Multidisciplinary investigation on the catfish parasite Hamatopeduncularia Yamaguti, 1953 (Monogenoidea: Dactylogyridae): description of two new species from India, and phylogenetic considerations

Abstract Hamatopeduncularia was erected with Hamatopeduncularis arii as the type species. This genus comprises monogenoidean species mostly found as ectoparasites of marine catfishes belonging to the Ariidae. There is a significant taxonomic ambiguity among Hamatopeduncularia species due to their morphological similarity, but so far only a few morphological studies have succeeded in addressing interspecific variation and relationships. Moreover, little molecular data is available for this genus. A multidisciplinary, integrated study consisting of morphological, morphometric and molecular analyses was conducted on different species of Hamatopeduncularia recovered from the gills of two marine catfishes, Arius jella Day and Plicofollis dussumieri (Valenciennes). Five species of Hamatopeduncularia, two of which represent new species, were investigated: H. arii, H. elongatum, H. thalassini, H. madhaviae sp. nov. and H. bifida sp. nov. Phylogenetic analysis was performed using the 18S rDNA sequence as a molecular marker. The most important results of the present work are: (1) the multidisciplinary description of two novel species; (2) the multidisciplinary redescription of two species and of the type species of the genus; (3) the first molecular characterisation of 18S rDNA sequences of five species of genus Hamatopeduncularia; and (4) molecular support for the monophyly of the genus. http://zoobank.org/urn:lsid:zoobank.org:act:1333F4CC-E497-4D0A-AD7D-276D44AE6413 http://zoobank.org/urn:lsid:zoobank.org:act:43D18F75-6F4A-4F9B-8C00-6234E5BA6526


Introduction
India, with its vast coastline of 6483 km, hosts approximately 23 species of catfishes belonging to the family Ariidae (Order Siluriformes). Though only 11 species of Ariidae are used in commercial fisheries, they form the major part of the biomass of marine fish landings (Sudarsan et al. 1990), therefore representing an important commercial fishery resource along the east coast of India. Fish landings along the Visakhapatnam Coast are mainly concentrated on two species of catfishes, Arius jella Day and Plicofollis dussumieri (Valenciennes), which act as hosts to many metazoan parasites, especially ectoparasites such as monogenoids.
Among them, Dactylogyridae Bychowsky, 1933 are now attracting significant interest within the scientific community due to their rich biodiversity, with more than 379 species distributed under 31 genera (Mendoza-Palmero et al. 2015). Among Dactylogyridae, many new species of Hamatopeduncularia have been described in recent years, but unfortunately most of them were solely based on morphological data (Domingues et al. 2016;Yamada et al. 2017). However, some descriptions of Dactylogyridae, including both morphological and molecular data, were recently performed (Acosta et al. 2017(Acosta et al. , 2018Francová et al. 2017;Verma et al. 2017a,b,c;Mendoza-Franco et al. 2018).
Hamatopeduncularia was first erected by Yamaguti (1953), with H. arii as the type species. It is a species-rich genus: to date, about 39 species have been described, 24 of them recovered as ectoparasites of fishes belonging to Ariidae. The presence of the haptoral digits, with each digit ending in a marginal hook (simple anchors without any expanded outer roots), is traditionally considered the synapomorphy of the genus. Discrimination among species of Hamatopeduncularia is traditionally performed based on the following morphological characters: male copulatory organ; haptoral components such as anchors, bars and hooks; and vaginal structures (e.g. Domingues et al. 2016). Nevertheless, this sometimes has caused serious problems in species discrimination, especially when similar morphological traits were reported (Shinn et al. 2010;Strona et al. 2013). Thus, in line with the most recent studies on Dactylogyridae (Acosta et al. 2017(Acosta et al. , 2018Verma et al. 2017a,b,c;Mendoza-Franco et al. 2018), a multidisciplinary characterisation of the genus including the study of molecular markers for species description is recommended.
In the present study, we recovered different morphotypes of these parasites from the marine catfishes A. jella and P. dussumieri caught during a 2-year period of investigation (January 2013-December 2014). To validate the taxonomic status of these Hamatopeduncularia parasites, we applied an integrative taxonomic investigation combining morphological-morphometric with molecular study, i.e. performing morphological observations, also with scanning electron microscope (SEM), processing morphometric data with multivariate analysis methods, and, for the first time on this genus, sequencing the 18S rDNA and producing a molecular phylogenesis on this basis. Morphometric data were also used for the quantitative characters, and multivariate analysis methods were applied to identify the species-specific pattern of traits for each investigated species.
A complete morphological account, supported by statistical evidence, was obtained for two novel Hamatopeduncularia species, H. madhaviae sp. nov. and H. bifida sp. nov., and for three already known species of the genus, H. arii, H. elongatum Lim, 1996 and H. thalassini Bychowsky and Nagibina, 1969. Results were compared with previous morphological works (Bychowsky & Nagibina 1969;Lim 1996;Yao et al. 1998;Domingues et al. 2016). The molecular characterisation that was performed supported the morphological results in most cases.

