Specimens were collected from a narrow artificial canal running from Wenshan lake in Shenzhen city, Guangdong Province, China (22°31'55"N, 113°56'21"E) on 10 May 2021 (Fig. 1). A 200-μm-mesh sieve was used to collect Cladophora algae, to which the worms were attached. The contents of the mesh sieve were washed into a bucket using habitat water, and then transported to the laboratory of Shenzhen University for further analysis and culturing. The flatworms were reared in a glass aquarium (21 cm × 15 cm; depth 18 cm) at room temperature (23–26 °C). The culture was aerated, and the flatworms were fed daily with Daphnia.

Figure 1. 

Locality and habitat of Dugesia ancoraria. A. Sampling locality in Guangdong Province, China; B, C. Habitat at sampling locality.

After starvation for three days, total DNA was extracted from three sexual individuals using the E.Z.N.A.^TM Mollusc DNA Isolation Kit (Omega, Norcross, GA, USA). Four gene fragments, namely 18S ribosomal gene (18S rDNA), 28S ribosomal gene (28S rDNA), ribosomal internal transcribed spacer-1 (ITS-1), and cytochrome C oxidase subunit I (COI), were amplified by polymerase chain reaction (PCR). We used 2×Taq Plus Master Mix II (Vazyme, China) to amplify 18S rDNA, 28S rDNA, ITS-1, and COI. Primers used for amplification and the PCR protocol are listed in Table 1. Forward and reverse DNA strands were determined by Sanger sequencing either at BGI (Guangzhou, China) or TsingKe Biotech (Beijing, China). All new sequences have been uploaded to GenBank, NCBI (Table 2).

To determine the phylogenetic position of the new species within the genus Dugesia, we generated datasets consisting of marker gene sequences (18S rDNA, 28S rDNA, ITS-1, and COI; see Table 2) of the new species Dugesia ancoraria and available sequences of other Dugesia species from GenBank, NCBI, as well as two outgroup species, viz., Recurva postrema Sluys & Solà, 2013 (ITS-1 sequence not available in GenBank, NCBI), and Schmidtea mediterranea (Benazzi, Baguñà, Ballester, Puccinelli & del Papa, 1975).

Nuclear ribosomal markers were aligned with MAFFT (online version 7: MAFFT alignment and NJ / UPGMA phylogeny (cbrc.jp), Katoh et al. 2017) using the E-INS-i algorithm, while mitochondrial coding gene COI was aligned by MASCE v2.03 (Ranwez et al. 2018). In order to check for the absence of stop codons, COI sequences were translated into amino acids by ORFFINDER in NCBI, applying genetic code 9 before alignment, after which regions of ambiguous alignments were removed by Gblocks v0.91b (Talavera and Castresana 2007), using the same parameters as specified in Li et al. (2019). For ribosomal DNA, sequences were excluded by ClipKIT (Steenwyk et al. 2020) with kpic-gappy mode to keep parsimony-informative and constant sites and to remove highly gappy sites. Final length of the alignments was 693 base pairs (bp) for COI, 1,388 bp for 18S rDNA, 1,383 bp for 28S rDNA, and 591 bp for ITS-1. To ensure sequences’ validity, the substitution saturation test (Xia et al. 2003; Xia and Lemey 2009) in DAMBE6 software (Xia 2017) was used to evaluate the nucleotide substitution saturation of four datasets, followed by the assembly of a multi-gene concatenated dataset (with the order 18S rDNA–28S rDNA–ITS-1–COI), using SequenceMatrix v1.8 (Vaidya et al. 2011). Sequences that were shorter or not available in GenBank were completed with “-”. We used PartitionFinder2 (Lanfear et al. 2017) to evaluate the best-fit evolution models by estimating independent models of molecular evolution for subsets of sites that were deemed to have evolved in similar ways.

Phylogenetic trees were constructed by Maximum Likelihood (ML) and Bayesian Inference (BI) methods. For ML, standard bootstrap analysis with 1,000 replications was performed by IQ-TREE v1.6.2 (Nguyen et al. 2015). BI was performed in MrBayes v3.2.6 (Ronquist et al. 2012) with two simultaneous runs of one cold and three hot chains. Each run for the concatenated dataset was performed for 1,000,000 generations, sampling every 1,000 generations. We checked the resulting parameter file of each run in TRACER v1.7.1 (Rambaut et al. 2018) to ensure that the effective sample size (ESS) values of each parameter were above 200.

