Lepotrema Ozaki, 1932 (Lepocreadiidae: Digenea) from Indo-Pacific fishes, with the description of eight new species, characterised by morphometric and molecular features

We review species of the genus Lepotrema Ozaki, 1932 from marine fishes in the Indo-West Pacific. Prior to the present study six species were recognised. Here we propose eight new species on the basis of combined morphological and molecular analysis: Lepotrema acanthochromidis n. sp. ex Acanthochromis polyacanthus from the Great Barrier Reef (GBR); Lepotrema hemitaurichthydis n. sp. ex Hemitaurichthys polylepis and H. thompsoni from Palau and French Polynesia; Lepotrema melichthydis n. sp. ex Melichthys vidua from Palau and the GBR; Lepotrema amansis n. sp. ex Amanses scopas from the GBR; Lepotrema cirripectis n. sp. ex Cirripectes filamentosus, C. chelomatus and C. stigmaticus from the GBR; Lepotrema justinei n. sp. ex Sufflamen fraenatum from New Caledonia; Lepotrema moretonense n. sp. ex Prionurus microlepidotus, P. maculatus and Selenotoca multifasciata from Moreton Bay; and Lepotrema amblyglyphidodonis n. sp. ex Amblyglyphidodon curacao and Amphipron akyndynos from the GBR. We also report new host records and provide novel molecular data for two known species: Lepotrema adlardi Bray, Cribb & Barker, 1993 and Lepotrema monile Bray & Cribb, 1998. Two new combinations are formed, Lepotrema cylindricum (Wang, 1989) n. comb. (for Preptetos cylindricus) and Lepotrema navodonis (Shen, 1986) n. comb. (for Lepocreadium navodoni). With the exception of a handful of ambiguous records, the evidence is compelling that the host-specificity of species in this genus is overwhelmingly oioxenous or stenoxenous. This renders the host distribution in three orders and ten families especially difficult to explain as many seemingly suitable hosts are not infected. Multi-loci molecular data (ITS2 rDNA, 28S rDNA and cox1 mtDNA) demonstrate that Lepotrema is a good generic concept, but limited variability in sequence data and differences in phylogenies produced for different gene regions make relationships within the genus difficult to define.


Introduction
Members of the digenean family Lepocreadiidae Odhner, 1905 are common parasites of fishes of the Indo-West Pacific region, particularly of coral reef fishes. Our systematic studies on this family in the waters around northern Australia and other sites in the region have hitherto been mostly reliant on comparative morphology Bray et al., 1993;Bray & Cribb, 1996a, b, c, d;Bray et al., 1996a, b;Bray & Nahhas, 1998;Bray & Cribb, 2002, 2003Bray & Justine, 2006;Bray et al., 2009a, b;Bray et al., 2010a, b;Bray & Justine, 2012). More recently, some progress has been made in our understanding of higher level lepocreadiid systematics (Bray et al., 2009b;Bray & Cribb, 2012;Bray et al., 2018) by the addition of molecular evidence. At the species level, the limits of the discriminating ability of morphological evidence has become increasingly apparent as we have attempted to elucidate the systematics of some of the larger lepocreadiid genera. Lepotrema Ozaki, 1932 is a case in point. It is encountered in a wide range of fish families in the orders Tetraodontiformes and Perciformes, with one record from a pleuronectiform. Most of the literature reports, including some from all three orders, are listed under the type-species, Lepotrema clavatum Ozaki, 1932, suggesting a very low level of specificity. Our molecular evidence presented here, based mainly on ITS2 rDNA and cox1 mtDNA sequences, together with sampling evidence, however, indicates that in general specificity is high. Most species appear to be oioxenic or stenoxenic, with no clear evidence that any individual species parasitizes multiple orders.
In addition to the molecular and host-specificity evidence presented here, we have found that it is usually possible to detect minor, but relatively consistent, morphometric distinguishing characteristics if the sample from a given host is of a reasonable size, i.e. more than three specimens. Members of Lepotrema are small, making the use of hologenophores problematical, as most distinguishing characters are ratios of measurements relative to bodylength. Nevertheless, most species are recognisable by combinations of morphometric characters, most readily visualised using graphs. In several cases, only one or two worms were recovered from a host species, and we have not been able to identify them to species.

Materials and methods
Digeneans collected from freshly killed fish were fixed by being pipetted into nearly boiling saline and immediately preserved in formalin or 70% ethanol (Cribb & Bray, 2010). Whole-mounts were stained with Mayer's paracarmine or Mayer's haematoxylin, cleared in beechwood creosote or methyl salicylate and mounted in Canada balsam. Measurements were made through a drawing tube on an Olympus BH-2 microscope, using a Digicad Plus digitising tablet and Carl Zeiss KS100 software adapted by Imaging Associates, and are quoted in micrometres, with the range and the mean in parentheses. Morphometric distinctions are derived from graphs produced using the Scatter plot function in Excel. The following abbreviations are used: NHMUK, the Natural History Museum, London, UK; MNHN JNC, Muséum National d'Histoire Naturelle, Paris, France; QM, Queensland Museum, Brisbane, Australia; WAM, Western Australian Museum, Perth, Western Australia.
The partial 28S rDNA sequences generated during this study were aligned with sequences of related lepocreadiids from GenBank using MUSCLE version 3.7 (Edgar, 2004) run on the CIPRES portal (Miller et al., 2010a), with ClustalW sequence weighting and UPGMA clustering for iterations 1 and 2. The resultant alignment was refined by eye using MES-QUITE (Maddison & Maddison, 2018). The ends of each sequence were trimmed and ambiguously aligned regions were identified and masked manually (those constituting more than three bases and present in greater than 5% of the sequences in the dataset). Bayesian inference analysis of the 28S dataset was performed using MrBayes version 3.2.6 (Ronquist et al., 2012), run on the CIPRES portal. The best nucleotide substitution model was estimated using jModelTest version 2.1.10 (Darriba et al., 2012); the TVM?I?C model was predicted the as the best estimator by the Akaike Information Criterion (AIC) and TPM2uf?I by the Bayesian Information Criterion (BIC). Bayesian inference analysis was run over 10,000,000 generations (ngen = 10,000,000) with two runs each containing four simultaneous Markov Chain Monte Carlo (MCMC) chains (nchains = 4) and every 1,000th tree saved. Bayesian inference analysis used the following parameters: ''nst = 6'', ''rates = invgamma'', ''ngammacat = 4'', and the priors parameters of the combined dataset were set to ''ratepr = variable''. Samples of substitution model parameters, and tree and branch lengths were summarised using the parameters ''sump burnin = 3,000'' and ''sumt burnin = 3,000''. Species of Mobahincia Bray, Cribb & Cutmore, 2018(MH157068), Multitestis Manter, 1931 and Neomultitestis Machida, 1982 (MH157072) were designated as functional outgroup taxa, following Bray et al. (2018).