Preparation for morphological characterisation
Freshly caught ariid fishes were brought to the laboratory where they were identified and examined for the presence of monogenoidean parasites. For the collection of monogenoidean worms, gill arches were removed and placed in a petri dish containing 8‰ saline water to avoid salinity stress, which may lead to contraction of forms. This saline solution was prepared by mixing one part seawater (salinity: 32‰) with three parts tap water. Sometimes an 8‰ solution of NaCl in water was also used (Cribb & Bray 2010). All gill parts were scraped with a needle to detach the parasites. The contents were transferred to an evaporating bowl and washed well by the process of decantation. This step was repeated until the contents appeared clear and the parasites were cleaned of gill mucus. Then the contents were taken in a small Petri dish and examined under a binocular stereo microscope (Nikon SMZ800) at 5× magnification. The detached monogenoids were picked up using a fine needle and transferred individually into a drop of glycerine on a slide for the preparation of semi-permanent material. The preparation was then covered with a coverslip and sealed for the examination of sclerotised structures. Some parasites mounted on slides with water droplet were also observed in vivo under a histological microscope (Nikon YS100) at 10-100× magnification. They were stained with alum carmine, dehydrated through ascending grades of alcohol and mounted in Canada balsam for preparation of permanent slides.
From these permanent slides schematic line drawings were made with a camera lucida connected to a microscope Nikon ECLIPSE 50i. Measurements were taken with the use of an ocular micrometre; a maximum number of 10 specimens for each species of parasite were used for this purpose. All measurements are given in micrometres (μm) and presented as the average, followed in parentheses by the range. Body length includes haptor. Measuring method and terminology concerning the hard structures of the haptor (i.e. anchor, marginal hooks, and dorsal bar) are according to Gusev (1977). Identification of the parasites was performed by studying forms in freshly dead and, sometimes, live condition and by observing details from mounted specimens under the light microscope. The utmost care was taken to study the observed modifications of haptoral armature and of copulatory organs, which play an important role in the identification of monogenoidean parasites. Papers by Yamaguti (1953), Tripathi (1959), Gusev (1976), Kearn and Whittington (1994) and Lim (1996) were considered both for species identification/description and establishment of novel taxa.
Good-quality permanent slides with representatives of each species of parasite have been deposited in the helminthological collections of the Central Zone Regional Centre, Zoological Survey of India, 68-169, Vijoy Nagar, Jabalpur -482 002, Madhya Pradesh; the assigned museum registration numbers are listed in the Results section.

