Zootaxa 1992: 37–52 (2009) ISSN 1175-5326 (print edition) www.mapress.com/zootaxa/ Article ZOOTAXA Copyright © 2009 · Magnolia Press ISSN1175-5334(online edition) Morphological disparity despite genetic similarity; new species of Lobosorchis Miller & Cribb, 2005 (Digenea: Cryptogonimidae) from the Great Barrier Reef and the Maldives TERRENCE L. MILLER1, ABIGAIL J. DOWNIE1 & THOMAS H. CRIBB1,2 1Parasitology Section, School of Molecular and Microbial Sciences, The University of Queensland, Brisbane, Queensland 4072, Aus- tralia. E-mail: [email protected]. E-mail: [email protected]. 2Centre for Marine Studies, The University of Queensland, Brisbane, Queensland 4072, Australia. E-mail: [email protected] Abstract Examination of the humpback red snapper, Lutjanus gibbus (Perciformes: Lutjanidae), from Lizard Island off the Great Barrier Reef and Rasdhoo Atoll, Maldives revealed the presence of a new species of Lobosorchis (Digenea: Cryp- togonimidae), L. polygongylus n. sp.Lobosorchis polygongylusn. sp. is distinguished from the type- and only other spe- cies, L. tibaldiae by the combination of body size, oral spine number (60–81 in L. polygongylus, 47–56 in L. tibaldiae) and number of testes (13–25 in L. polygongylus, 9 in L. tibaldiae). Bayesian inference analysis using data from the inter- nal transcribed spacers 1 and 2 (ITS1 and ITS2), 5.8S and the large subunit (LSU) nuclear ribosomal DNA of L. polygongylus,L. tibaldiae and species of Beluesca,Caulanus,Chelediadema,Neometadena,Latuterus,Retrovarium and Siphoderina was performed to explore phylogenetic relationships of species of Lobosorchis with other cryptogonimid taxa. Despite the significant morphological differences between Lobosorchis polygongylus and L. tibaldiae, these two species differed consistently by only 5 base pairs (bp) over the entire ITS region (3 bp in ITS1, 0 bp in 5.8S and 2 bp in ITS2) and 1 bp in the LSU rDNA regions examined. The ITS2 rDNA region was sequenced from metacercariae obtained from the fins, flesh or body cavities of a number of fishes belonging to the Blenniidae, Pomacentridae and Tetraodonti- dae and analysed using minimum evolution analysis with L. polygongylus and L. tibaldiae. This revealed the presence of two additional genotypes (putative Lobosorchis sp. A and B), which consistently differed from L. polygongylus by 1 and 4 bp, L. tibaldiae by 1 and 4 bp and from each other by 3 bp over the ITS2 dataset. Although these genetic differences are relatively small, when evaluated in light of the differences observed between L. polygongylus and L. tibaldiae (which are morphologically quite distinct) and differences seen in other congeneric cryptogonimid taxa, the ITS2 rDNA data alone suggest that at least two more species of Lobosorchis are present at Lizard Island. These data also suggest that the ITS2 rDNA region alone is suitable for resolving operational taxonomic units (OTUs) at the species level within the Cryp- togonimidae based on what was observed in this and other cryptogonimid systems. A morphological description of meta- cercariae of L. tibaldiae obtained from two species of Pomacentridae at Heron Island, off the Great Barrier Reef is also provided. Key words: Biodiversity; Blenniidae; Cryptogonimidae; Digenea; Great Barrier Reef; Lizard Island; Lobosorchis; Lut- janidae; Lutjanus gibbus; Lutjanus; Maldives; Opisthorchiata; Opisthorchioidea; Pomacentridae; ribosomal DNA; Tet- raodontidae Introduction The humpback red snapper, Lutjanus gibbus (Forsskål) (Lutjanidae), is a relatively common inhabitant of Indo-West Pacific coral reef ecosystems from the Red Sea and East Africa to the Line and Society islands, north to southern Japan and south to Australia (Allen 1985). The diet of this long-lived species includes a vari- ety of fishes, benthic crustaceans, echinoderms and mollusks, similar to its sister taxon, the two-spot red snap- per, L. bohar (Forsskål) (Miller & Cribb 2007a; Froese & Pauly 2008). Lutjanus gibbus is known to harbour a Accepted by N. Dronen: 11 Nov. 2008; published: 2 Feb. 2009 37 number of digenetic trematode species. An unidentified species of Stephanostomum Looss, 1899 (Acanthoc- olpidae Lühe, 1906) was reported from L. gibbus off Kuwait by Al-Yamani and Nahhas (1981). Parukhin (1971) reported the faustulid Bacciger bacciger (Rudolphi, 1819) Nicoll, 1914 from the Red Sea. Unidenti- fied sclerodistomid (Prosogonotrema Vigueras, 1940 sp.) and lepocreadiid (Pseudocreadium Layman, 1930 sp.) species were reported from the Persian Gulf and Gulf of Oman by El-Naffar et al. (1992). Rigby et al. (1999) reported an unidentified species of Hemiuridae Looss, 1899 from French Polynesia. A few species of Opecoelidae Ozaki, 1925 have also been reported from L. gibbus: Hamacreadium lethrini Yamaguti, 1934 was reported by Fischthal and Kuntz (1964) off the Philippines; H. mutabile Linton, 1910 was reported from the Red Sea by Parukhin (1970), French Polynesia by Rigby et al. (1999) and the Great Barrier Reef by Bray and Cribb (1989); an unidentified species of Helicometra Odhner, 1902 was reported off French Polynesia by Rigby et al. (1999). The only cryptogonimid known from L. gibbus is Retrovarium sphericum (Nahhas, Sey & Nishimoto, 1998) Miller & Cribb, 2007, which was reported off Kuwait by Nahhas et al. (1998). The monotypic Lobosorchis Miller & Cribb, 2005 (Digenea: Cryptogonimidae) was proposed by Miller and Cribb (2005) for L. tibaldiae Miller & Cribb, 2005 recovered from the Spanish flag snapper, Lutjanus carponotatus (Richardson), and the dory snapper, L. fulviflamma (Forsskål), off Heron and Lizard Islands on the Great Barrier Reef and New Caledonia and distinguished from all other cryptogonimid genera based on its combination of having oral spines, multiple testes in two symmetrical groups, and vitelline follicles that are distributed mainly in the forebody. A recent survey of the digenetic trematode fauna inhabiting Indo-West Pacific lutjanids revealed the presence of a new species of Lobosorchis from Lutjanus gibbus off Rasdhoo Atoll in the Maldives and Lizard Island off the Great Barrier Reef, Australia. Here we describe this new spe- cies and provide an amended diagnosis of Lobosorchis. We also compare morphometric and molecular data obtained from the internal transcribed spacer (includes ITS1, 5.8S and ITS2) and the large subunit (LSU) ribo- somal DNA (rDNA) regions of this new taxon with the type-species L. tibaldiae and putative Lobosorchis metacercariae recovered from three species of Blenniidae, three Pomacentridae spp., and one species of Tet- raodontidae off the Great Barrier Reef. Bayesian Inference analysis of a combined ITS and LSU dataset was performed to explore the phylogenetic relationships of species of Lobosorchis with other recently reported cryptogonimid taxa (Miller & Cribb 2007b; c; d). Minimum evolution analysis was also conducted on the ITS2 dataset (which included data from metacercariae obtained from intermediate hosts at Lizard Island) to attempt to elucidate the life-cycles of and explore the diversity of Lobosorchis species on the Great Barrier Reef (GBR). Material and methods Host and parasite collection Fish were collected using baited line, clove oil, seine or spear from the following localities: Heron Island (23°26’S; 151°54’E) in the southern GBR, Lizard Island (14°40’S; 145°27E) in the northern GBR and Rasd- hoo Atoll (4°16’N; 72°58’E), Maldives. Total numbers of species of Lutjanidae collected during this study are listed in Miller and Cribb (2007b). Fish were killed by neural pithing and the intestine immediately removed, washed in vertebrate saline (0.85%), and examined for the presence of endohelminths. Adult trematodes obtained from the intestine or pyloric caeca were washed in saline and killed by pipetting them into nearly boiling saline. Specimens for morphological