Results
Family Lepocreadiidae Odhner, 1905 Genus Lepotrema Ozaki, 1932 The genus Lepotrema was erected by Ozaki (1932) for L. clavatum Ozaki, 1932 from a monacanthid, the threadsail filefish Stephanolepis cirrhifer (Temminck & Schlegel) (as Monacanthus c.), from the Japanese coast. The exact site of collection was not given, but this fish species is said to be distributed between Otaru and Nagasaki. A ''genital sucker'' is described ''at the bottom of the chamber (genital atrium) lying directly inside of the end part of the metraterm''. This feature, along with the distinctly dorsal excretory pore, represent the main distinguishing features of the genus. Yamaguti (1934) recognised the genus and species and reported it from the type-host in the Inland Sea, Japan, as well as in the Korean black scraper Thamnaconus modestus (Günther) (as Cantherhines unicornu) (Monacanthidae) and the cinnamon flounder Pseudorhombus cinnamoneus (Temminck & Schlegel) (Paralichthyidae) from Japanese waters. Some measurements were given but no illustration. The morphological features of the genus were not discussed. Four years later, Yamaguti (1938) re-examined the specimens from S. cirrhifer and T. modestus and synonymised the genus with Lepocreadium Stossich, 1903. He stated ''Although Ozaki distinguished his genus Lepotrema from the known members of the Lepocreadiinae by the position of the genital pore and the possession of a genital sucker, the genital pore usually lies to one side of the median line in this subfamily as defined by Odhner and the ''genital sucker'' of Ozaki is not a sucker in the true sense of the word, but a bulb-like muscular thickening of the metraterm''. Hanson (1955) followed this generic designation in reporting (but not illustrating) Lepocreadium clavatum in 'Melichthys buniva', apparently a misapplied name for the black triggerfish Melichthys niger (Bloch) (Balistidae) (see Randall, 2007), from Hawaii. She also described Lepocreadium incisum Hanson, 1955 from the same fish species, reporting a ''bulb-like muscular thickening of metraterm as described by Yamaguti'' and a ''subterminal'' excretory pore. Pritchard (1963) reported Lepocreadium clavatum from M. niger, the pinktail triggerfish Melichthys vidua (Richardson), the brown-and-white butterflyfish Hemitaurichthys zoster (Bennett) (Chaetodontidae) and the Hawaiian dascyllus Dascyllus albisella Gill (Pomacentridae), all from Hawaii, but again without any illustration. Yamaguti (1970) re-recorded Lepocreadium clavatum from M. vidua from off Hawaii, describing it in detail, but illustrating only the terminal genitalia and the proximal female system. The terminal part of the metraterm is described as an indistinct ''bipartite spherical bulb of lamellar muscle fibers''. He described a new species, Lepocreadium xanthichthydis Yamaguti, 1970 from the sargassum triggerfish Xanthichthys ringens (Linnaeus) (Balistidae) from off Hawaii with the metraterm provided with a ''bulb of lamellar muscle fibers'', and a ''dorsoterminal'' excretory pore. Dyer et al. (1988) reported Lepocreadium clavatum in the white-banded triggerfish Rhinecanthus aculeatus (Linnaeus) (Balistidae) off Okinawa, Japan, but again without any illustration. Bray et al. (1993) reported Lepocreadium clavatum from two pomacentrids, the spiny chromis Acanthochromis polyacanthus (Bleeker) and the banded parma Parma polylepis Günther, from off Heron Island on the southern Great Barrier Reef, providing illustrations of individuals from both fishes. The terminal part of the metraterm was decribed as a ''large, circular, folded muscular pad (not a sphincter)'' and the excretory pore as ''mid-dorsal, about halfway between caecal ends and posterior extremity''. They also decribed Lepocreadium adlardi  from the Bengal sergeant Abudefduf bengalensis (Bloch) (Pomacentridae), with the distal extremity of the metraterm ''clamped in prominent folded muscular pad'' and the excretory pore ''middorsal, close to level of posterior extremity of caeca''. Bray et al. (1993) first considered re-recognising the genus Lepotrema based on the structure of the distal metraterm and the dorsal excretory pore. This action was then taken by Bray & Cribb (1996c) in a review of the genus. They transferred Lepocreadium incisum, Lepocreadium xanthichthydis and Lepocreadium adlardi to Lepotrema making new combinations. They also noted the similarity of Preptetos cylindricus Wang, 1989 and Lepocreadium navodoni Shen, 1986 to members of the genus, but refrained from making new combinations ''pending further study''. They reported Lepotrema clavatum from the broom filefish Amanses scopas (Cuvier) (Monacanthidae) (illustrated) and the halfmoon triggerfish Sufflamen chrysopterum (Bloch & Schneider) (Balistidae) (not illustrated) from off Heron Island and described Lepotrema canthescheniae Bray & Cribb, 1996 from the endemic large-scaled leatherjacket Cantheschenia grandisquamis Hutchins from off Heron Island.  erected Lepotrema monile  from Ward's damsel Pomacentrus wardi Whitley (Pomacentridae) from off Heron Island. This species is problematical in that the distal metraterm is surrounded only by a ''distinct, but narrow, sphincter''. The excretory pore, however, is ''dorsal, between ends of caeca''. Machida & Kuramochi (1999) recognised the validity of Lepotrema in reporting (but not illustrating) L. clavatum in T. modestus and the redtoothed triggerfish Odonus niger (Rüppell) (Balistidae) off Japan. Machida & Uchida (2001) made the first report of L. clavatum from a pomacanthid when they recorded it from the Japanese swallow Genicanthus semifasciatus (Kamohara) from off Japan. They gave some measurements but did not describe the metraterm or excretory pore or give an illustration. In his review of the family Bray (2005) recognised the genus and Bray et al. (2009b) included 'L. clavatum' from Acanthochromis polyacanthus from off Lizard Island on the northern Great Barrier Reef in a molecular phylogeny of the Lepocreadioidea. The 28S rDNA and mitochondrial NADH dehydrogenase subunit 1 (ND1) reported by Bray et al. (2009b) for 'Lepotrema clavatum' are the only molecular data presently available for this genus.
The only evidence of the life-cycle of Lepotrema is supplied by Kondo et al. (2016), who described metacercariae of L. clavatum from three cnidarians, the moon jellyfish Aurelia aurita (Linnaeus) (s.l.), the Japanese sea nettle Chrysaora pacifica (Goette) and the ghost jellyfish Cyanea nozakii Kishinouye, from the Seto Inland Sea, Japan. They also reported metacercariae and juveniles of L. clavatum from juvenile Pacific rudderfish Psenopsis anomala (Temminck & Schlegel) (Centrolophidae) and Thamnaconus modestus. The first intermediate hosts are unknown.

Overview of new findings
In the present study we examined new specimens consistent with the concept of Lepotrema in the possession (especially) of a distinct folded muscular bulb on the distal metraterm and a postero-dorsal excretory pore. These were from 29 host/parasite/ locality combinations. These forms are superficially highly similar to each other. The specimens were therefore assessed iteratively by morphology and analysis of ITS2 rDNA and cox1 mtDNA sequences for as many combinations for which suitable specimens were available. As discussed in greater detail below, we found that for host/locality combinations for which there were multiple specimens and multiple sequences, there was a strong tendency for genetic and morphological distinctions to be detectable. Using this rationale, we here characterise nine species on the basis of combined morphological and molecular data; of these seven are described as new. One further species is described as new on the basis of morphological data only. In addition, six existing species for which no molecular data are available are recognised. Finally, we summarise reports, old and new, of five host/parasite combinations which may well comprise further new species but for which the evidence is presently inadequate.