After having been starved for three days, the mitochondrial DNA of an asexual specimen of D. ancoraria was extracted (due to absence of sexual specimens at that time) using Animal mitochondrial DNA column extraction kit (PCR Grade; BioLebo Technology, Beijing China), followed by amplification of mitochondrial DNA using a REPLI-g Midi Kit (QIAGEN, Hilden, Germany). We compared the COI gene of the three sexual individuals with that of the asexual individual for mitochondrial extraction via megablast and, thus, found that they were perfectly identical. Paired-end sequencing was conducted on the Illumina Hiseq 2500 platform (BGI, Guangzhou, China). The mitogenome sequences were assembled using MitoFinder (Allio et al. 2020). The functional regions of these genes were annotated and verified according to Huang et al. (2022). However, both MITOS (online version: MITOS Web Server (uni-leipzig.de), Bernt et al. 2013) and tRNAscan-SE (Lowe and Chan 2016;) failed to annotate trnT in the mitogenome of D. ancoraria. Therefore, trnT was identified manually on the basis of homology comparisons with other species in the family Dugesiidae Ball, 1974. Seven species belonging to Dugesiidae, five to Geoplanidae Stimpson, 1857, and two species belonging to Planariidae Stimpson, 1857 were chosen for the construction of the mitochondrial tree (Table 3). As outgroup taxon we used the maricolan species Obrimoposthia wandeli (Hallez, 1906), which is a member of Uteriporidae Diesing, 1862 (Table 3). Multiple sequences alignments (MSA) for protein coding genes (PCGs) and ribosomal genes were carried out using MASCE v2.03, which translates the nucleotides to amino acids before alignment. Hereafter, MSAs were trimmed by Gblocks v0.91b. Substitution saturation tests for each PCG were performed using DAMBE6. Subsequently, SequenceMatrix v1.8 was used to combine the alignments. The best-fit model for each PCG was selected by PartitionFinder2. ML analysis was conducted by IQ-TREE v1.6.2. For BI, MrBayes v3.2.6 was applied with 2,000,000 generations, sampling every 2,000 generations. Gene rearrangement scenarios including reversals, transpositions, reverse transpositions, reversal and tandem-duplication-random-losses (TDRL) among all species were analysed using the software CREx (Bernt et al. 2007) on the CREx web server (http://www.pacosy.informatik.uni-leipzig.de/crex) based on common intervals.

For the morphological analysis, the flatworms were starved for three days prior to the preparation of histological sections according to procedures described by Song et al. (2020). Briefly, histological sections were made at intervals of 6 μm and were stained with modified Cason’s Mallory-Heidenhain stain solution (see Yang et al. 2020). Hereafter, slides were mounted onto glass slides with neutral balsam (Yuanye Biotechnology, Shanghai, China) and sealed with a coverslip. Preparations registered with PLA codes were deposited in the Institute of Zoology, Chinese Academy of Sciences (IZCAS), while histological slides registered with RMNH.VER. codes will be deposited at Naturalis Biodiversity Center, Leiden, The Netherlands.

au: auricle; bc: bursal canal; ca: common atrium; cb: copulatory bursa; cm: circular muscle; d: diaphragm; db: distal bulge; du: duct; e: eye; ed: ejaculatory duct; esv: extension seminal vesicle; go: gonopore; hb: hunchback bump; ie: inner epithelium; lm: longitudinal muscle; lod: left oviduct; lvd: left vas deferens; ma: male atrium; od: oviduct; oe: outer epithelium; ov: ovary; pg: penis glands; ph: pharynx; pp: penis papilla; rod: right oviduct; rvd: right vas deferens; sg: shell glands; sv: seminal vesicle.

The phylogenetic trees obtained by BI and ML from the concatenated dataset (with the order 18S rDNA–28S rDNA–ITS-1–COI) showed similar topologies and supported nodes (Fig. 2). In this tree, the terminals for D. ancoraria grouped together and did not group with any other species of Dugesia included in our molecular analysis. It is noteworthy that the new species D. ancoraria formed a well-supported clade with D. notogaea Sluys & Kawakatsu, 1998 and D. bengalensis Kawakatsu, 1972 (Fig. 2, 100% bootstrap [bs] in ML; Suppl. material 1, 1.00 posterior probability [pp] in BI). Dugesia notogaea is an Australian species, while D. bengalensis inhabits a portion of India. As suggested by the high support value (99% bs; 1.00 pp), these three species are closely related and share a sister-group relationship with D. adunca Chen & Sluys, 2022 from Guangxi province. Then, these four species are supported as sister to a small clade composed of D. ryukyuensis Kawakatsu, 1976 and D. batuensis Ball, 1970 (96% bs; 1.00 pp).