Morphometric analysis
To investigate differences and relationships among the five species of Hamatopeduncularia, we collected morphometric data for 53 characters, including details of general morphology, copulatory structures and haptoral hard parts, from 10 specimens for each species. These data were subsequently analysed with multivariate statistics such as Euclidean cluster analysis, multidimensional scaling (MDS) and principal component analysis (PCA). These analyses were performed using PRIMER v. 6 (Plymouth Routines in Multivariate Ecological Research) and STATISTICA v. 8 software.
The dendrogram clustering method uses the similarities or distances between the species taken as samples while forming the clusters. Similarities are a set of rules that serve as criteria for grouping or separating the species, based on multiple dimensions. The most straightforward way of computing distances between the objects in a multidimensional space is to compute Euclidean distances. A correlation matrix (morphological measurements vs. parasite species) was prepared based on obtained data, and an overall mean was derived for each parameter. The data were then analysed by hierarchical clustering through group average linking, following Euclidean dendrogram cluster analysis procedures implemented in PRIMER v. 6. The output of this analysis is shown as a dendrogram of relationships between groups/species. Ordination plots (MDS) were then produced from these similarity measures through PRIMER dendrogram protocol. Significance tests of samples were made using the analysis of similarity (ANOSIM) randomisation test (Clarke & Green 1988).
PCA is a statistical ordination technique that reduces the number of dimensions (variables) in a set of data, finds linear combinations of the more correlated variables that explain most of the variance, and eventually assigns a component score to each original individual (Foottit & Sorensen 1992). The morphometric data were subjected to PCA using the software STATISTICA v. 8. PCA was used in estimating the morphometric variation among species and to identify the variables that substantially contribute to this variation. The results are presented in the form of graphs and tables. Finally, a PCA plot was obtained using the components showing high variance.

SEM preparation of Hamatopeduncularia arii
Freshly collected H. arii samples were used to obtain SEM preparations according to the protocol by Fisher et al. (2012) with slight modifications. Parasites were washed thoroughly in 8‰ saline water until all the debris was removed, fixed in 2.5% glutaraldehyde for 2 hours at 4°C, washed with 0.1 M phosphate buffer (3 changes of 15 min each at 4°C), and dehydrated in an ascending series of acetone. Then specimens were dried in a critical point drying apparatus,  The prepared specimens were placed in a JSM-6610LV SEM for observation.