Molecular data
We generated 32 5.8S-ITS2-28S rDNA sequences and 31 partial cox1 mtDNA sequences for 15 host/locality combinations. Eleven genotypes/clades were present in both the ITS2 and cox1 datasets, several of which exhibited some low-level intra-genotypic variation. The complete ITS2 region ranged between 275-293 bp in length (as calculated by the ITS2 Database). The ITS2 alignment (including flanking 5.8S and 28S regions) comprised 11 genotypes (each represented by 1-6 replicates) and consisted of 466 bp. The level of distinction between ITS2 genotypes in the final dataset ranged from 1-9 bp (see Table 1); a phylogram representing these differences is shown in Fig. 1A. All cox1 sequences were 475 bp long and the final alignment contained no indels. The number of base differences between species ranged between 19-84 bp, and within a species between 0-14 bp (Table 1). A phylogram representing these differences is shown in Fig. 2. A striking aspect of this dataset is that genotypes in the ITS2 dataset differed by very few bases; several of the genotypes (ultimately considered to relate to different species) differed by just one base. However, in the cox1 dataset the same clades were represented with much greater levels of difference between them. Most of the ITS2 genotypes and cox1 clades related to single host/locality combinations. ABGD analysis of the cox1 dataset suggested the presence of 9-11 species. The initial partition identified nine groups; these nine groups matched the species recognised by morphology, except for the grouping of specimens from Palau, French Polynesia and Moreton Bay as a single unit. The recursive partition identified 11 groups; these groups matched the species recognised by morphology, except for the division of samples from Palau and French Polynesia. 28S rDNA sequence data were generated for all 11 genotypes and were 1,343-1,344 bp long; the final dataset (including the only Lepotrema sequence data available on GenBank and outgroup taxa) was 1,339 bp long. The level of differences between species ranged between 0-23 bp; two genotypes that had no bases different in the final dataset (those from P. wardi and those from Cirripectes spp.) differed by a single indel only. A phylogram representing analyses of the 28S dataset is shown in Fig. 1B.

Remarks
Hitherto, this, the type-species of Lepotrema, was known mainly from its original description. We reproduce here one new illustration (Fig. 3), a copy of the original illustration (Fig. 4) and give measurements for three specimens from the type-host, probably from the type-series. Molecular and morphometric results presented in this paper provide substantial doubt to the other records of this species (especially those from non-tetraodontiforms), given the overall pattern of oioxenous or stenoxenous hostspecificity recognised here. Although we do not have molecular data for this species from its type-host and locality, despite the examination of nine specimens of the type-host from off Minabe, Wakayama Prefecture, Japan, we can be confident that some subsequent reports of this species were mistaken. The clearest evidence for this comes from the forms from Acanthochromis polyacanthus, Amanses scopas and Melichthys vidua from the Great Barrier Reef Barker et al., 1994;Bray & Cribb, 1996c, 2002Bray et al., 2009b), originally identified  Ozaki, 1932; 3, Ventral view of voucher in probable type-series; 4, Ventral view, redrawn from Ozaki (1932); 5-7, Lepotrema adlardi ; 5, ex Abudefduf bengalensis, Ningaloo Reef, ventral view; 6, ex Abudefduf bengalensis, off Heron Island, ventral view; 7, ex Abudefduf bengalensis, off Lizard Island, ventral view. Scalebars: 200 lm as L. clavatum, but which our molecular results indicate are separate species, differing by 98-100 bp in the partial cox1 dataset; clearly they could not relate to the true L. clavatum and it is our view that they do not represent L. clavatum and are distinct. These records, along with the record from Parma polylepis (see Lepotrema sp. 4 below), have been deleted from the list of hosts for this species. We base our ideas on the morphology of this worm on the original description (Ozaki, 1932) and our observations of ten worms, including three ovigerous worms from the type-host, probably the type series. A prepharynx was not described or illustrated by Ozaki (1932), but our observations indicate that a distinct prepharynx is present. The distinctive characters of this species include its large size, large oral sucker, and the relatively short pre-bifurcal and pre-vitelline distances ( Table 2). In our view, it is highly probable that several more of the reports of L. clavatum will prove spurious, but more work is necessary to explore this.

Remarks
This species is morphologically and genetically distinct, being narrow, with an even narrower, long forebody, a long prepharynx and a pre-vitelline distance similar to the forebody length (Figs. 5-7). New measurements are given in Table 3. This is the first report of L. adlardi from the northern Great Barrier Reef (Lizard Island) and the Indian Ocean (Ningaloo Reef). The small samples sizes available (Table 3) give no evidence of morphological variation between the localities. The sequenced specimens come only from the type-locality. The species appears strictly oioxenic to Abudefduf bengalensis. Thirty-two specimens of A. bengalensis have been examined in Moreton Bay, but this species has never been recovered from there. It has also never been found in the banded sergeant A. septemfasciatus (Cuvier) (8 specimens examined, from Lizard Island and French Polynesia), the scissortail sergeant Abudefduf sexfasciatus (Lacépède) (65 specimens from many localities), the blackspot sergeant A. sordidus (Forsskål) (6 specimens from various localities), the Indo-Pacific sergeant A. vaigiensis (Quoy & Gaimard) (12 specimens from the GBR and Moreton Bay) or Whitley's sergeant A. whitleyi Allen & Robertson (308 specimens from various localities). Forty-one other pomacentrid species have been investigated without the recovery of L. adlardi.
Ovary trilobate, immediately pre-testicular, close or adjacent to ventral sucker. Laurer's canal opens at about level of anterior edge of anterior testis. Seminal receptacle dorsal or dorso-lateral to ovary. Mehlis' gland dorsal to ovary or anterior part of anterior testis.
Uterus intercaecal, mostly pre-testicular, passes ventrally to ovary, overlaps ventral sucker. Eggs tanned, operculate. Metraterm shorter than cirrus-sac, distal extremity with large folded muscular pad. Vitellarium follicular, reaching from posterior edge of pharynx to posterior extremity, fields may be confluent in forebody (as narrow band) and post-testicular region; lateral and ventral to caeca.
Excretory pore dorsal, in anterior post-testicular region; vesicle reaches to testes, not traced further.