Figure 2. 

Maximum likelihood phylogenetic tree topology inferred from the concatenated dataset (18S rDNA, ITS-1, 28S rDNA and COI). Numbers at nodes indicate support values (bootstrap/ posterior probability). Asterisks (*) indicate support values lower than 50% bs/0.50 pp, or posterior probability not applicable to this node, because of different topologies of trees generated by BI and ML methods. Scale bar: substitutions per site.

The complete, circular mitochondrial genome of Dugesia ancoraria is 17,705 bp in length, and includes 12 of the 13 protein-coding genes of mitochondrial genomes (atp8 was not found), two ribosomal RNA (rRNA) genes, and 22 transfer RNA (tRNA) genes, which are arranged as follows: cox1-E-nad6-nad5-S2-D-R-cox3-I-Q-K-atp6-V-nad1-W-cox2-P-nad3-A-nad2-M-H-F-rrnS-L1-Y-G-S1-rrnL-L2-T-C-N-cob-nad4l-nad4. GC content is 23.77%, while a positive GC skew ([G-C]/[G+C] = 0.323) indicated the occurrence of more Gs than Cs (Fig. 3).

Figure 3. 

Arrangement of the mitochondrial genome of Dugesia ancoraria. Outer circle: annotation of genes, with protein-coding genes, ribosomal RNAs and transfer RNAs represented by cyan, orange, and red, respectively. Intermediate circle: sequencing coverage, with green colour indicating coverage greater than 95% average coverage. Inner circle, with the blue colour indicating GC content and the thin orange circle indicating 50% of GC content. The picture in the middle represents an individual of D. ancoraria.

Both the ML and BI trees obtained from 12 protein coding genes (PCGs) have highly supported clades, excepting one node with a bootstrap support lower than 70%. Since the topologies of the ML and BI trees are basically identical, we integrated them into one phylogenetic tree. In the integrated tree, Crenobia alpina (Dana, 1766) and Phagocata gracilis (Haldeman, 1840) together form a clade that shares a sister-group relationship with a clade that is composed of two smaller clades, one comprising land planarians (Geoplanidae) and the other constituted by dugesiid freshwater planarians (Dugesiidae). The latter family forms a well-supported monophyletic group, in which D. ancoraria is sister to D. ryukyuensis with high support values (100% bs; 1.00 pp) (Fig. 4).

Figure 4. 

Possible mechanisms of mitochondrial gene rearrangement in Continenticola estimated using CREx, with reference to phylogenetic relationships. On the left-hand side, phylogenetic tree obtained from Maximum likelihood and Bayesian analysis of the concatenated dataset for the protein-coding and rRNA genes within mitochondrial genomes; numbers at nodes indicate support values (pp/bs); Asterisks (*) indicate that bootstrap is not applicable to these nodes because of different topologies of trees generated by BI and ML methods. Pink lines connect species that share the same gene order. On the right-hand side, changes of gene order in mitochondrial genomes in several species of triclad; protein-coding genes in blue, tRNA genes in white, rRNA genes in grey; lines with different colours indicate transpositions of different genes; red lines indicate tandem duplication random loss (TDRL) events between species, while the orange colour indicates where TDRL events occurred.