Molecular and phylogenetic analysis
After the collection of parasites from the gill filaments, each specimen to be processed for molecular analysis was observed and identified under a Nikon YS100 stereo zoom microscope, and then preserved in 95% ethanol at -20°C. DNA was extracted using a NucleoSpin™ PlantII DNA extraction kit (Macherey-Nagel GmbH & Co., Düren NRW, Germany) according to the manufacturer's instructions, and stored at -20°C until further processing. The 18S rDNA was amplified according to Modeo et al. (2013b) using the following primers: Forward 18S F9 as forward primer (5ʹ -CTGGTTGATCCTGCCAG -3ʹ) (Medlin et al. 1988) and 18S R1513 Hypo as reverse primer (5ʹ -TGATCC TTCYGCAGGTTC -3ʹ) (Petroni et al. 2002).
The polymerase chain reaction (PCR) cycle, performed according to the manufacturer's instructions (Takara Bio, USA), was set up in a Gene Amp system C1000™ Thermal Cycler (Bio-Rad, Hercules, CA) with the following steps: 3 min at 94°C for the initial denaturation, followed by 35 cycles of 30 sec at 94°C for denaturation, 30 sec at 55°C for primer annealing, 2 min at 72°C for extension, and a final extension for 6 min at 72°C. An aliquot (5 μL) of each amplicon was checked on 1% agarose gel. Gel electrophoresis run was performed for 40 min at 150 V along with a 100-10,000-bp DNA molecular weight marker and, later, stained with ethidium bromide for 30 min. PCR bands were visualised under a UV light transilluminator and photographed with a Gel Doc system. PCR product was purified using a Euro Gold Cycle Pure Kit (Euroclone SpA, Italy) and stored at -20°C until it was sent for sequencing.
Semi-nested PCR was performed using the primer combinations 18S F9 + 18S R1052 and F300 + 18S R1513 Hypo for the PCR samples of H. elongatum and H. bifida sp. nov. to obtain more abundant amplification products of the target region [18S R1052: 5ʹ -AACTAAGAACGGCCATGCA -3ʹ -Rosati et al.  . PCR cycles were set as previously, except for the extension time: 1.5 min. PCR products were sequenced with the GATC Biotech sequencing services, using three internal primers as in Nitla et al. (2018) The resulting 18S rDNA electropherograms were checked and assembled using Chromas Lite 2.1 software. The sequences thus obtained were deposited in the National Center for Biotechnology Information (NCBI) GenBank database (for the accession numbers see Table I). Hamatopeduncularia sequences from the present study and the latest sequences available in the NCBI database belonging to the subclass Polyonchoinea were automatically aligned with the ARB software package (Ludwig et al. 2004) against the SILVA 102 SSU rRNA database (Pruesse et al. 2007). Then, alignment of monogenoidean sequences was manually refined. For phylogenetic analyses 85 sequences were employed: we selected 55 sequences belonging to the order Dactylogyridea, the present five sequences, plus 25 more as outgroup (23 Gyrodactylidea plus two incertae sedis). The alignment was reduced in length according to the shortest sequence, producing an 1806-character matrix. Maximum likelihood (ML) analyses (PHYML 5.3.2; Guindon & Gascuel 2003) and Bayesian inference (BI) analyses (MrBayes 3.2; Ronquist et al. 2012) were performed, with the GTR + I + G substitution model, as indicated by the Akaike information criterion (AIC), calculated by jModelTest 2.2 (Darriba et al. 2012). ML analysis was performed with 1000 pseudoreplicates, while for BI analysis, three different Markov chain Monte Carlo runs were used, with one cold chain and three heated chains each, running for 1,000,000 generations. The ARB NJ algorithm (Ludwig et al. 2004) with the "similarity" correction was employed to calculate the similarity matrix.

Results
The description/redescription of the five species of parasites recovered on the marine catfishes A. jella and P. dussumieri from the Visakhapatnam Coast, Bay of Bengal, between January 2013 and December 2014 follows. For each parasite species, a detailed account, the indication of the host from which it was isolated, some ecological notes (prevalence of infection and mean intensity), and some remarks concerning previous descriptions are provided. All measurements are in µm.

Site of infection
Gill lamellae.

Voucher material
Slide A-18,080 with 10 specimens.

Mean intensity
Eighteen parasites per infected host (total number of recovered H. arii specimens: 1063).

Redescription
Based on 10 adult specimens. Fusiform parasite measuring 1342    144-192) long, 308 (192-420) wide, digitate, provided with 12 short finger-like projections known as digits (Lim 1996); 14 marginal hooklets, 12 of which are located on haptoral digits and one pair located near ventral anchors. A short peduncle present between proper body and haptor. The two ventral anchors are supported by a simple ventral bar having a length of 7 (5-8) and a maximum width of 77 (72-96). Ring pads found on the inner roots of the ventral anchors. A pair of dorsal anchors also present, with inner length 47 (36-52), inner root length 20 (12-24), outer root length 12 (8-16), and a point 25 (24-28) long. Bar supporting the dorsal anchors with a length of 7 (5-8) and a width of 75 (72-96). Dorsal bar with anchor-shaped appendix. The haptoral hard parts of H. arii were also examined under SEM (Figure 2).

Remarks
The present form has been identified as H. arii based on the structure of the copulatory complex, position of gonads, and structure and armature of haptor, which are very similar to those previously reported (Yamaguti 1953;Bychowsky & Nagibina 1969;Paperna 1977;Lim 1996). In the specimens here investigated, both ring pads and appendix on dorsal bar were observed (see Discussion).

Site of infection
Gill lamellae.

Mean intensity
Seven parasites per infected host (total number of recovered H. elongatum specimens: 1935).