Remarks
This species is characterised by molecular means (Table 1) and distinguished from similar congeners by the following morphological characteristics (Table 3). Lepotrema clavatum is larger, with a relatively longer forebody, relatively larger oral sucker and pharynx, a longer cirrus-sac, a shorter post-testicular region, longer ventral sucker to bifurcal distance and ventral sucker to ovary distances, a slightly shorter previtelline distance, a smaller sucker ratio and slightly longer caeca. Lepotrema incisum has deeply incised testes, a relatively larger pharynx, a longer previtelline distance and cirrus-sac, a relatively shorter post-testicular region and smaller eggs. Lepotrema monile lacks a strong muscular pad around the metraterm, has a small sphincter and has a relatively shorter cirrus-sac.
The two worms sequenced from off Heron Island have identical ITS2 and cox1 sequence data. 28S sequence data for the specimens from off Heron Island differed from samples from off Lizard Island (Gen-Bank: FJ788483.1) by 2 bp; this is a level greater than between conspecific Lepotrema samples infecting Hemitaurichthys polylepis from Palau and French Polynesia, which had identical 28S data across these regions. In addition, specimens of L. acanthochromidis n. sp. from off Heron Island tend to have slightly longer eggs than those from off Lizard Island. However, given the host and generally similar morphology and the lack of ITS2 and cox1 data for samples from off Lizard Island, we take a conservative approach and recognise both sets of samples as the same species. This anomaly is worthy of further study.
This species has strongly oioxenous specificity for Acanthochromis polyacanthus, having been found in that species at least 40 times on the GBR but never in 1,228 individuals of 55 other pomacentrid species Figs. 8-9 Lepotrema acanthochromidis n. sp. 8, ex Acanthochromis polyacanthus, off Heron Island, holotype, ventral view; 9, ex Acanthochromis polyacanthus, off Lizard Island, ventral view. Scale-bars: 200 lm examined on the GBR. Acanthochromis polyacanthus can be found together with Abudefduf bengalensis (the host of L. adlardi) and Pomacentrus wardi (the host of L. monile, see below) but there is no evidence of any sharing of the three Lepotrema species by these three pomacentrid species. Notably, according to Cribb et al. (1994), A. polyacanthus is also the only (but frequently infected) pomacentrid host for a bivesiculid, Bivesicula unexpecta .
Lepotrema hemitaurichthydis n. sp. Description  [Based 18 whole-mounted specimens, 17 ex H. polylepis, 13 from off Palau, 3 from off Tubuai and 1 from off Rimatara, and 1 ex H. thompsoni; measurements in Table 4.] Body oval or slightly pyriform, slightly wider in hindbody. Tegument finely spined, spines reaching to, or close to, posterior extremity. Oral sucker large, transversely oval, subterminal. Ventral sucker rounded to oval, usually smaller than oral sucker, pre-equatorial. Prepharynx absent to short. Pharynx large, oval to subglobular. Oesophagus short, narrow. Intestinal bifurcation in posterior forebody. Caeca broad, reach to about middle of post-testicular region or beyond.
Ovary trilobate, immediately pre-testicular, adjacent or close to ventral sucker. Laurer's canal opening dorsally at, or close to, sinistral edge of anterior testis. Seminal receptacle dorsal or dorso-lateral to ovary or overlapping anterior testis. Mehlis' gland dorsal to ovary. Uterus intercaecal, mainly pre-testicular, passes ventrally to ovary, overlaps ventral sucker. Eggs tanned, operculate. Metraterm shorter than cirrus-sac, distal extremity with large folded muscular pad. Vitellarium follicular, reaching from pharynx to just into post-testicular region or close to posterior extremity, fields confluent or nearly so in forebody and post-testicular region; lateral and ventral to caeca.
Excretory pore dorsal, in mid post-testicular region; vesicle reaches to testes, not traced further.

Remarks
This species is characterised by molecular means and distinguished from similar congeners by the following morphological characteristics (Table 5). Molecular results are available only for worms from Palau. The Heron Island worms are treated as the same species here, but some morphometric differences are evident and discussed. Lepotrema clavatum has a relatively longer forebody, a less distinct preoral lobe, longer prepharynx and oesophagus (going by the ''typeseries''), a longer intestinal bifurcation to ventral sucker distance, a larger anterior testis, and relatively smaller pharynx and post-testicular and post-caecal distances. Lepotrema acanthochromidis n. sp. has a less distinctive pre-oral lobe, a relatively smaller pharynx, a longer prepharynx, a relatively longer oesophagus and pre-vitelline distance, a relatively shorter cirrus-sac and the gonads tend to be slightly larger. Lepotrema hemitaurichthydis n. sp. has a less distinct pre-oral lobe, a relatively longer oesophagus, a relatively shorter pre-vitelline distance, a longer ventral sucker to ovary distance, the gonads are slightly larger, the post-testicular distance is shorter and the caeca are longer. Lepotrema incisum has deeply incised testes. Lepotrema monile has a small sphincter rather than a strong muscular pad around metraterm.
The Palau and Heron Island specimens differ slightly, with the Palau specimens having a relatively longer pre-oral lobe, shorter oesophagus, longer ventral sucker to ovary distance and smaller gonads. These differences are not of the magnitudes that distinguish this form from other species, so they are considered intraspecific variation here.
We have examined only a single specimen of M. vidua at each of Palau and Heron Island, both being infected with multiple specimens of L. melichthydis n. sp. Four individuals of M. vidua examined in French Polynesia were not infected. In addition, we have examined 344 specimens of 13 species of Balistidae in the Indo-West Pacific region without finding infections of this species. We thus infer that it is oioxenous to M. vidua, or potentially stenoxenous to Melichthys, which has just three recognised species. Pritchard (1963) and Yamaguti (1970) reported L. clavatum from M. vidua from Hawaiian waters. Yamaguti (1970) described the worms and illustrated the terminal genital and the ''ovarian complex'', but not the whole worm and Pritchard (1963) did not describe the worm. These records may represent L. melichthydis n. sp., but using the few measurements supplied by Yamaguti (1970) the oesophagus in the Hawaiian form is much longer than found in any of the worms we have studied. The range of sucker-width ratios derived from the measurements given by Yamaguti (1970) is very large (1:0.80-1.33), whereas in all our specimens the ventral sucker is distinctly smaller than the oral sucker at a ratio of 1:0.76-0.89 There is a clear need for sequencing and morphological study of specimens from a range of Hawaiian fish species.
Ovary strongly or weakly trilobate or more or less globular, immediately pre-testicular, separated from ventral sucker. Laurer's canal opens dorsally, sinistrally to anterior testis, sometimes dorsal to left caecum. Seminal receptacle dorsal or dorso-lateral to ovary. Mehlis' gland dorsal to ovary. Uterus intercaecal, mostly pre-testicular, passes ventrally to ovary, overlaps ventral sucker. Eggs tanned, operculate. Metraterm shorter than cirrus-sac, distal extremity with large folded muscular pad. Vitellarium follicular, follicles sparse, reaching from pharynx to posterior extremity, confluent in post-testicular region; lateral and ventral to caeca.
Excretory pore dorsal, in posterior post-testicular region; vesicle reaches to ovary.

Remarks
This species is characterised by molecular means (Table 1) and distinguished from similar congeners by the following morphological characteristics (Table 5). Lepotrema clavatum is distinctly larger with mainly larger features, a longer prepharynx and oesophagus (cf. ''type-series''), relatively smaller suckers and pharynx, relatively larger testes and relatively shorter cirrus-sac and post-testicular and post-caecal distances. Lepotrema acanthochromidis n. sp. is usually distinctly larger, with a longer pre-oral lobe, relatively slightly smaller suckers and pharynx, a longer oesophagus, a relatively slightly shorter cirrus-sac and larger gonads. Lepotrema hemitaurichthydis n. sp. is usually distinctly larger and broader, with a tendency to have a longer oesophagus, a longer pre-vitelline distance, a relatively slightly shorter cirrus-sac, larger testes and longer caeca. Lepotrema incisum has deeply incised testes. Lepotrema melichthydis n. sp. is distinctly larger, with a distinctly longer pre-oral lobe, a relatively shorter prepharynx, relatively smaller suckers, a relatively shorter cirrus-sac and a relatively longer post-testicular distance. Lepotrema monile has a small sphincter rather than a strong muscular pad around metraterm.
We have examined nine individuals of Amanses scopas on the Great Barrier Reef; five of these were infected with L. amansis n. sp. No infection of this or any other species of Lepotrema has been seen by us in 130 individuals of 28 other species of Monacanthidae examined in the region, apart from the records of L. canthescheniae in the southern Great Barrier Reef/ New South Wales endemic fish Cantheschenia grandisquamis.
Testes 2, small, oval, entire, virtually tandem to slightly oblique, in mid hindbody. Post-testicular region long. External seminal vesicle oval, but usually obscured by eggs. Cirrus-sac claviform, mainly dorsal to ventral sucker. Internal seminal vesicle oval to rounded. Pars prostatica vesicular. Ejaculatory duct long, muscular. Genital atrium distinct. Genital pore sinistral, at level of the anterior margin of the ventral sucker or just anterior.
Excretory pore dorsal, in anterior post-testicular region; vesicle reaches to at least ovary.