The gene order of rRNAs, PCGs and tRNAs of D. ancoraria and other species used in our phylogenetic analysis are shown in Fig. 4. Some tRNAs absent in previous publications, such as those from Platydemus manokwari de Beauchamp, 1963 and Parakontikia ventrolineata (Dendy, 1892), had been successfully annotated using MITOS (Bernt et al. 2013). Our results show that the orders of PCGs and ribosomal genes are conserved among species belonging to the triclad suborder Continenticola and are arranged as follows: cox1-nad6-nad5-cox3-atp6-nad1-cox2-nad3-nad2-rrnS-rrnL-cob-nad4l-nad4. In contrast, the order of tRNAs is highly variable. Within the cluster of Dugesiidae species, D. ancoraria shares an identical gene order with D. constrictiva. An analysis of gene order rearrangements with CREx suggests that only one transposition (trnN) occurred from D. ancoraria to D. ryukyuensis and also one transposition (trnE) from D. ryukyuensis to D. japonica. Except for a transposition of trnE, a tandem-duplication-random-loss (TDRL) event is required for the transformation from D. japonica to Schmidtea mediterranea (Benazzi et al., 1975). With respect to the Geoplanidae, Amaga expatria Jones & Sterrer, 2005 and Obama sp. share the same gene order. Besides a transposition of trnF, a transposition of trnM and trnH linkage is needed to go from C. alpina to Bipalium kewense Moseley, 1878. Two inverse transpositions, namely trnL2 and trnT, occurred from Bipalium kewense to Obama sp. and Amaga expatria, and from Obama sp. to Platydemus manokwari, resulting in almost the same gene order shared by B. kewense and P. manokwari, with the only exception being the position of trnC (Fig. 4).

Order Tricladida Lang, 1884

Suborder Continenticola Carranza, Littlewood, Clough, Ruiz-Trillo, Baguñà & Riutort, 1998

Family Dugesiidae Ball, 1974

Genus Dugesia Girard, 1850

Dugesia ancoraria Zhu & Wang, sp. nov.

https://zoobank.org/137337E1-0D52-4288-A3A3-7E8C80241A74

Material examined

Holotype : PLA-0251, a narrow artificial canal of Wenshan lake, Shenzhen city, Guangdong Province, China, 22°31'55"N, 113°56'21"E, 10 May 2021, coll. MY Xia and co-workers, sagittal sections on 14 slides.

Paratypes : PLA-0252, ibid., sagittal sections on 12 slides; PLA-0253, ibid., transverse sections on 35 slides; RMNH.VER.21525.1, ibid., sagittal sections on 12 slides.

Habitat

Specimens were collected from a narrow artificial canal running from Wenshan lake (22°31'55"N, 113°56'21"E), which is located in Shenzhen city, Guangdong Province, China (Fig. 1A). The animals were collected from Cladophora algae, as well as the stone wall of the canal, which had a water depth of 20–30 cm; water temperature was about 23 °C. Thirty specimens were collected, none of which was sexually mature. However, after six months of rearing under laboratory conditions, about 20 specimens eventually attained sexual maturity.

Diagnosis

Dugesia ancoraria is characterised by the following characters: highly asymmetrical penis papilla, provided with a hunchback-like dorsal bump; vasa deferentia opening symmetrically into the mid-lateral section of the more or less ellipsoidal seminal vesicle, which may give rise to a narrow dorsal extension; long and narrow duct connecting seminal vesicle with small diaphragm; ejaculatory duct with a subterminal opening through the ventral surface of the penis papilla; asymmetrical oviducal openings, with the right oviduct opening into a section of the bursal canal that bends ventrally to communicate with the common atrium; the left oviduct opens into the bursal canal at the point where the latter meets the common atrium.

Etymology

The specific epithet is derived from Latin adjective ancorarius, of the anchor, and alludes to the penis papilla, which has a hunchback-shape, reminiscent of an anchor, more or less.

Description

Sexualized specimens measured 8.43–9.11 mm in length and 1.13–1.18 mm in width (n = 4; Fig. 5A, B). Head of low triangular shape with blunt auricles. At the level of the auricles there is a pair of black, bean-shaped eyes, located in pigment-free areas. The distance between the eyes and the lateral body margin is about 0.36–0.46 mm, while the size of the eyecups varies between 210–230 μm. Each eyecup contains numerous retinal cells.

Figure 5. 

External morphology of Dugesia ancoraria. A. Living sexual animal in dorsal view; B. Living sexual animal in ventral view; C. Anterior end, dorsal view; D. Ventral view of rear end, showing pharynx, mouth and gonopore. Scale bars: 500 μm.

The ground colour of the dorsal surface is brown, dotted with dark brown and white specks; ventral surface much paler than dorsal surface; the body margin is pale (Fig. 5A, B).

The cylindrical pharynx is positioned at about 1/2 of the body and measures about 1/5 of the total body length; the mouth opening is situated at the posterior end of the pharyngeal pocket. The musculature of the pharynx consists of an outer, subepithelial layer of circular muscle, followed by a layer of longitudinal muscle, while the inner musculature is composed of a thick, subepithelial layer of circular muscle, followed by 2–3 layers of longitudinal muscle. The gonopore is situated at about 1/5 of the length of the body, as measured from the posterior body margin (Fig. 5D).

