree and measurements mostly overlap but wide ranges for some features are still observed (see also Additional file 3: Table S3), which hampers the differentiation from other congeneric species. Paying attention to the characters traditionally used in the taxonomy of Microcotyle, the number of clamps (82-132) and the number of testes (9-24) of M. erythrini (s.s.), combining the information from all descriptions in confirmed hosts (see Table 2 and Additional file 3: Table S3), resemble or overlap with those of several species reported in the Mediterranean (M. donavini van Beneden & Hesse, 1863 and M. pomatomi Goto, 1899) and in other sparid hosts (M. isyebi and M. visa). Regarding the traits more recently used to differentiate the species of Microcotyle, such as the genital atrium armature and combining the information from all descriptions in confirmed hosts (see Table 2 and Additional file 3: Table S3), M. erythrini (s.s.) resembles other species with large number of   [24,31]; numbers estimated from the drawing for M. pomatomi) (see Additional file 3: Table S3).
According to the combination of the characters listed above, M. isyebi, M. pomatomi and M. visa appear most similar morphologically to M. erythrini (s.s.). Microcotyle pomatomi, the only species described and reported from a non-sparid host (Pomatomus saltatrix (L.); Pomatomidae), is difficult to differentiate due to the numerous circumglobal records and descriptions which have increased abnormally the ranges for the metrical data of this species. Moreover, the only Mediterranean description of M. pomatomi (off Turkey, Sezen & Price, 1967 in [32]) is particularly similar to M. erythrini (s.s.). Detailed morphological and molecular studies are needed to differentiate the two species. The other two species, both sparid parasites, were described as hardly morphologically distinguishable from M. erythrini. Microcotyle visa was differentiated from M. erythrini by the smaller clamp size, larger pharynx and greater number of testes; however, these differences are not completely sufficient to differentiate species as all they overlap (even with those of M. erythrini (s.s.)) [2]. No diagnostic morphological differences were provided by Bouguerche et al.
Mouth subventral, within funnel-shaped vestibule with pair of septate buccal suckers. Pharynx subspherical; oesophagus short, connected to intestinal bifurcation at posterior margin of genital atrium, or more posterior. Caeca extend up to haptor peduncle, with inner and external lateral ramifications (external more profuse).

Remarks
Both morphological and molecular data reported in the present paper agree with the original description of M. isyebi based on material from B. boops off Algeria [3] and from the Spanish Mediterranean [33]. Parona & Perugia [5] and Akmirza [6] also provided morphological data from specimens identified as M. erythrini ex B. boops but these were not considered as species diagnostic in the present study as they represent pooled information for parasites ex B. boops and another host, P. acarne; Microcotyle spp. in sparids are highly host species-specific (see [2,3] and the present study). In the present study, no specimens of Microcotyle spp. were found in P. acarne.
Some comments on the original diagnosis of the species can be added in light of the data from the description of López-Román & Guevara Pozo [33] and the present study. The range for the number of clamps seems too wide in the original description of M. isyebi based on material collected off Algeria (54-102) compared with that reported by López-Román & Guevara Pozo e Germarium. f Egg. g Detail of abopercular egg filament. Abbreviations: co, copulatory organ; mc, main chamber of the genital atrium; p, small posterior chambers ("pockets" sensu Mamaev [44]). Scale-bars: a, 500 µm; b, d-f, 100 µm; c 50 µm [33] (90-100) and the present study (80-110) based on material collected off Spain. These numbers should be re-examined, especially because this trait is particularly differential among the species of Microcotyle and, as previously reported, it could help differentiate M. isyebi from other similar species such as Microcotyle sp. ex H. dactylopterus [24] and M. whittingtoni n. sp. (see [3] and Remarks to the new species below). Particular attention must be paid to the lower range for clamp number as a small number of clamps is often related to young or damaged specimens. The number of spines in the main chamber of the genital atrium of M. isyebi is also clearly lower in the original description than in the present material (136-230 vs 253-356) (see [3] and Table 2 Table S4). The presence of posterior small chambers of the genital atrium ("pockets") was also reported as diagnostic in the original description of M. isyebi; however, this feature requires a further comment. According to Bouguerche et al. [3], "pockets" are absent in M. archosargi, M. lichiae Ariola, 1899 and M. pomatomi; however, this difference seems to be valid only for M. lichiae, as these small chambers exist in M. archosargi and M. pomatomi according to the drawings in [25] and [32], respectively.