Remarks
The present form was identified as H. elongatum based on the similarities of copulatory tube and haptor armature proposed by Lim (1996)

Site of infection
Gill lamellae.

Remarks
This parasite was identified as H. thalassini mainly based on size of anchors, presence of specialised haptoral structures such as ring pads, appendix, shape of the copulatory organ, accessory piece with spatulated end, and vaginal armament, which resemble the corresponding characters of the previous description by Bychowsky and Nagibina (1969). The present study constitutes both the first report of this species from the catfish A. jella and the first report from Visakhapatnam coast, Bay of Bengal, India (see Discussion).

Site of infection
Gill lamellae.

Mean intensity
Ten parasites per infected host (total number of recovered H. bifida sp. nov. specimens: 2204).

Etymology
The term "bifida" (N.L. fem. adj., meaning "bifurcated") refers to the bifurcated tip of the accessory piece of the male copulatory organ.

Remarks
This parasite is characterised by a copulatory organ provided with a tubular accessory piece with a bifurcated tip, slightly tilted at a right angle, and a broad base. The haptoral armature consists of six pairs of thin, long peduncles, each carrying a marginal hook, a "V"-shaped dorsal transverse bar and a pair of cement glands. It further differs in the structure of the accessory piece, which is bifid and not reported in other species of Hamatopeduncularia so far. Additionally, the length of the peduncle connecting the proper body with haptor is considerably long in this species. Although a certain grade of variation is present in haptoral armature (Figures 7 and 8), based on the differences with other species of Hamatopeduncularia the present morphospecies is considered new and designated as H. bifida (see Discussion).

Site of infection
Gill lamellae.

Mean intensity
Thirteen parasites per infected host (total number of recovered H. madhaviae sp. nov. specimens: 752).

Etymology
The term "madhaviae" is in honour of Professor Rokkam Madhavi, in recognition of her contribution to helminthology.

Remarks
This parasite species considerably differs from all other species of the genus reported so far in the structure of haptoral armature and copulatory organs; thus, we are proposing it as the new species H. madhaviae (see Discussion).