analysis were then stored in 10% formalin and specimens for DNA extraction and analysis were stored in 95–100% ethanol at -20°C. Metacercariae were obtained from the musculature and fins of fish examined using fine-tipped forceps. When possible, metacercariae were gently excised from their cysts using forceps for putative identification using a dissecting microscope, then heat fixed in nearly boiling saline and stored in 10% formalin for morphological analysis or 95–100% ethanol at - 20°C for molecular analysis. If metacercariae were damaged or excision unsuccessful, they were immediately stored in ethanol for molecular analysis. 38 · Zootaxa 1992 © 2009 Magnolia Press MILLER ET AL. Morphological and molecular sample processing Specimens for morphological and molecular analysis (DNA extraction, amplification of the entire internal transcribed spacer (ITS1, 5.8S and ITS2) and large subunit (LSU) nuclear ribosomal DNA regions, and sequencing) were processed according to the protocols reported by Miller and Cribb (2007b; d).The ITS and LSU regions were sequenced for the type-species of Lobosorchis,L. tibaldiae, from the type-host and locality and submitted to GenBank under the following accession numbers: ITS (FJ154899), LSU (FJ154901). Gen- Bank accession numbers for the remaining species of Lobosorchis sequenced in this study are provided under the appropriate specimens in the Results section. The consensus sequences for each taxon utilized in this study were constructed from multiple replicates (each replicate contig being both a forward and reverse sequence from a single individual). Additional replicate contigs were sequenced from specimens obtained from differ- ent host/parasite/location combinations whenever possible. All morphometric measurements were made using an ocular micrometer and are in micrometers, with the mean followed by range in parentheses. Length-width or sucker-width ratios in the diagnoses refer to the length divided by the width or oral sucker width divided by the ventral sucker width expressed as a whole number. Type- and voucher specimens were deposited in the Queensland Museum, Brisbane, Australia. Comparative DNA analyses Three rDNA datasets were analysed independently in this study. These included the entire ITS (ITS1, 5.8S and ITS2), ITS2 only (because only ITS2 data was obtained for some metacercariae included here) and LSU regions. The three rDNA region datasets from taxa sequenced in this study were initially aligned using ClustalX version 2.0.9 (Larkin et al. 2007) under the following parameters: pairwise alignment parameters = gap opening 10.00, gap extension 0.10, DNA weight matrix International Union of Biochemistry (IUB); mul- tiple alignment parameters = gap opening 10.00, gap extension 0.20, delay divergent sequences 30%, DNA weight matrix IUB. The resulting sequence alignments were exported from ClustalX in FASTA and NEXUS formats, and refined by eye using MacClade version 4.08 (Maddison & Maddison 2005). After alignments of the rDNA regions were edited, the ends of each fragment were trimmed to match the shortest sequence in the alignment. Distance matrices for the three rDNA region datasets (entire ITS, ITS2 and LSU) were constructed with the absolute pairwise character difference and the percentage of uncorrected “p” pairwise character dif- ferences. Pairwise comparisons of absolute sequence divergence for all taxa were calculated with gaps treated as missing data. The entire ITS, ITS2 only and LSU rDNA regions for species of Lobosorchis were also aligned with those reported for species of the cryptogonimid genera Beluesca Miller & Cribb, 2007, Caulanus Miller & Cribb, 2007, Chelediadema Miller & Cribb, 2007, Latuterus Miller & Cribb, 2007, Neometadena Hafeezullah & Siddiqi, 1970, and Retrovarium Miller & Cribb, 2007by Miller and Cribb (2007b; c; d) for comparative pur- poses and to explore levels of interspecific