Remarks
This species is characterised by molecular means (Table 1) and distinguished from similar congeners by the following morphological characteristics (Table 5). Lepotrema clavatum is larger with mainly larger characters, including the oral sucker, pharynx and testes, it has a longer forebody, a distinct pre-oral lobe (in examined specimens), a shorter pre-vitelline distance, a relatively longer cirrus-sac, and shorter ventral sucker to ovary, post-testicular and post-caecal distances. Lepotrema acanthochromidis n. sp. has a longer pre-oral lobe, a shorter ventral sucker to ovary distance and larger testes. Lepotrema adlardi is narrower, particularly in the forebody, the forebody is longer, the pre-oral lobe is distinct, it has longer prepharynx and pharynx, the intestinal bifurcation is more posterior, the extension of the vitellarium into forebody is much shorter and the ventral sucker to ovary distance is longer. Lepotrema amansis n. sp. has relatively larger suckers and pharynx, a longer cirrussac and shorter ventral sucker to ovary and posttesticular distances. Lepotrema hemitaurichthydis n. sp. is broader, with a longer forebody and pre-oral lobe, slightly larger suckers, a larger pharynx, a shorter ventral sucker to ovary distance, larger testes and a shorter post-testicular distance. Lepotrema incisum has deeply incised testes. Lepotrema melichthydis n. sp. has a distinct pre-oral lobe, a relatively shorter prepharynx and oesophagus, a larger pharynx, shorter pre-vitelline and ventral sucker to ovary distances and a relatively smaller ovary. Lepotrema monile has a small sphincter rather than a strong muscular pad around metraterm.
Lepotrema cirripectis n. sp. has been detected in five of 11 individuals of two species of Cirripectes examined at Lizard Island and in one species of Cirripectes examined at Heron Island. No specimens relating to Lepotrema have been found by us in 240 individuals of 28 species of other blenniid genera from the region, allowing the inference that this species is stenoxenous for the genus Cirripectes. Heron Island specimens are all smaller than the Lizard Island specimens, but with similar proportions. Description (Fig. 22) [Based on 6 whole-mounted specimens; measurements in Table 6.] Body pyriform, widest in hindbody. Tegumental spines reaching close to posterior extremity. Eye-spot pigment scattered around pharynx and oral sucker regions in some, but not all, specimens.
Oral sucker large, subglobular, subterminal. Ventral sucker oval, of similar size to oral sucker, just preequatorial. Prepharynx short, in posterior cavity of oral sucker, thick-walled. Pharynx large, oval. Oesophagus short, narrow. Intestinal bifurcation in posterior forebody. Caeca broad, reach to about middle of post-testicular region.
Testes 2, oval, entire or slightly irregular, symmetrical or slightly oblique, in mid hindbody. External seminal vesicle large, but usually obscured by eggs. Cirrus-sac claviform, mainly dorsal to ventral sucker. Internal seminal vesicle oval. Pars prostatica vesicular. Ejaculatory duct long, thick-walled. Genital atrium distinct. Genital pore sinistral, ventral to sinistral caecum, at about level of anterior margin of the ventral sucker.
Ovary trilobate, immediately pre-testicular, close to ventral sucker. Laurer's canal not detected. Seminal receptacle dorsal or dorso-lateral to ovary. Mehlis' gland dorsal to ovary. Uterus intercaecal, pre-testicular, passes ventrally to ovary, overlaps posterior edge of ventral sucker. Eggs tanned, operculate. Metraterm shorter than cirrus-sac, distal extremity with large folded muscular pad. Vitellarium follicular, reaching from anterior part of pharynx to posterior extremity, confluent in forebody and post-testicular region; lateral and ventral, but not dorsal to caeca.
Excretory pore dorsal, in anterior post-testicular region; vesicle reaches to testes, not traced further.

Remarks
This species is not characterised by molecular means but is distinguished from similar congeners by the following morphological characteristics (Table 6).
It is the only species with more or less symmetrical testes, and, probably as a result, it tends to be broader than the other species. Lepotrema clavatum is a larger worm but is relatively slightly narrower, it has a longer forebody, a less distinct pre-oral lobe, a longer prepharynx (cf. ''type series''), a relatively smaller pharynx, a relatively longer oesophagus (cf. ''type series''), a more anterior intestinal bifurcation and vitelline extent, a relatively smaller ventral sucker, a longer ventral sucker to ovary distance, a shorter posttesticular distance and longer caeca. Lepotrema acanthochromidis n. sp. has smaller suckers, a longer prepharynx, a smaller pharynx, a longer oesophagus, a more anterior intestinal bifurcation, the ovary is  Bray & Cribb, 1996c). Scale-bars: 200 lm  (779) Pre-oral lobe 17-39 (26) 10-28 (20) 10-20 (14) Post-testicular distance 129-237 (185) 148-233 (187) 102-197 (165) 115 213-232 (227) separated from the ventral sucker and the gonads are smaller. Lepotrema amansis n. sp. is distinctly narrower, with a shorter pre-oral lobe, a longer prepharynx, a relatively longer cirrus-sac, a longer ventral sucker to ovary distance, relatively distinctly smaller gonads and a shorter post-testicular distance. Lepotrema cirripectis n. sp. is narrower, with a shorter pre-oral lobe, a longer prepharynx, a shorter pharynx, longer intestinal bifurcation to ventral sucker and ventral sucker to ovary distances, smaller testes and longer caeca. Lepotrema hemitaurichthydis n. sp. has a shorter pre-oral lobe, smaller suckers, longer prepharynx and cirrus-sac, the ovary separated from the ventral sucker, the gonads are smaller, the posttesticular region is shorter, and the caeca are longer. Lepotrema incisum has deeply incised testes. Lepotrema melichthydis n. sp. has a slightly longer preoral lobe, slightly smaller suckers, a shorter previtelline distance, a longer ventral sucker to ovary distance and relatively slightly smaller gonads. Lepotrema monile has a small sphincter rather than a strong muscular pad around metraterm. Lepotrema justinei n. sp. has been found only in S. fraenatum from off New Caledonia, where three of 14 fish examined were infected. We have not seen this species in eight S. fraenatum examined from off Heron Island or in seven examined at Ningaloo Reef. In addition, examination of a further 167 individuals of S. bursa (Bloch & Schneider, 1801) and S. chrysopterum from the Indo-West Pacific region (together with many other balistids) have not been infected with this species. The single specimen reported as Lepotrema clavatum from S. chrysopterum by Bray & Cribb (1996) from off Heron Island is clearly distinct (see below as Lepotrema sp. 3).
Ovary trilobate, immediately pre-testicular, separated from ventral sucker. Laurer's canal opens dorsally to anterior part of anterior testis, not easily detected. Seminal receptacle dorsal or dorso-lateral to ovary. Mehlis' gland dorsal to ovary. Uterus intercaecal, pre-testicular, passes ventrally to ovary. Eggs tanned, operculate. Metraterm shorter than cirrus-sac, distal extremity with large folded muscular pad. Vitellarium follicular, reaching from level of posterior part of pharynx, oesophagus or intestinal bifurcation to posterior extremity, confluent in forebody and posttesticular region; lateral and ventral, but not dorsal to caeca.
Excretory pore dorsal, in anterior post-testicular region or at level of caecal ends; vesicle reaches to testes, not traced further.