Morphometric analysis
Euclidean cluster analysis. The Euclidean cluster analysis identified three clusters at a distance of 9 with stress at 0, which were separated clearly with ANOSIM: Global R = 1.0; P = 0.1%. The first cluster (Group I) was represented by H. elongatum and H. bifida sp. nov.; the second cluster (Group II) was an assemblage of two species, H. thalassini and H. arii; and the third cluster (Group III) included only one species, H. madhaviae sp. nov. (Figure 7). The plot obtained as result of multidimensional scaling supports the Euclidean cluster analysis (Figure 8).
Hamatopeduncularia elongatum and H. bifida sp. nov. in Group I showed similarities in the structure of their haptoral hard parts, with simple bars and anchors. In Group II, H. arii and H. thalassini possessed similar haptoral armature (i.e. bars with expanded wings, anchors with ring pads on the inner root of ventral anchors, and anchor-shaped appendix). Hamatopeduncularia madhaviae sp. nov., forming the separate cluster Group III, was characterised by a long dorsal bar, a curved ventral bar, a simple, short copulatory tube, and a narrow peduncle.
Principal component analysis (PCA). In Euclidean cluster analysis, the discrimination of species is mainly based on their morphometric characters presented in a two-dimensional model with fairly convincing groups in the form of a Euclidean dendrogram and mean dimensional scaling. However, PCA is more commonly used to understand morphological variations at each parameter level. Based on the morphometric variations of the data, principal components were identified in the PCA. Those having an eigenvalue > 1 were utilised to describe the data. The correlation values for each parameter tested with each axis and their discrimination into principal components (PC1, PC2, PC3) are provided in Table III and Figure 9A. Group I (eigenvalue 21.6 and total variance 40.7%) isolated the species H. elongatum and H. bifida sp. nov. from other species based on the following traits: body length and width, distance between anterior cleft and eyespots, distance between eyespots, pharynx width, oesophagus, caeca, testis length and width, ovary length and width, seminal receptacle length  and width, vagina length and width, copulatory tube length, haptor length and width, dorsal bar length, dorsal and ventral inner root, ventral outer root, 2 nd and 6 th pairs of marginal hooks. Group II (eigenvalue 13.3 with a variance of 25%) corresponded to the species H. thalassini and H. arii. Principal components supporting this group were pharynx length, seminal vesicle length and width, egg length and width, prostate gland length and width, dorsal and ventral bar length, ventral bar width, dorsal anchor length, dorsal base, ventral shaft, 1 st , 5 th and 7 th pairs of marginal hooks. Hamatopeduncularia madhaviae sp. nov. constituted a separate cluster, Group III, with an eigenvalue of 11.9 and variance of 22.4%. Principal components that supported this group were copulatory tube width, accessory piece length and width, dorsal outer root, dorsal main piece, dorsal shaft, dorsal recurved point, ventral anchor length, ventral main piece, ventral base, ventral recurved point, 3 rd and 4 th pairs of marginal hooks. The total variance for the first three axes together scored 88.2%. The morphological characters responsible for the discrimination of the three groups in the cluster analysis were identified in the PCA. Hence, ordination analysis is recognised as the best tool to explain the projection of individual variable scores and represented in the form of variance (%).
The total eigenvalues for each principal component are shown in Table IV. The results show the highest total value of 21.6 for PC1, and eigenvalues for PC2 and PC3 of 13.3 and 11.9, respectively. The total variance for the first three axes together scores 88.2%, and there is another axis with an eigenvalue of 6.3 with a total variance of 11.8. Compared with the other three axes, the values obtained for the fourth axis are low. Hence, this axis was merged with the third axis in this analysis ( Figure 9A). The results obtained from the PCA are in close association with those from the cluster analysis, and the morphological characters which are responsible for the discrimination of three groups in the cluster analysis were identified in the PCA, as major parameters that support the differentiation of the species from one another. The scores of individual morphological characters obtained through the PCA are shown in Table III. The score plot revealed clear separation among the species based on morphometric characters (Table V). The scatter diagram of PCA 1 versus PCA 2 shows the relationships among the species ( Figure 9B). The sister genus of Hamatopeduncularia was not properly resolved in our analysis at the molecular level, being represented either by genus Bychowskyella or by genus Mizelleus, with genera Mymarothecium and Anacanthorus as weaker options; however, the family Dactylogyridae comprises more than 200 genera and our analysis unfortunately could include only a small fraction of them, i.e. those for which the molecular marker sequence is available. Identity values between the 18S rDNA sequences of the five Hamatopeduncularia species herein studied and those of their closest relatives according to the present molecular investigation ranged between 86.4% and 90.5% (Table VI). Hamatopeduncularia, Bychowskyella and Mizelleus, together with Mymarothecium and Anacanthorus, formed a monophyletic clade sister to all other sequenced members of the family Dactylogyridae.  However, the susceptibility of fish hosts to parasites can be influenced by various factors such as host sex, ontogenetic alterations in behaviour, physiology and ecology (Takemoto et al. 1996), and the evolutionary history of the host-parasite relationship is even more important.