In the original description of the species, M. isyebi was differentiated from M. pomatomi and from Microcotyle sp. ex H. dactylopterus [24] by traits with overlapping ranges (the number of clamps and spines of the genital atrium for Microcotyle sp. ex H. dactylopterus) or almost overlapping ranges (the number of clamps and testes for M. pomatomi). Microcotyle pomatomi and Microcotyle sp. ex H. dactylopterus [24] require further taxonomic research; M. pomatomi has numerous descriptions and synonyms worldwide which have expanded extremely the ranges for morphological features (see [32]; also the only Mediterranean record by Sezen & Price (1967) in [32]), and the morphology Microcotyle sp. ex H. dactylopterus has been only briefly described [24,31].
Bouguerche et al. [3] reported that M. isyebi is almost indistinguishable from M. erythrini. As mentioned above, examination of mature, entire, uncontracted, unstretched and unfolded specimens of this species would be helpful to define or shorten some of the descriptive morphological ranges. Other morphological traits suggested in the present study reveal additional differences.
Mouth subterminal, ventral, with 2 septate buccal suckers within funnel-shaped vestibule; oesophagus short; intestinal bifurcation at level of posterior margin of genital atrium or just proterior. Caeca with inner and profuse external lateral ramifications extending into haptor.
The combination of characters for M. whittingtoni n. sp. previously mentioned in the remarks is also present in Fig. 6 Microcotyle whittingtoni n. sp. ex Dentex dentex (L.) from off Guardamar del Segura, Spain. Holotype. a Whole mount. b Anterior end. c Clamp. d Genital atrium, including copulatory organ. e Germarium. f Egg. g Detail of abopercular egg filament. Abbreviations: co, copulatory organ; mc, main chamber of the genital atrium; p, small posterior chambers ("pockets" sensu Mamaev [44]). Scale-bars: a, 500 µm; b, d-f, 100 µm; c, [24]. Ariola [35] differentiated M. lichiae from other Microcotyle spp. predominantly by its large body size, asymmetrical haptor and concentric arrangement of the spines in the genital atrium. Additionally, M. whittingtoni n. sp. differs from M. lichiae by the shape of the haptor and the genital atrium; in fact Ariola [35] described a genital atrium with five rings of concentric spines in M. lichiae, unique among Microcotyle spp. Moreover, M. lichiae parasitizes a non-sparid host (Carangidae). More specimens of M. lichiae must be examined in order to determine the taxonomic status of this species. Regarding the outstandingly greater body length of M. lichiae (8000 vs 2719-4569 µm), this must be considered with caution as the size of polyopisthocotylean monogeneans is strongly related to host size, and thus not a reliable character in the taxonomy of polyopisthocotyleans [32]. Regarding M. isyebi and Microcotyle sp. ex H. dactylopterus [24], it is difficult to depict differential features for M. whittingtoni n. sp. other than parasitism in different hosts. The only clear differential morphological trait of M. whittingtoni n. sp. is the possession of more robust clamps, distinctly different from those in M. isyebi (ratio "c" sclerite maximum width/total clamp length, 0.100-0.146 vs 0.027-0.88). Another difference can be found in the posterior extremities of vitelline fields, always asymmetrical in M. whittingtoni n. sp., while M. isyebi includes some specimens with symmetrical fileds (21% according to the present study, see species description above). The data on Microcotyle sp. ex H. dactylopterus is limited, and some characters, such as clamp or egg morphology are not reported, therefore, until more specimens are analysed, the only evidence to differentiate these species is parasitism in different host species and molecular delineation ( [24]; present study).
The specimens recorded by González González [36] in D. dentex from the Balearic Islands (identified as M. erythrini) belong to M. whittingtoni n. sp. due to the congruent morphology, host, and geographical distribution. The description by González González [36] agrees well with the description of the new species, except for the greater number of clamps (110-120 vs 60-78). However, the number of clamps in the specimen of the drawing and photograph in González González [36] has 60 clamps (figures 6 and 8 in [36]) in agreement with the description of the new species.