Remarks
This species is characterised by molecular means (Table 1) and distinguished from similar congeners by the following morphological characteristics (Table 6). Lepotrema clavatum is larger with a slightly longer forebody, a shorter pre-oral lobe, longer prepharynx (cf. ''type-series'') and oesophagus, a shorter previtelline distance, a longer cirrus-sac, a relatively shorter ventral sucker to ovary distance, slightly larger testes, a shorter post-testicular distance and longer caeca. Lepotrema acanthochromidis n. sp. is very similar but the prepharynx tends to be longer, and there is a tendency for a slightly shorter pharynx and ventral sucker to ovary distance. Lepotrema adlardi is slightly narrower, with a distinctly narrower forebody, a smaller oral sucker, a longer prepharynx, a shorter intestinal bifurcation to ventral sucker distance, a much longer pre-vitelline distance and a slightly shorter cirrus-sac and ventral sucker to ovary distance. Lepotrema amansis n. sp. is smaller, but relatively slightly narrower, it has a less distinct pre-oral lobe, a slightly shorter pre-vitelline distance, a shorter ventral sucker to ovary distance and a smaller posterior testis. Lepotrema cirripectis n. sp. has a less distinct pre-oral lobe, slightly smaller suckers, a longer prepharynx, a smaller pharynx, the ventral sucker to ovary distance tends to be longer, the testes tend to be smaller as does the cirrus-sac but the post-testicular distance may be slightly larger. Lepotrema hemitaurichthydis n. sp. is very similar, but is wider, with possibly a slightly longer prepharynx, possibly a slightly longer oesophagus, a longer ventral sucker to ovary distance and slightly longer caeca. Lepotrema justinei n. sp. is broader, with a relatively slightly larger ventral sucker, a shorter pre-bifurcal distance, a distinct separation of ventral sucker and ovary, symmetrical testes and smaller gonads. Lepotrema melichthydis n. sp. has a more prominent pre-oral lobe, a shorter previtelline distance and a slight tendency for the ventral sucker to ovary distance to be smaller.
Numerous specimens of Lepotrema moretonense n. sp. have been found in all seven P. microlepidotus and both P. maculatus examined from Moreton Bay. In addition, a single gravid adult and a single immature specimen (that was sequenced for the ITS2 and cox1 datasets) have been found in two of the 36 individuals of Selenotoca multifasciata (Scatophagidae) examined from Moreton Bay. We think that these data suggest that this species is effectively stenoxenous to the genus Prionurus. We have only examined a single uninfected individual of a Prionurus species elsewhere and so cannot comment on the geographical distribution of this species.
Excretory pore dorsal, in mid post-testicular region; vesicle reaches to mid-region of anterior testis.

Remarks
This species is characterised by molecular means (Table 1) and distinguished from similar congeners by the following morphological characteristics (Table 7).
Lepotrema clavatum is larger, with a slightly longer forebody, a longer pre-oral lobe, a larger oral sucker, a shorter oesophagus, a longer intestinal bifurcation to ventral sucker distance, a much longer cirrus-sac, larger gonads, a shorter post-testicular distance and longer caeca. Lepotrema acanthochromidis n. sp. has a longer pre-oral lobe, the oesophagus tends to be longer, the intestinal bifurcation is further from the ventral sucker, the cirrus-sac is longer and post-testicular region is slightly shorter. Lepotrema adlardi is slightly narrower, particularly in the forebody, with a longer pre-oral lobe and prepharynx, a much longer pre-vitelline distance and a longer cirrus-sac and caeca. Lepotrema amansis n. sp. never gets as large, usually has a less distinct pre-oral lobe, wider suckers, a slightly longer pharynx, a longer intestinal bifurcation to ventral sucker distance, a slightly shorter pre-vitelline distance, a longer cirrussac and a shorter post-testicular distance. Lepotrema cirripectis n. sp. has a slightly shorter prepharynx and oesophagus and a smaller pharynx, the intestinal bifurcation is further from the ventral sucker and has a longer cirrus-sac, ventral sucker to ovary distance and caeca. Lepotrema hemitaurichthydis n. sp. has a slightly longer forebody, the intestinal bifurcation is further from the ventral sucker, the cirrus-sac is longer, the testes are larger, post-testicular region is shorter and the caeca are longer. Lepotrema justinei n. sp. is wider, with a longer pre-oral lobe, relatively slightly larger suckers, an indistinct prepharynx, a longer pharynx, a shorter oesophagus, a longer cirrus-sac, no distinct separation of ventral sucker and ovary and larger gonads. Lepotrema melichthydis n. sp. has a more distinct pre-oral lobe, a shorter prepharynx, a longer pharynx, a shorter oesophagus, the intestinal bifurcation is further from the ventral sucker, the pre-vitelline distance shorter and the cirrussac and caeca are longer. Lepotrema moretonense n. sp. has a more distinct pre-oral lobe, a shorter prepharynx, a longer intestinal bifurcation to ventral sucker distance, a larger oral sucker and pharynx, a longer cirrus-sac and a shorter post-testicular distance. Bray et al. (1993) described and illustrated this form, based on two specimens. They stated that it was ''similar to L. clavatum in many metrical features and possess a distinct folded muscular metraterm pad. They differ in the short cirrus-sac, just overlapping the ventral sucker, and the straight ejaculatory duct''.
Infections of this form have been found, all as single infections, in five of 41 A. curacao examined from off Heron Island, but in none of 30 examined from off Lizard Island. It has also been found in just one of nine Amphiprion akindynos examined at Heron Island. The species has not been detected in over 1,300 individuals of 40 other pomacentrid species examined in the region.

Remarks
In this species, along with only L. xanthichthydis, the vitellarium reaches to the oral sucker. These worms differ in cirrus-sac length, the lack or presence of uterine coils in the forebody, and egg-length (Table 6; see also Fig. 28).
The original description of L. canthescheniae was based on three specimens from two individual C. grandisquamis from off Heron Island (Bray & Cribb, 1996c). We have since examined nine more individuals of C. grandisquamis (especially to obtain material for sequencing) but all have been uninfected. Total prevalence now stands at two of 42. We suspect that this species is oioxenous to C. grandisquamis, but the low prevalence detected is puzzling. This species is clearly morphologically distinct from the only other species of Lepotrema we have found in a monacanthid, L. amansis n. sp.  Syn. Lepocreadium sp. from Pomacentrus cf. wardi of Bray et al. (1993)  Localities: Off Lizard Island (14°40 0 S, 145°28 0 E) (ex P. amboinensis and P. chrysurus); off Heron Island (ex S. apicalis). Voucher material: Ex P. amboinensis (QM G237567); ex P. chrysurus (QM G237565-6); ex S. apicalis (QM G237568). Prevalence: Ex P. amboinensis (in 1 of 54 fish examined); ex P. chrysurus (in 1 of 26 fish examined); ex S. apicalis (in 2 of 27 fish examined). Representative DNA sequences: ITS2 rDNA, two identical replicates (one submitted to GenBank MH730009); cox1 mtDNA, two identical replicates (both submitted to GenBank MH730048-49); 28S rDNA, one sequence (submitted to GenBank MH730024).