As for the prevalence and mean intensity of infection for the different investigated monogenoids, H. madhaviae shows the highest prevalence value (43%) and H. arii shows the highest mean intensity value (18) with respect to the other investigated species.
Species identification of monogenoidean parasites is traditionally based only on morphological characters, including the haptoral armature and copulatory organ, which are considered taxonomic markers (Gerasev 1992;Lim 1995Lim , 1996. In the present study, the five species of Hamatopeduncularia were investigated through a multidisciplinary integrated study approach: morphological analysis was further supported by morphometric, molecular/phylogenetic analyses, and SEM analysis (this only for H. arii). Previous SEM investigations are not available; nevertheless, observed qualitative characters can be compared regardless of the particular technique used as they are not linked to any measurement unit. Our SEM investigation of H. arii (Figure 2), which is the type species of the genus, showed the generic diagnostic characters (digitate haptor, armed with marginal hooks; two pairs of anchors and two bars; outer root of anchors usually not expanded into wings), and also species-specific characters such as the appendix on dorsal bar (Yamaguti 1953;Bychowsky & Nagibina 1969;Lim 1996).
Hamatopeduncularia arii, reported from A. jella in the present study, resembled previous descriptions in all aspects (Yamaguti 1953;Bychowsky & Nagibina 1969;Paperna 1977;Lim 1996) except for a few morphological and morphometric variations. Both ring pads and the appendix on the dorsal bar were observed in the present specimens. However, the ring pads were not mentioned by Yamaguti (1953), Bychowsky and Nagibina (1969) or Paperna (1977), and the appendix on the dorsal bar was not observed by Lim (1996). Hamatopeduncularia arii was first reported by Yamaguti in 1953 from Arius spp. from Borneo. In 1969, Bychowsky and Nagibina redescribed this species from Arius falcarius (now Arius arius), Arius leiotetocephalus (now Plicofollis nella) and Arius maculatus from the South China Sea. Paperna (1977) found it on Arius thalassinus (now Netuma thalassina) from Kenya. Lim (1996) reported this species from A. maculatus, with characters similar to those of the original description except for the presence of ring pads on the inner root of the ventral anchors. As for the Indian region, Gupta and Khullar (1967) reported H. arii from Arius sp.
Hamatopeduncularia elongatum was reported from A. jella in the present study and found to be in congruence with the description provided for the species by Lim (1996) (i.e. nature of copulatory tube and haptor armature). It was first reported by Lim (1996)    Hamatopeduncularia thalassini was reported from the host A. jella and first described by Bychowsky and Nagibina (1969) from Tachysurus thalassinus (now N. thalassina) from the South China Sea (Hainan Island); later Paperna (1977) briefly described this species with a few minor differences from the original account of Arius sp. Eventually, Lim (1996) provided a complete description of the parasite, again from N. thalassina.
Hamatopeduncularia bifida sp. nov. resembles H. wallagonius Singh et al. 1995, H. malayanus Lim, 1996and H. isosimplex Lim, 1996 in the general appearance of morphological characters, but differs from them in possessing a "V"-shaped dorsal bar and a straight ventral bar slightly bent in the centre, vs "V"-shaped dorsal and ventral bars. It further differs in the structure of the accessory piece, which is not reported as bifid in other forms. Besides, the length of the peduncle connecting the proper body with the haptor is considerably long in the present specimen. Interestingly, H. bifida sp. nov. formed a cluster with H. elongatum in the morphometric analysis, though a lot of variation is seen in the copulatory structure (Figures 7 and 8). Based on the differences noted with other species of the genus reported so far, the present species is considered a new species.
Hamatopeduncularia madhaviae sp. nov. from the host P. dussumieri differs from other species in the shape and structure of its copulatory organ. However, it shows some similarities with H. papernai Lim, 1995 andH. ritai Rastogi et al., 2005 in characters such as anchors, ventral bar, copulatory tube and marginal hooks, and their arrangement. The copulatory tube is slightly bent at the distal end in both H. papernai and H. madhaviae sp. nov. However, the only difference between the latter two species is the proximal end of the copulatory tube that is "C" (hook) shaped in H. madhaviae sp. nov., and closed cone-shaped in H. papernai. It differs from H. ritai in having an almost straight dorsal bar slightly bent in the middle, and in the absence of accessory piece and sleeve sclerites, whereas in H. ritai two accessory pieces and dorsal anchors with sleeve sclerites are present, and the dorsal bar is "V" shaped. As the present species showed significant morphological variation from other species reported so far, it is considered a new species.