variation. The ITS and LSU rDNA regions were then combined and assigned partitions in a single NEXUS file. Minimum evolution and Bayesian inference analyses of the combined (entire ITS and LSU) dataset of these taxa were performed using PAUP* version 4.0b10 (Swofford 2003) and MrBayes version 3.1.2 (Ronquist & Huelsenbeck 2003). Modeltest version 3.7 (Posada & Crandall 1998) was used to estimate the best substitution model for the combined dataset. Bayesian inference analysis was conducted on the combined dataset using the GTR+I+G model predicted as the best estimator by the Akaike Information Criterion (AIC) in Modeltest. The ITS and LSU regions were partitioned in the combined dataset to allow estimates of each model parameter for each rDNA region. Bayesian inference analysis was run over 1,000,000 generations (ngen=1000000) via four simultaneous Markov Chain Monte Carlo (MCMC) chains (nchains=4) and every 100th tree saved (samplefreq=100). Bayesian analyses used the following parameters: nst=6, rates=invgamma, ngammacat=4, and the MCMC parameters were left at the default set- tings, and the priors parameters of the combined dataset were set to ratepr=variable. Samples of substitution model parameters, and tree and branch lengths were summarized using the parameters ‘sump burnin=3000’, NEW SPECIES OF LOBOSORCHIS (CRYPTOGONIMIDAE) Zootaxa 1992 © 2009 Magnolia Press · 39 ‘sumt burnin=3000’. These ‘burnin’ parameters were chosen because the log likelihood scores ‘plateaued’ well before 300,000 replicates in the BI analyses. Minimum evolution analysis was also conducted for comparative purposes on the ITS2 only dataset inde- pendently (which included putative Lobosorchis spp. sequences obtained from metacercariae off the GBR) using PAUP*. Nodal support for minimum evolution analyses of the combined (ITS and LSU) and ITS2 only datasets were inferred by bootstrap analysis using a heuristic search of 10,000 replicates. Results Morphological data Lobosorchis Miller & Cribb, 2005 Type-species:L. tibaldiae Miller & Cribb, 2005. Diagnosis: Body oval; length/width ratio c. 1.4–1.9. Tegument armed with small to minute spines. Oral sucker distinctly wider than long, with enlarged oral spines, opens nearly terminally. Ventral sucker unspecia- lised, embedded in ventrogenital sac. Ratio oral/ventral sucker width c. 2.5–3. Forebody occupies c. 30–40% of body length. Prepharynx short. Pharynx wider than ventral sucker. Oesophagus short. Intestinal bifurcation between ventral sucker and pharynx. Caeca blind, terminate close to posterior end of body. Testes multiple, in two symmetrical groups of 4–5 or more per group, total number of testes 9–25, in mid-hindbody. Seminal ves- icle tubulosaccular, between ovary and ventral sucker. Common genital pore immediately anterior to ventral sucker. Gonotyl a small lobe arising from anterior wall of ventrogenital sac. Ovary deeply lobed, ventral to or immediately anterior and adjacent to testes. Laurer's canal present. Seminal receptacle saccular, dextral or sinistral to seminal vesicle, between ovary and ventral sucker. Vitelline follicles confined almost entirely to forebody, confluent dorsally, extend from slightly posterior to ventral sucker to near posterior margin of oral sucker. Uterine coils restricted to hindbody, extend from posterior end of body to ventral sucker. Excretory vesicle Y-shaped; arms reach pharynx. It is to be treated as masculine. Lobosorchis polygongylus n. sp. (Fig. 1) Type host:Lutjanus gibbus (Forsskål), Perciformes, Lutjanidae, humpback red snapper. Type locality: Rasdhoo Atoll (4º16’N; 72º58’E), Maldives. Additional localities: Lizard Island, Great Barrier Reef (14º40’S; 145º27’E), Queensland, Australia. Site: Intestine and pyloric caeca. Prevalence: 5 of 5 (100%) at Rasdhoo Atoll; 2 of 7 (29%) at Lizard Island. Molecular sequence data: ITS, 3 individuals from Lizard Island, 4 individuals from the Maldives; LSU, 3 individuals from Lizard Island, 3 individuals from the Maldives. Deposited specimens: holotype G230563, 12 paratypes G230564-G230575. GenBank accession numbers: ITS (FJ154900), LSU (FJ154902). Etymology: The epithet polygongylus is derived from the Greek poly, meaning many and the Greek gongylos, meaning ball-shaped, round or spherical, referring to the multiple testicular follicles present in this species. 