Remarks
This species is distinct in that it lacks a prominent folded muscular pad at the distal metraterm, having a reduced muscular sphincter. It differs genetically from all other Lepotrema species for which molecular data are available. New measurements are given in Table 7; see also  It also differs from other similar species in the following characteristics. Lepotrema clavatum is larger, with a longer forebody, a shorter pre-vitelline distance, a longer cirrus-sac, a shorter ventral sucker to ovary distance, slightly larger gonads, a shorter posttesticular region and longer caeca. Lepotrema adlardi is narrower and mostly larger, with a longer forebody, slightly smaller suckers, a distinctly longer prepharynx, a longer oesophagus and pre-vitelline distance, a shorter ventral sucker to ovary distance and a slightly smaller ovary. Lepotrema acanthochromidis n. sp. has a slightly longer intestinal bifurcation to ventral sucker distance, a longer cirrus-sac and a shorter ventral sucker to ovary distance. Lepotrema amansis n. sp. has a slightly longer forebody, larger suckers and pharynx, a longer intestinal bifurcation to ventral sucker distance, a shorter pre-vitelline distance, a longer cirrus-sac, a shorter ventral sucker to ovary distance and post-testicular region and possibly slightly shorter eggs. Lepotrema amblyglyphidodonis n. sp. has a less distinct pre-oral lobe and a shorter ventral sucker to ovary distance. Lepotrema cirripectis n. sp. has a less distinct pre-oral lobe, a longer intestinal bifurcation to ventral sucker distance and cirrus-sac and possibly slightly longer caeca. Lepotrema hemitaurichthydis n.
sp. has a slightly longer forebody, slightly larger suckers, a longer intestinal bifurcation to ventral sucker distance, a longer cirrus-sac, shorter ventral sucker to ovary distance and post-testicular region, longer caeca and possibly slightly longer eggs. Lepotrema justinei n. sp. is broader, with a distinct pre-oral lobe, a shorter prepharynx, a larger pharynx, shorter pre-vitelline distance and ventral sucker to Figs. 29-35 Lepotrema spp. 29, 30, Lepotrema monile Bray & Cribb, 1998. 29, ex Pomacentrus wardi, off Heron Island, ventral view (redrawn from Bray & Cribb, 1998 ovary distances, larger gonads and shorter caeca. Lepotrema melichthydis n. sp. has a longer pre-oral lobe, a shorter prepharynx, a longer pharynx and intestinal bifurcation to ventral sucker distance, a shorter pre-vitelline distance, a longer cirrus-sac, a shorter ventral sucker to ovary distance, a smaller ovary and possibly slightly shorter eggs. Lepotrema moretonense n. sp. has a shorter prepharynx, a larger pharynx, a longer cirrus-sac, a shorter ventral sucker to ovary distance, a larger posterior testis and a shorter post-testicular region. We have collected L. monile in seven of 61 P. wardi examined at Heron Island . It has only been detected in one of 17 Stegastes apicalis from off Heron Island, but in none of 74 individuals of 12 other species of Pomacentrus examined from off Heron Island. However, it has been found in P. amboinensis (see Sun et al., 2012) and P. chrysurus (once each) from off Lizard Island, in each case the identification is based of morphology only. This species is interpreted as predominantly stenoxenous to the genus Pomacentrus. The low recorded prevalence in two of the species of Pomacentrus suggests that this species might well be found in more species of Pomacentrus given sufficient sampling.

Remarks
One specimen is available. Measurements of the specimen are given in Table 8; see also Fig. 31. It is broadly oval with almost symmetrical testes. In its nearly symmetrical testes this specimen resembles L. justinei n. sp., but the folded muscular pad on the metraterm is far smaller. Dyer et al. (1988) reported L. clavatum from R. aculeatus from off Okinawa, Japan, but without any descriptive matter, meaning that it is not possible to speculate rationally on whether that form is the same as the Lizard Island form, the Palau form (see below) or distinct. We have examined 38 individuals of this fish from off Lizard Island and another 23 from other sites in the region, but apart from Lepotrema sp. 2 (see below), only the single infection has been detected.

Remarks
Two specimens are available (Table 8; Fig. 32). They appear distinctly different from the specimen (Lepotrema sp. 1) reported from the same host at Lizard Island. One of two individuals of R. aculeatus examined from off Palau was infected.

Remarks
Only one specimen is available (Table 8; Fig. 33). We have now examined 67 individuals of S. chrysopterum at Heron Island with only one infection detected. This species is clearly different from Lepotrema justinei n. sp., the other form found in Sufflamen. It is a smaller, much narrow worm with almost tandem testes and the ovary is distinctly separated from the ventral sucker.

Remarks
One Lepotrema specimen was found in 119 specimens of this host, 42 of which were from off Heron Island. This specimen (Fig. 35) has a relatively exceptionally large ventral sucker which appears to distinguish it from all other recognised species of Lepotrema. It also clearly distinct from all the species for which molecular data exists on the basis of ITS2 and cox1 data. In the 28S phylogram (Fig. 1B) this form is sister to Lepotrema amblyglyphidodonis n. sp., but this relationship has poor support.

Remarks
Important differentiating characters include the shape, width and the deeply incised testes (Table 9; Fig. 36). The host-species was quoted by Hanson (1955) as Melichthys buniva, but according to Randall (2007) this is a synonym of an Atlantic species, and is a misidentification of M. niger.

Remarks
Important differentiating characters include the forebody length, the anterior position of the bifurcation and the vitelline extent, the cirrus-sac not reaching into the hindbody, the uterine coil in the forebody and the egg length (46-56) (Yamaguti, 1970) (Table 9; Fig. 37).

Remarks
Although erected in the genus Preptetos, this species is consistent with Lepotrema and we formally propose the new combination here. It is much bigger than any other described species of Lepotrema (see Wang, 1989) (Table 9; Fig. 38). We note that we have examined 23 specimens of Monacanthus chinensis from Moreton Bay without finding this or any other species of Lepotrema.

Remarks
Although erected in the genus Lepocreadium, this species is consistent with Lepotrema and we formally propose the new combination here. It appears to be of an unusual pyriform shape with the narrower part anterior and the cirrus-sac beside the ventral sucker (Table 9; Fig. 39). In the original paper it is compared with L. clavatum and L. xanthichthydis, both considered in Lepocreadium by Shen (1986), but considered Lepotrema here. The cirrus-sac is described as to the right of the ventral sucker, an unusual feature for Lepotrema. It is noteworthy that T. modestus has also been reported as a host of L. clavatum. It seems possible that either L. navodonis could be a synonym of L. clavatum, or that L. clavatum may not really infect T. modestus.