Morphometric analysis
Based on morphometric data, both Euclidian analysis ( Figure 7) and PCA (Figure 9) differentiated the five investigated Hamatopeduncularia species, H. arii, H. thalassini, H. bifida sp. nov. and H. elongatum (all found on the host A. jella), and H. madhaviae sp. nov. (retrieved on P. dussumieri), into three different clusters. Thus, the present morphometric study provides a solid basis for the separation of species within the same genus.
The structure of both haptoral and copulatory apparatus helped in defining these three groups. Within each of the two groups comprising two parasite species, all members exhibited similarities in their haptoral apparatus structures and clustered together in both the Euclidean cluster analysis and PCA: H. arii and H. thalassini (Group II) have similar haptoral sclerite morphologies; H. bifida sp. nov. and H. elongatum (Group I) show similar length of hooks and anchors. Nevertheless, within each of these two groups the two species of parasites can be easily distinguished by differences in the structure of the copulatory apparatus. Moreover, H. madhaviae sp. nov. from the host P. dussumieri, belonging to the third, separate cluster (Group III), showed differences as well in both its haptoral sclerite dimensions and copulatory apparatus structure.
Previous studies indicated that the haptoral morphologies of species are generally associated with attachment site on the host (niche) (Simkova et al. 2002). In the present study, two out of the four species found on A. jella (i.e. H. arii and H. thalassini) with similar haptoral morphology clustered together in the phylogenetic analysis, whereas the other two (H. elongatum and H. bifida sp. nov.), with similar haptoral sclerites, formed a second, separate cluster. Thus, it is suggested that this difference might be related to their specific niche on the host gill, i.e. attachment specificity on a particular area of the host gill.
Our cluster analysis identified an interesting set of associations when grouping parasite species according to their haptoral sclerite variability. Attachment organs of parasites, especially ectoparasites, have been considered important for host specificity. The monogenoidean haptor has been found to be highly specialised and therefore could constrain the parasites to certain sites on their hosts' gills (Rohde 1989;Lambert & El Gharbi 1995). The present study supports the hypothesis that phylogenetically close parasite species may possess similar attachment organs, which might serve to better adapt them to their hosts. The morphology of two sclerotised organs, namely the attachment organ (the haptor) and the reproductive organ (including the copulatory piece and the vagina), mainly helps in the identification of monogenoidean parasites. The morphology of the haptor is considered useful for parasite determination at the generic level, while the reproductive organ seems to be more suitable for identification at the specific level, probably because of its higher rate of change (Pouyaud et al. 2006;Wu et al. 2007Wu et al. , 2008. three genera parasitise catfishes belonging to the same order, i.e. Siluriformes, and show the dactylogyrid type of seminal vesicle (Gusev 1976).
Thus, considering both morphological and molecular results, we confirm that Hamatopeduncularia belongs to the family Dactylogyridae. Moreover, our phylogeny results are coherent with those provided by Mendoza-Franco and colleagues (Mendoza-Franco et al. 2018), whose analysis placed the 28S rDNA sequence of Hamatopeduncularia bagrae in association with the genera Thaparocleidus, Quadricanthus and Bichowskyella, within Dactylogyridae.
The monophyly of the family Dactylogyridae is well supported by high statistical value (1.00/100) in the present study. In our analysis, Ancyrocephalinae proved to be a polyphyletic taxon, with members scattered among different clades of Dactylogyridae (Klassen 1994;Lim 1998;Mollaret et al. 2000;Simkova et al. 2003Simkova et al. , 2006Plaisance et al. 2004Plaisance et al. , 2005Mendoza-Palmero et al. 2015;Müller et al. 2016), and Pseudodactylogyrinae appeared to be non-monophyletic. Future studies are needed to address the issue of non-monophyly of subfamilies highlighted by our phylogenetic analysis based on the molecular marker 18S rDNA; an integrative morphological-molecular study approach is recommended.