Multivariate morphometric analysis
The PCA using seven morphometric variables associated with body shape produced a plot of the 86 specimens (one extreme outlier was removed prior to analyses) in the first plane of the PCA showing the morphological variability between the species of Microcotyle from the Spanish Western Mediterranean (Fig. 7). The first two axes of the PCA explained 73.81% of the variation in the dataset. The first axis explained 55.32% of the variation and showed a separation between M. erythrini (s.s.) and M. isyebi, while M. whittingtoni n. sp. overlapped with the other two species. The specimens of M. erythrini ex P. erythrinus and Pa. pagrus showed a wider variation, whereas, for M. whittingtoni n. sp. ex D. dentex and M. isyebi ex B. boops the variation was lower. The first axis was positively correlated with the maximum body width (0.844), body length without the haptor (0.736), body width at level of the genital atrium (0.726) and body width at testis area (0.777), and negatively correlated with the length of the anterior haptor lobe (−0.856) and haptor length (−0.803). The second axis which was negatively correlated with body width at the level of the buccal suckers (−0.838) showed intraspecific separation between the specimens of M. erythrini ex P. erythrinus and ex Pa. pagrus.

Discussion
No type-species was selected for the genus Microcotyle in the original definition by van Beneden & Hesse [4], which included the descriptions of two species, M. donavini and M. erythrini (also M. canthari and M. labracis, but these species currently belong to the genera Neobivagina and Serranicotyle, respectively). Later, Sproston [37] selected M. donavini as the type-species for the genus but at the time of the erection of the genus, these two species were the first morphological references. First descriptions of new Microcotyle species were based on vague morphological differences (mostly the number of clamps and testes [5,38,39]). Many more species have been described since then worldwide and several genera of the subfamily Microcotylinae have been erected, and M. erythrini has continued being considered valid [2,3]. Recently, Bouguerche et al. [2] and Bouguerche et al. [3] provided molecular evidence that despite the validity of this species, several Microcotyle spp. from different host species have been wrongly identified as M. erythrini because of their morphological homogeneity. These authors referred to a M. erythrini complex of cryptic species and suggested that a molecular re-evaluation may reveal additional parasite diversity [3]. Caution must be taken in order to select representative specimens in perfect conditions of maturity, completeness and constitution. Whittington [40] stated that to separate monogeneans of a species complex with high levels of diversity "it is vital to ensure that there is a useful trail of high quality parasite material" for taxonomic studies, also stressing the importance of supporting the results with molecular genetic analyses. The present study shows that morphological differences between M. erythrini (s.s.) and similar species can be found: a new species of Microcotyle is described in D. dentex, together with the redescription of M. erythrini (s.s.) (including a new host record, Pa. pagrus) and a new geographical record of M. isyebi with additional morphological information, all supported by molecular evidence.
Molecular analyses of the cox1 gene showed clear differences between Microcotyle spp. distinctly separating the three species described here. Previous studies have suggested levels of intraspecific variation lower than 5% for species of mazocraeids and microcotylids (up to 5.6% and up to 4.5%, respectively, Yan et al. [41]; Mladineo et al. [42]). Based on cox1 sequences, M. whittingtoni n. sp. appears markedly distinct, since the genetic distance from the remaining congeners was higher than 10.8%; specifically, the two isolates ex D. dentex differed from M. erythrini (s.s.) by 10.8-13.5% and from M. isyebi by 14.4-15.2%. The available 28S rDNA sequences for Microcotyle spp. are scarce as this region is not commonly used as a marker for interspecific differences.