The globular ovaries are located at 1/6 – 1/7 of the distance between the brain and the root of pharynx. From the ovaries, the nucleated oviducts run ventrally in a caudal direction and open separately and asymmetrically into the female reproductive apparatus. Posterior to the gonopore, the right oviduct turns antero-medially and then opens into a section of the bursal canal that bends ventrally to communicate with the common atrium. The left oviduct opens into the bursal canal at the point where the latter meets the common atrium. (Figs 6A, 9B).

Figure 6. 

Dugesia ancoraria, holotype PLA-0101, sagittal sections, anterior to the right. A. Photomicrograph showing copulatory bursa, ejaculatory duct, penis papilla, and seminal vesicle; B. Photomicrograph showing copulatory bursa, bursal canal, and gonopore; C. Photomicrograph showing copulatory bursa, and bursal canal. Scale bars: 100 μm.

A large sac-shaped copulatory bursa is situated immediately behind the pharyngeal pocket and occupies the entire dorso-ventral space; it is lined with a layer of vacuolated, nucleated cells (Figs 6B, 9B). From the bursa, the bursal canal runs in a caudal direction dorso-laterally to the male copulatory apparatus. At the level of the gonopore, the bursal canal curves rather sharply downwards, thus giving rise to a more or less vertically oriented section that opens through the dorsal wall of the common atrium (Figs 6C, 9B).

The bursal canal is lined by a nucleated, columnar glandular epithelium, which is underlain with a layer of longitudinal muscles, followed by 1–4 layers of circular muscles. Along the ventral coat of muscle, ectal reinforcement is present in the form of a single layer of longitudinal muscle running from about the opening of the canal into the common atrium to about 1/3 of the length of the bursal canal (Fig. 9B). Shell glands discharge their cyanophil secretion into the most ventral section of the vertically running portion of the bursal canal, with some glands even discharging into the common atrium (Figs 6B, C, 9B).

The large, near-globular testicular follicles are situated dorsally and extend posteriorly from a short distance behind the brain to well beyond the copulatory apparatus. The male atrium comprises most of the dorso-ventral space of the body (Figs 6A, 9A, B). The large and oval-shaped penis bulb is composed of intermingled longitudinal and circular muscle fibres. The penis papilla has a more or less oblique, postero-ventral orientation or even a vertical orientation, and has a striking shape (Figs 6A, 9A). The papilla is markedly asymmetrical as a result of the course of ejaculatory duct, which opens to the exterior through the postero-ventral wall of the penis papilla. Furthermore, near its root, the papilla has a dorsal bump, which gives it a hunchback appearance (Figs 6A, 9A). The degree of development of this dorsal bump differs between specimens. In the holotype it is highly developed (Fig. 6A), while in paratype PLA-0104 it is somewhat smaller, albeit still well-developed (Fig. 8A), but in paratype PLA-0102 the bump is practically absent (Fig. 7A). In addition, the asymmetrical appearance of the penis papilla is enhanced by the fact that the distal portion of the dorsal lip of the papilla gives rise to another bulge, which may be swollen or drawn-out to a greater or lesser extent. In paratype RMNH.VER.21525.1, it is a rather long-drawn bulge (Fig. 8A), whereas in the holotype and paratype PLA-0102 it is more rounded (Fig. 7A). The papilla is covered by a thin, nucleated epithelium, which is underlain with a well-developed, subepithelial layer of circular muscle, followed by a layer of longitudinal muscle at the ventral root.

Figure 7. 

Dugesia ancoraria, paratype PLA-0102, sagittal sections. A. Photomicrograph showing ejaculatory duct, penis papilla, and seminal vesicle; B. Photomicrograph showing copulatory bursa, bursal canal, and gonopore. Scale bars: 100 μm.

Figure 8. 

Dugesia ancoraria, paratype RMNH.VER.21525.1, sagittal sections. A. Photomicrograph showing ejaculatory duct, penis papilla, and seminal vesicle; B. Photomicrograph showing copulatory bursa, common atrium, and gonopore; C. Photomicrograph showing copulatory bursa, and bursal canal. Scale bars: 100 μm.