Phylogenetic results
Neighbour-joining phylograms produced for the ITS2 and cox1 datasets demonstrate the complete lack of intraspecific variation in ITS2 rDNA (except for L. hemitaurichthydis) and its frequency among cox1 sequences (6 of 9 species for which at least two sequences were obtained). The topologies from the two analyses were substanially different (Figs. 1A and 2, respectively), with only one relationship shared between the two analyses; L. hemitaurichthydis n. sp. and L. moretonense n. sp. were sister taxa in analyses for both regions. Due to the high level of difference between the two phylograms, interpretation of relationships within the genus are instead based on the partial 28S rDNA dataset; this region has been shown to be more reliable for inference of phylogenetic relationships (Blasco-Costa et al., 2016), and has been used for inferring relationships within the Lepocreadiidae in several studies (Bray et al., 2009b;Bray & Cribb, 2012;Bray et al., 2018). Similar to the ITS2 dataset, the 28S dataset contained few variable base positions, with two clearly distinct species (L. cirripectis n. sp. and L. monile) having identical sequences in the final dataset; they differ by a single indel in the partial 28S sequence alignment. Preliminary analysis that included all available lepocreadoid taxa showed that all Lepotrema species sequenced formed a single well-supported clade. Thus, due to the limited number of differences in the Lepotrema dataset, a reduced analysis relative to three closely related genera was conducted to limit the loss of informative characters. The phylogram produced by Bayesian inference analysis of this dataset is shown in Fig. 1B. Lepotrema as a genus was again wellsupported, as was a clade containing all Lepotrema  (Hanson, 1955) ex Melichthys buniva, Hawaii, ventral view (redrawn from Hanson, 1955); 37, Lepotrema xanthichthydis (Yamaguti, 1970) ex Xanthichthys ringens, Hawaii, ventral view (redrawn from Yamaguti, 1970); 38, Lepotrema cylindricum (Wang, 1989) n. comb. ex Monacanthus chinensis or Navodon septentrionalis, off Fujian, China, ventral view (redrawn from Wang, 1989); 39, Lepotrema navodonis (Shen, 1986) n. comb. ex Thamnoconus modestus, off Zhejiang, China, ventral view (redrawn from Shen, 1986). Scale-bars: 200 lm species to the exclusion of L. melichthydis n. sp. Lepotrema acanthochromidis n. sp., L. hemitaurichthydis n. sp. and L. moretonense n. sp. formed a stongly-supported clade, sister to a clade consisting of L. amblyglyphidodonis n. sp. and Lepotrema sp. 5. These five species formed a strongly-supported clade, sister to the poorly-supported clade of L. adlardi, L. amansis n. sp., L. cirripectis n. sp. and L. monile.
geographical considerations. Most of the species are known from just one site, or perhaps two sites within the GBR. Indeed, Cribb et al. (2016) observed that such limited reporting of species is general for the fish trematode fauna of the Indo-Pacific. A handful of species of Lepotrema have been reported from multiple localities. Specimens of L. cirripectis n. sp. from off Heron and Lizard Islands had identical ITS2 sequences and minor variation in cox1 sequences; this cox1 variation was at a level much lower than between clearly distinct species. Lepotrema adlardi has also been reported from the northern and southern GBR and off Western Australia, although it has only been sequenced from the southern GBR. More interestingly, H. melichthydis n. sp. is here reported from both Palau and the southern GBR, in the same fish species; slight morphological differences are interpreted as intraspecific variation. The most intriguing case of apparent widespread distribution is that of L. hemitaurichthydis n. sp. which is here reported from the off Australs and Marquesas Archipelagos in French Polynesia and from off Palau. In this case one host species is the same, the morphology is similar (although with some possible distinctions), and the ITS2 and cox1 sequences differ by 1 bp (in addition to a single indel) and 14 bp, respectively. Importantly, the intraspecific cox1 variation is greater than that for any 'good' species whereas it is considerably less than the interspecific variation between any combination of recognised species. In the face of this somewhat conflicting information, we propose a conservative approach, interpreting the forms from H. polylepis as a single species that demonstrates geographical genetic variation. This approach (interpreting low-level genetic differences in worms from the same or very similar hosts over geographic range as intra-specific variation) has been adopted for several trematode taxa ] although of these, only the study of the Monorchiidae incorporated both ITS2 and cox1 sequence data. Ultimately, we think that we do not yet know enough about the nature of the distribution of trematodes in the Indo-Pacific to be able to reliably interpret circumstances such as these. Identification of the true host and geographical distribution of the type-species of Lepotrema, L. clavatum, remains important. On the basis of the patterns of host-specificity reported here, we think it unlikely that any species of the Balistidae, Chaetodontidae, Paralichthyidae, Pomacanthidae and Pomacentridae will prove to be typical hosts of this species. However, this does remain to be demonstrated and it should not be considered a fait accompli. As shown by Wee et al. (2017), host specificity of species of a single genus may vary quite dramatically and without evident explanation. A special problem with L. clavatum relating to the ''type-series'' is that these worms were flattened, whereas none of the material we collected was treated in this way. Certainly we advocate that molecular data should be a significant part of the argument when the overall status of L. clavatum receives further attention.

Significance of host specificity
Three families dominate the host records of Lepotrema spp. Using the number of records as an indicator it can be seen that 28% are from the Pomacentridae, 27% from the Balistidae and 23% from the Monacanthidae. Other reef fish families such as the Chaetodontidae (8%), Acanthuridae (3%) and Pomacanthidae (1.7%) are also represented, and the Blenniidae, which occurs frequently on coral reefs, has 5% of records. Records in the Paralichthyidae (1.7%) and the Scatophagidae (1.7%) indicate that the genus is not solely reefassociated. Nevertheless, it is clearly mostly associated with fishes on coral reefs. Two orders predominate with the Tetraodontiformes (50% of records) and the Perciformes (with 48%). The single report from a pleuronectiform appears anomolous and may be accidental, although the worms are described as ovigerous, but relatively small (Yamaguti, 1934).
This host distribution can best be considered in the context of the recent report that metacercariae of Lepotrema clavatum infect medusae of several cnidarian species in Japanese waters (Kondo et al., 2016). These authors further demonstrated that Thamnaconus modestus and Psenopsis anomala, which both live in association with jellyfish as juveniles, had both nematocysts and juvenile L. clavatum in their guts. In combination these data establish convincing evidence for at least one mode of transmission of this species. In this context it is noteworthy that Miyajima et al. (2011) demonstrated that S. cirrhifer, the typehost of L. clavatum, will feed willingly on medusae in captivity. Unfortunately, it is not obvious that feeding on medusae explains the distribution of other Lepotrema species. Although it is generally acknowledged that reports of ingestion of medusae by fishes are inadequate, of the fish reported as doing so in the reviews of Arai (2005) and Ates (1988), just two, Melichthys niger and Xanthichthys ringens, are known hosts for species of Lepotrema. For some of the remaining fish reported as hosts here, it seems plausible that medusae may be featured in their diet. Species of Abudefduf, Amblyglyphidodon, Hemitaurichthys and Prionurus all often feed in mid-water, although it is unclear whether medusae are, in fact, an important part of their diet. In contrast, species of Amanses, Cantheschenia, Cirripectes, Pomacentrus, Rhinecanthus and Sufflamen are typically demersal feeders that would not be expected to feed on medusae consistently. In this context it is noteworthy that Amanses scopas is an obligate coral feeder (Bacchet et al., 2006); Ward (pers. comm.) has identified fragments of the corals Acropora cytherea (Dana) and Isopora sp. from the intestine of specimens of A. scopas collected from off Heron Island. It seems possible, therefore, that species of this genus may have a predilection for cnidarians, both polyps and medusae, as second intermediate hosts. Regardless of the range of second intermediate hosts for species of this genus, the apparent randomness of distribution among apparently comparable fishes remains baffling. Lepotrema cylindricum (Wang, 1989) n. comb.