We have delimited the valid morphological ranges of M. erythrini (s.s.) and similar species in sparids; in addition, we suggest the use of new diagnostic characters and morphological tools for assessment of multivariate patterns (e.g. PCA). One of the issues in defining the differential traits of similar species of microcotylids is to avoid abnormally wide morphological ranges by selecting only representative specimens: (i) not including specimens potentially belonging to other species (e.g. morphologically similar parasites from other host species not confirmed by molecular analysis); (ii) selecting morphologically optimal specimens (mature, unbroken, uncontracted, unstretched, not wrinkled and unfolded); and (iii) characterizing these specimens accurately to ensure that the diagnostic species-specific characters are properly described. More than 150 years after the original description of M. erythrini, numerous descriptions of this species from P. erythrinus and other hosts have provided extremely wide ranges of morphological information for this species, making it almost impossible to find differentiating features. By defining M. erythrini (sensu stricto) here, we aimed to characterize the species by much narrower morphological ranges only considering valid genetically tested specimens from confirmed hosts. Regarding the optimal specimen selection, only completely mature adults should be representative for standardized taxonomic descriptions. Worms with fully developed both male and female reproductive systems must be selected, as testes in young adults are early functional and vas deferens is full of sperm while no developed oocytes exist in the germarium. Also, the preservation and completeness of the specimens must be ensured.
Knowledge of the three-dimensional structure of the monogeneans in fresh preparations is crucial to understand the morphology of the specimens mounted in Canada balsam as under the coverslide they are represented in a two-dimensional view; knowledge of the natural shapes allows detecting possible folds and missing parts. For example, the haptor of Microcotyle spp. has a dorsal lobe and a ventral lobe (sometimes notched anteriorly), both with clamps; when the specimens are mounted (usually in ventral view), dorsal and ventral projections are folded and often the ventral lobe overlaps the haptor peduncle and/or the posterior end of the body (Fig. 1c). When measuring the haptor, this morphology must be considered in order to measure the body length and the haptor dimensions (see Fig. 1c, d). Moreover, when clamps are counted, possible gaps in the sequence of clamp frills (Fig. 1e, arrowhead) or possible missing pieces of haptor (Fig. 1e, arrow) must be considered, taking into account that the most distal clamps at the ends of the haptoral lobes are smaller.
In general, we recommend the revision and adequate counting of some discrete characters as the number of clamps or testes in the descriptions of several previously described species of Microcotyle, as some ranges are often abnormally high (e.g. 59-126 clamps for M. visa or 9-24 testes for M. erythrini, see [2,3]). We must be particularly rigorous with this consideration as these traits are key in the species diagnoses of polyopisthocotyleans. For example, the number of clamps herein reported for M. whittingtoni n. sp. is 60-78, as the much higher range reported for M. erythrini of González González [36] in Dentex dentex (110-120) was not confirmed and the drawing and the photomicrograph show that their specimens had 60 clamps [36]. Regarding testes, it should be considered that they are flattened and stacked in at least Fig. 7 Plot of 87 specimens of Microcotyle spp. in the first plane of the PCA. Ellipses indicate 95% confidence intervals two dorsoventral levels, so they must be detected and counted at different depth levels.
Some traditionally used morphological traits are intrinsically highly variable, and must be considered with extreme caution when used for taxonomy. Total length has been considered to characterize species such as M. archosargi and M. lichiae which are, in general, much larger than M. erythrini (s.s.) and similar species; however this trait is uncertain as monogenean sizes are known to be highly dependent on host size [32,43]; e.g. Mendoza-Franco et al. [25] described smaller specimens of M. archosargi, lowering the range of body length to numbers that overlap with most of the similar species in sparid fishes (see Additional file 3: Table S3). In the case of M. lichiae, the original description was based on a single specimen and there are no data for its intraspecific variability. The number of spines in the genital atrium has been used more recently to differentiate species of Microcotyle; this character is often highly variable (e.g. for M. isyebi Bouguerche et al. [3] reported 154-267 vs 272-395 in the present study) and can depend on the condition of the specimen (e.g. incorrect fixation or genital atrium more or less evaginated) or discordances related to the observers. The "pockets" of the genital atrium (sensu Mamaev [44]; posterior small chambers), typical of the genus, have also provided taxonomic information. Mamaev [44] already indicated that the presence or absence of spines in these "pockets" was a good diagnostic character. For example, Bouguerche et al. [3] also stated that M. isyebi shared the presence of genital atrium "pockets" with the other Microcotyle species parasitic in sparid fishes (i.e. M. archosargi, M. erythrini, M. isyebi and M. visa) and not in species parasitic in fishes of other families (i.e. M. donavini and M. lichiae; these authors also listed M. pomatomi and Microcotyle sp. ex H. dactylopterus [24] but "pockets" are present in these species, see Remarks to M. isyebi above). "Pockets" are often not described and sometimes not clearly drawn (e.g. M. pomatomi [32]), as sometimes they can be unarmed or armed with a few spines [44]. Moreover, when the genital atrium is evaginated, chambers often become indistinguishable in ventral view. Their absence implies a different general structure of the genital atrium, a feature used for differentiation at the generic level within the subfamily Microcotylinae [30,45]. Our last considerations of the traditionally used diagnostic traits refer to the dimensions of the soft muscular organs such as the pharynx or the genital atrium, both contractile and highly variable depending on the specimen, often mentioned in species descriptions (e.g. [2,43]). All these soft organs can entail diagnostic evidence, but reliable differences should be outstanding, mostly referred to their volume or area, and if possible, relative to the specimen size.