The vasa deferentia have expanded to form spermiducal vesicles that are packed with sperm. At the level of the penis bulb, the ducts recurve, while decreasing in diameter, run postero-medially for some distance and, thereafter, recurve anteriad before opening separately into the mid-lateral portion of the seminal vesicle (Figs 6A, 9A, B). The vesicle has a more or less ellipsoidal shape, while its dorsal wall may form a narrow extension, which was present in all specimens examined, excepting paratype RMNH.VER.21525.1. The seminal vesicle is lined by a ciliated, nucleated epithelium. A long and narrow duct connects the seminal vesicle with the small diaphragm, which communicates with the ejaculatory duct. The diaphragm receives the abundant secretion of erythrophil penial glands. From the diaphragm, the ejaculatory duct curves strongly postero-ventrally to open subterminally through the ventral epithelium of the penis papilla, thus giving rise to a highly asymmetrical papilla with a large dorsal lip and a small ventral lip (Figs 6A, 9A). Particularly the blunt tip of the dorsal lip of the penis papilla is penetrated by the numerous openings of orange-staining glands.

Figure 9. 

Dugesia ancoraria, Sagittal reconstruction of the copulatory apparatus of the holotype. A. Male copulatory apparatus; B. Female copulatory apparatus. Scale bars: 100 μm.

The male atrium is lined by a nucleated epithelium. The dorsal part of the male atrium is surrounded by a layer of circular muscle, followed by 1–2 layers of longitudinal muscle, while a subepithelial layer of circular muscle, followed by a layer of longitudinal muscle constitutes the musculature on the ventral part of the atrium. The male atrium communicates with the common atrium via a broad opening. The common atrium is lined with a nucleated epithelium, which is underlain by 2–3 layers of circular muscle (Figs 6A, 9).

In our phylogenetic tree (Fig. 2), the terminals for D. ancoraria grouped together, while they did not group with any other species of Dugesia included in our molecular analysis. Thus, the molecular analysis already suggested that D. ancoraria concerns a new species of Dugesia, which was supported by the morphological study (see below).

In the phylogenetic trees obtained from the concatenated dataset (Fig. 2), the sister-group relationship between D. ancoraria on the one hand and D. notogaea and D. bengalensis on the other hand is consistent and is supported by high bootstrap values, strongly suggesting that these three species form a monophyletic group. It is noteworthy that D. ancoraria from southern China, shares only a distant relationship to other Dugesia species from China, including D. constrictiva, D. verrucula, D. majuscula, D. circumcisa, D. semiglobosa, D. umbonata, D. gemmulata and D. tumida, but is most closely related to D. notogaea from Australia and D. bengalensis from India. The clade comprising D. ancoraria, D. notogaea and D. bengalensis shares a sister-group relationship with D. adunca, then is sister to a small clade comprising D. ryukyuensis from Japan and D. batuensis from peninsular Malaysia, and then further clusters with D. deharvengi, which is basically in agreement with the results of Chen et al. (2022). However, according to Liu et al. (2022), D. deharvengi shares a sister-group relationship with D. notogaea first, and then clusters with a group comprising D. ryukyuensis and D. batuensis, which could be due to the absence in the species phylogeny of the COI sequence of D. bengalensis. Actually, in the phylogenetic tree generated solely on COI sequences, D. adunca shared a sister-group relationship with D. deharvengi, albeit with low support, while D. ryukyuensis and D. batuensis clustered with a clade consisting of D. majuscula, D. verrucula, D. constrictiva and D. tumida with very low support (data not shown here), instead of clustering with D. notogaea and D. bengalensis, as in concatenation-based analyses. Furthermore, since different genes may exhibit highly variable rates of evolution, phylogenies inferred from single genes, with only limited evolutionary information, are often inconsistent. Altogether, our results suggested that concatenation-based analyses resulted in more resolved phylogenetic trees.

With respect to the geographical distribution within China of other species of Dugesia, in relation to the distribution of D. ancoraria, the following should be noted. Dugesia japonica has a wide distribution, as it has been reported from the eastern, southern and northern regions of China, while the remaining 11 species are found only in southern China. Among these species, D. semiglobosa and D. majuscula were documented in Hainan province, D. circumcisa and D. adunca in Guangxi province, with these two provinces being relatively close to the locality of D. ancoraria in Guangdong. Dugesia tumida and D. sinensis occur also in Guangdong Province, while they share only a very distant relationship with the new species D. ancoraria. Other species, like D. umbonata, were found in Jiangsu province and D. gemmulata in Guizhou. Although these two species are geographically far distant from each other, they share a close relationship.