40 · Zootaxa 1992 © 2009 Magnolia Press MILLER ET AL. FIGURE 1.Lobosorchis polygongylusn. sp. from the intestine of Lutjanus gibbus off Rasdhoo Atoll, Maldives. Ventral view of holotype. Scale bar = 200 μm. NEW SPECIES OF LOBOSORCHIS (CRYPTOGONIMIDAE) Zootaxa 1992 © 2009 Magnolia Press · 41 Description: Based on 13 specimens. Body oval, longer than wide, 959 (598–1,176) long by 580 (374–688) wide; length/width ratio 1.7 (1.4–1.9). Oral sucker 123 (78–154) long by 217 (165–254) wide. Oral spines 70 (60–81), length 19 (8–27). Ventral sucker 68 (53–80) long by 78 (59–88) wide. Ratio oral/ventral sucker width 2.8 (2.5–3). Forebody occupying 34 (31–36)% of body length. Prepharynx shorter than oesoph- agus, 15 (6–22) long. Pharynx 92 (62–115) long by 93 (70–109) wide. Ventral sucker/pharynx width ratio 0.8 (0.8–0.9). Oesophagus 36 (19–48) long. Intestinal bifurcation between ventral sucker and pharynx. Intestinal caeca blind, 668 (390–852) long, terminate close to posterior end of body. Testes multiple, in two symmetrical groups, total number 19 (13–25), in mid-hindbody, individual testes 87 (45–131) long by 99 (43–179) wide. Seminal vesicle tubulosaccular, between ovary and ventral sucker. Gonotyl a small lobe arising from anterior wall of ventrogenital sac. Genital pore immediately anterior to ventral sucker. Ovary deeply lobed, immedi- ately anterior and adjacent to testes, 171 (122–202) long by 190 (134–256) wide. Laurer's canal present. Sem- inal receptacle saccular, dextral or sinistral to seminal vesicle, between ovary and ventral sucker. Vitelline follicles confined almost entirely to forebody, confluent dorsally, extend from slightly posterior to ventral sucker to near posterior margin of oral sucker. Uterine coils restricted to hindbody, extend from posterior end of body to ventral sucker. Eggs small, darkly tanned, 18 (15–21) long by 8 (7–9) wide. Excretory vesicle Y- shaped, bifurcates dorsal to ovary; arms extend to pharynx, 757 (442–944) long. Excretory pore opens at pos- terior end of body. Lobosorchis tibaldiae Miller & Cribb, 2005 Metacercariae (Fig. 2) Hosts: Neoglyphidodon melas (Cuvier), Perciformes, Pomacentridae, bowtie damselfish; Pomacentrus mel- anochir (Bleeker), Perciformes, Pomacentridae, Indonesian damsel. Locality: Heron Island, Great Barrier Reef (23º26’S; 151º54’E), Queensland, Australia. Site: Flesh. Deposited specimens: vouchers G230576–G230578. Description: Based on 3 specimens. Body oval, longer than wide, 388 (328–426) long by 238 (163–289) wide; length/width ratio 1.7 (1.5–2). Oral sucker 64 (63–65) long by 111 (80–128) wide. Oral spines 60 (54–65), length 11 (8–15). Ventral sucker 42 (40–45) long by 42 (38–43) wide. Ratio oral/ventral sucker width 2.6 (2.1–3). Forebody occupying 37 (35–40)% of body length. Prepharynx 11 (6–13) long. Pharynx 39 (35–45) long by 49 (45–53) wide. Ventral sucker/pharynx width ratio 0.9 (0.8–0.9). Oesophagus 21 (16–26) long. Intestinal bifurcation dorsal to or immediately anterior to ventral sucker. Intestinal caeca blind, 398 (286–494) long, terminate midway between ventral sucker and posterior end of body or close to posterior end of body. Testes multiple, in two symmetrical groups of 4–5, in mid-hindbody, individual testes 21 (14–29) long by 22 (14–29) wide. Gonotyl a small lobe arising from anterior wall of ventrogenital sac, often undevel- oped. Genital pore immediately anterior to ventral sucker. Ovary deeply lobed, immediately anterior and adja- cent to testes, 22 (16–26) long by 26 (18–35) wide. Laurer's canal present. Vitelline system undeveloped. Excretory vesicle Y-shaped, bifurcates immediately posterior to or dorsal to ovary; arms extend to pharynx, 275 (230–307) long. Excretory pore opens at posterior end of body. Molecular data Lobosorchis sp. A Metacercariae Hosts:Ecsenius stictus (Springer), Perciformes, Blenniidae, Great Barrier Reef blenny; Salarias alboguttatus (Kner), Perciformes, Blenniidae, white-spotted blenny; Salarias fasciatus (Bloch), Perciformes, Blenniidae, 42 · Zootaxa 1992 © 2009 Magnolia Press MILLER ET AL. FIGURE 2. Lobosorchis tibaldiae Miller & Cribb, 2005 metacercaria from the flesh of Neoglyphidodon melas off Heron Island, Great Barrier Reef, Australia. Scale Bar = 200 μm. NEW SPECIES OF LOBOSORCHIS (CRYPTOGONIMIDAE) Zootaxa 1992 © 2009 Magnolia Press · 43 jewelled blenny; Canthigaster bennetti (Bleeker), Tetraodontiformes, Tetraodontidae, Bennett’s sharpnose puffer. Locality: Lizard Island, Great Barrier Reef (14º40’S; 145º27’E), Queensland, Australia. Site: Flesh. Molecular sequence data: ITS2, 5 individuals. GenBank accession number: ITS2 (FJ154903). Lobosorchis sp. B Metacercariae Hosts:Dischistodus perspicillatus (Cuvier), Perciformes, Pomacentridae, white damsel. Locality: Lizard Island, Great Barrier Reef (14º40’S; 145º27’E), Queensland, Australia. Site: Flesh. Molecular sequence data: ITS2, 1 individual. GenBank accession number: ITS2 (FJ154904). ITS rDNA Alignment of the ITS rDNA region (ITS1, 5.8S and ITS2) between L. polygonylus and L. tibaldiae only revealed that these two taxa differed by 5 bp in total (3 bp in ITS1, 0 bp in 5.8S and 2 bp in ITS2). The three base differences in the ITS1 between L. polygongylus and L. tibaldiae were due to (in consecutive order) an adenine insertion, an adenine – guanine transition and a cytosine – thymine transition. The two base pair dif- ferences in the ITS2 dataset were both due to adenine – cytosine transversions. No intraspecific variation was observed in any of the taxa sequenced over the ITS rDNA region. Variation in the 5’ half of the ITS1 made alignment in this region between the all of the cryptogonimid taxa examined here impossible, so only the 3’ half of the ITS1 was included for comparative purposes as in Miller and Cribb (2007b; c; d), because this region was easily alignable. Alignment of the ITS dataset contain- ing the other cryptogonimid taxa, which included the 3’ end of the ITS1, the entire 5.8S and ITS2 and 13 bp of the 5’ end of the LSU yielded 874 characters for analysis. Lobosorchis polygongylus differed from L. tibal- diae by 3 bp in the 5’ trimmed ITS dataset (1 bp in the ITS1 and 2 bp in the ITS2). LSU rDNA Sequencing of the LSU rDNA yielded an average of approximately 870 bp for all taxa. The aligned and trimmed sequences incorporated a total of 859 characters (base pairs and gaps) for analysis. Lobosorchis polygongylus differed from L. tibaldiae by 1 bp in the LSU dataset. This single base difference was due to a cytosine – thymine transition. No intraspecific variation was observed in any of the taxa sequenced over the LSU rDNA region. Combined ITS and LSU rDNA dataset analyses Bayesian inference analysis of the combined ITS (included 3’ end of ITS1, 5.8S and ITS2) and LSU rDNA dataset resulted in a phylogram in which Lobosorchis polygongylus and L. tibaldiae form a well-supported clade sister to Neometadena ovata (Hafeezullah & Siddiqi, 1970) Miller & Cribb, 2008 (Fig. 3). All genera were well resolved with high posterior probabilities in the Bayesian analysis. Species of Retrovarium formed 44 · Zootaxa 1992 © 2009 Magnolia Press MILLER ET AL. a clade sister to Chelediadema marjoriae Miller & Cribb, 2007. Caulanus thomasi Miller & Cribb, 2007 formed a strong clade with species of Latuterus and species of Beluesca and Chelediadema (both found in haemulids) were distinctly genetically distant as observed in Miller and Cribb (2007c; d). Minimum evolution analysis (ME score = 796.87) of the combined ITS and LSU dataset resulted in a phy- logram identical in topology to that observed in the Bayesian inference analysis. Nodal support was also rela- tively high for the observed clades (Fig. 3). ITS2 only dataset analysis Alignment of the ITS2 rDNA region for Lobosorchis polygongylus, L. tibaldiae, putative Lobosorchis sp. metacercariae obtained from the body cavity or flesh of intermediate hosts on the GBR, and the remainder of the cryptogonimid taxa yielded 309 bp (base pairs and gaps) for analysis. This alignment revealed the pres- ence of two additional genotypes (no ‘intraspecific’ variation was observed within these genotypes), one asso- ciated with metacercariae obtained from three species of Blenniidae and one species of Tetraodontidae (putative Lobosorchis sp. A) and the other genotype associated with the specimen obtained from the pomacen- trid Dischistodus perspicillatus (putative Lobosorchis sp. B). These two additional genotypes consistently dif- fered from L. polygongylus by 1 and 4 bp, L. tibaldiae by 1 and 4 bp and from each other by 3 bp over the ITS2 dataset. Lobosorchis polygongylus and L. tibaldiae differed from one another by 2 bp in the ITS2 rDNA region. The three base pair differences between putative Lobosorchis sp. A and B were all due to cytosine – thymine transitions. Minimum evolution analysis (ME score = 208.42) of the ITS2 dataset resulted in a phylogram with spe- cies of Lobosorchis forming a well-resolved clade (Fig. 4). All genera were well resolved, similar to Bayesian inference analysis of the combined ITS and LSU dataset, but intergeneric relationships were poorly resolved as evidenced by the low nodal support observed. Discussion Differential diagnosis Lobosorchis polygongylusn. sp. can be easily distinguished from L. tibaldiae by the combination of body size, oral spine number (60–81 in L. polygongylus, 47–56 in L. tibaldiae) and number of testicular follicles (13–25 in L. polygongylus, 9 in L. tibaldiae). Species of a few cryptogonimid genera have multiple (= 9) testes (Acanthosiphodera Madhavi, 1976, Iheringtrema Travassos, 1947, Novemtestis Yamaguti, 1942, Polyorchi- trema Srivastava, 1939 and Siphodera Linton, 1910), but are all easily distinguished from L. polygongylus based on testes number (all of the species in these genera except for species of Polyorchitrema have only 9 testes), presence or absence of oral spines (species of Iheringtrema and Siphodera lack oral spines) and vitell- ine follicle distribution (species of Acanthosiphodera and Novemtestis have vitelline follicles confined almost entirely or entirely posterior to the ventral sucker). The only other cryptogonimids that have been reported with more than approximately 9 testes are species of Polyorchitrema, which may have 8–50 testes (Miller & Cribb 2008), but these differ significantly from L. polygongylus in lacking oral spines, the vitelline follicles are distributed in lateral groups posterior to the ventral sucker, the ovary is entire to slightly lobed and the uterus is confined mainly anterior to the testes. Morphological disparity in light of genetic similarity Despite the extensive morphological differences between L. polygongylus and L. tibaldiae, these two spe- cies differed from each other by only 5 bp (0.5% sequence divergence) in the ITS and 1 bp (0.1%) in the LSU NEW SPECIES OF LOBOSORCHIS (CRYPTOGONIMIDAE) Zootaxa 1992 © 2009 Magnolia Press · 45 FIGURE 3. Phylogram of relationships between species of Lobosorchis Miller & Cribb, 2005 and the other cryp- togonimid taxa included based on Bayesian inference analysis of the combined ITS (includes partial ITS1 and complete 5.8S and ITS2) and LSU rDNA dataset. Posterior probabilities are shown above the nodes and bootstrap support values shown below. Phylogram is midpoint rooted. 46 · Zootaxa 1992 © 2009 Magnolia Press MILLER ET AL.