The use of the correct tools and procedures can allow that the currently genetically differentiated species (M. erythrini, M. isyebi and M. whittingtoni n. sp.) become pseudocryptic with defining diagnostic characters or combinations of characters. When the morphometric data of individual worms was integrated in the PCA, the resulting components could not be useful to separate species but provided useful information on specimen groupings based on their shape. The results of the PCA in the present study illustrated that additional diagnostic information can be extracted from the general form of the worms, particularly regarding the relative dimensions and arrangement of the haptor and the remaining of the body. In view of this evidence, we suggest new diagnostic characters revealing previously unnoticed morphological differences: (i) haptor dimensions including anterior and posterior lobes (the larger values for haptor length to body length ratio and for anterior/posterior haptoral lobe length ratio differentiate M. erythrini (s.s.) from M. isyebi and M. whittingtoni n. sp.); (ii) thickness of the clamps (the higher ratio between "c" sclerite maximum width/total clamp length differentiates M. whittingtoni n. sp. from M. isyebi and M. erythrini (s.s.)); (iii) relative size and shape of spines of the "pockets" of the genital atrium (spines of the "pockets" in M. whittingtoni n. sp. appear to be longer and more curved than those of M. isyebi and M. erythrini (s.s.)); (iv) extension and symmetry of the posterior extremities of vitelline fields (posterior extremities of vitelline fields always asymmetrical in M. whittingtoni n. sp. vs occasionally symmetrical in M. isyebi and M. erythrini (s.s.)); and (v) shape of the tip of the abopercular filament of the egg; the solid (capitated or pointed) tips of the abopercular filaments differentiate M. erythrini (s.s.) from M. isyebi and M. whittingtoni n. sp. We propose that the region that can provide more taxonomic information is the haptor, taking into account its three-dimensional structure as an oval to fusiform (when pointed at both ends) "foot" holding a body perpendicularly inserted, directly or through a peduncle (Fig. 1c-e). In this way the total and relative haptor dimensions must include both lobes (anterior and posterior), and one of them is often unnoticed in mounted specimens because they fold over the body (see Fig. 1c, d). In fact, some authors have described the haptor of some Microcotyle species as triangular (e.g. [3,26,33,34]) only referring to the lobe not folded over the body. In this way, M. erythrini (s.s.) can be defined by its relatively longer ventral lobe, the one that is usually unnoticed as it is adhered to the body in permanent mounts. As a note of caution, we must stress the need of examination of adult specimens only, as the relative dimensions of the haptor are known to change significantly during the development (see, for example Machkewskyi et al. [43]). The shape and size of the clamps also provides useful taxonomic information. These structures are usually described only as Microcotyle-type, and the width and length are provided (sometimes wrongly addressed, see Additional file 3: Table S3, Additional file 4: Table S4 and Fig. 1a for correct measuring). However, within this morphological description, some variations can be found. A more detailed study of clamp features can provide further taxonomic information. For example, the accessory sclerite ('e') is herein described as trifid or trident-shaped for all three species analysed, but it is mostly not described and not drawn, and the few authors drawing the sclerite represent it as single or lancet-shaped (e.g. [24,46]). We also suggest that more attention should be paid to the thickness of the clamps: among the three species herein analysed, M. whittingtoni n. sp. shows noticeably thicker clamps; we explored this attribute through the width of the anterolateral sclerite ('c') in absolute value and in relation to clamp length, as this region of the clamp appeared to be constant in all the specimens observed. The number of the spines in the "pockets" of the genital atrium is sometimes reported separately, but no specific information on the shape of these spines is usually found, except sometimes detailing that they are equal to those in the main chamber of the genital atrium [2,3,24]. Interestingly, these spines were observed to be longer than the spines of the main chamber in all three species herein described, and those in M. whittingtoni n. sp. were distinctly curved; the lack of information from other species prevented us to reach to further taxonomic conclusions, but we encourage the authors to provide specific information on the spines in the "pockets" in their descriptions of species of Microcotyle.
In the specimens of Microcotyle from the Spanish Western Mediterranean we observed some differences in the extension of the posterior extremities of vitelline fields (also including the extension of the caeca, as they accompany the vitellarium): extending into the haptor or peduncle in M. erythrini (s.s.) and into the haptor in M. whittingtoni n. sp. and prehaptoral in M. isyebi. However, this trait was not here suggested to characterize M. isyebi as according to the original description the posterior extension of the left caecum (and consequently the accompanying vitelline fields) extends into haptor "for a short distance" of the specimens from off Algeria [3].
This character may be dependent on the degree of contraction of the specimen, and therefore all specimens should be fixed and mounted in a similar way to be comparable. Other aspect related to the posterior extension of the vitelline fields of the vitellarium is their symmetry. We observed that the posterior extensions of the vitelline fields were always asymmetrical in M. whittingtoni n. sp., while in the other two species we found both specimens with symmetric and asymmetric vitelline fields. Gill polyopisthocotyleans show more or less distinct asymmetry related with the side of the gill filament they attach to [47,48]; interestingly Bouguerche et al. [3] reported that left caecum (and consequently the accompanying vitelline fields) was longer in M. isyebi, while in all the species herein observed included specimens with both dextral or sinistral asymmetry.
Mamaev [44] described the eggs of Microcotyle spp. as two-filamented, with usually long opercular and shorter abopercular filament, but no further morphological details are normally provided in the species descriptions. The examination of the new specimens from the Spanish Western Mediterranean also revealed differential details regarding the eggs such as the different shapes of the end of the abopercular filament: solid (pointed or capitate) in M. erythrini (s.s.) (Fig. 4f ) and hollow (bifid or cupshaped) in M. isyebi and M. whittingtoni n. sp. (Figs. 5f,  6f ). Other possible differential details were observed such as the type of connection between the egg and the opercular filament: abruptly connected in M. erythini (s.s.) and M. isyebi (Figs. 4f, 5f ) and inserted through a gradual transition in M. whittingtoni n. sp. (Fig. 6f ). This trait is not used for diagnosis in the present study as it requires a more standardized description. More detailed descriptions are recommended as this trait can be taxonomically useful and other authors, e.g. Sproston [37], have already reported interspecific differences regarding the egg shape. The information on this trait can be limited as the egg shape varies depending on the condition and presence of uterine eggs.

Conclusions
The present study suggests new diagnostic morphological traits to differentiate Microcotyle spp. in Mediterranean sparids and shed light on the case of M. erythrini species complex changing its previously considered cryptic status. More detailed descriptions are recommended, including molecular data, preferably of more informative gene markers regarding the interspecific differences in the polyopisthocotyleans such as cox1 [41,42,49], but also 28S rDNA sequences as they can provide useful complementary information. This study also shows that M. erythrini (s.s.) is not species-specific (even not genusspecific) to its hosts, as it parasitizes Pa. pagrus in addition to the type-host, P. erythrini; therefore, although the host species must continue as referential in the taxonomy of Microcotyle spp., a new host record does not necessarily mean a new species. However, further studies are needed in order to establish the morphological traits defining the microcotylids, especially for genera such as Microcotyle, with numerous species reported worldwide.