The close relationship between Chinese D. ancoraria and Australian D. notogaea, with the latter being the sister-species of Malaysian D. bengalensis, is interesting from a historical biogeographic perspective. This pattern of relationships basically agrees with that uncovered by Solà et al. (2022), in which D. notogaea also fell into an Asian clade, including specimens from China, Malaysia, Thailand, Japan, and Indonesia. These authors surmised that this pointed to anthropochore dispersal of Dugesia from Asia to Australia, as they considered Wallace’s Line to indicate an unsurmountable biological barrier for natural arrival of the genus in Australia (see also Ali and Heaney 2022). Excluding unlikely jump dispersal, earlier hypotheses that Dugesia naturally spread from Southeast Asia to Australia (Sluys et al. 1998) foundered on the paleogeographical evolution of the Indo-Australian archipelago. According to paleogeographical reconstructions, the river systems of Asia on the one hand and those of Australia/New Guinea on the other hand have never been in contact, not even during the Pleistocene when the sea level was much lower (Sluys et al. 2007). Nevertheless, the distribution of Dugesia is repeated by the equally remarkable distribution in Asia and Australasia of portions of the camaenid land snails (Scott 1997; Cuezzo 2003), while the earliest nautiloids from Australia share major characteristics with Asiatic species (Stait and Burrett 1987). Evidently, similarity in these distributional patterns does not imply that they originated during the same period in geological history. Future studies on Dugesia from China and the Indo-Australian archipelago would be very interesting, as these may shed light on the biogeographic history of the region.

In our mitogenome analysis, trnT was the only tRNA that could not be automatically annotated by MITOS. However, through translating nucleotides of 12 PCGs to amino acid with Expacy (http://web.expasy.org/translate/), 63 sites of threonine, which is coded by ACN, were found. These results thus support the existence of trnT, which is required for the reading of the triplet of the genetic code (ACN). Therefore, we annotated trnT manually, based on homology comparisons with other species in the family Dugesiidae. Specifically, we aligned the complete mitogenome of D. ancoraria with three species belonging to family Dugesiidae, namely D. japonica, D. ryukyuensis and D. constrictiva and found a homologous sequence (60 bp) among these five species, which is particularly conserved at 5’ end (AGAA) and 3’ end (TTCTT). In addition, the position of trnT of all reported species belonging to Dugesiidae is very conserved and is situated between trnL2 and trnC. Interestingly, the putative trnT in D. ancoraria is also located between trnL2 and trnC, providing another line of evidence in support of the annotation. Furthermore, we predicted the secondary structure of trnT manually and presented this through RNAalifold WebServer (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi) and were surprised to find that the predicted result is not a typical cloverleaf structure with an absence of the DHU stem. By comparing the free energy between the two predicted structures, we found that the free energy with DHU stem (-1.20 kcal/mol) is higher than the one without DHU stem (-4.50 kcal/mol). The absence of the DHU stem in trnT also occurred in several other species of Tricladida, such as Dugesia japonica, D. ryukyuensis, Crenobia alpina, Obama sp. and Schmidtea mediterranea. (Sakai and Sakaizumi 2012; Solà et al. 2015; Ross et al. 2016). To the best of our knowledge, P. gracilis is the only Tricladida species reported to possess trnT with a DHU loop. Therefore, absence of the DHU stem in trnT could be a common phenomenon among species of the Tricladida.

In addition, some common features can be discovered in the mitochondrial gene order of the investigated species. Among the five species of Geoplanidae, locations of trnT are variable, in that transpositions of trnT occur in each of two adjacent species in the mitogenome tree, from B. kewense to P. ventrolineata. Besides, trnF is located at 3’ downstream of nad4, with the only exception being Crenobia alpina. The gene rearrangement of the maricolan Obrimoposthia wandeli differs considerably from species belonging to the suborder Continenticola. Therefore, transformation of gene order in O. wandeli to species of the Continenticola may require multiple rearrangements, including reversals, transpositions, and TDRL. Similarly, several TDRL events are required to go from the Geoplanoidea gene order to those of the Dugesiidae species. It is also noteworthy that two species (D. ancoraria and D. constrictiva) with identical mitochondrial gene order occur in two separate clades. Since gene rearrangements appear to be rare events that may not arise independently in separate lineages (Boore 1999), it is likely that the ancestor of Dugesia (indicated by a triangle in Fig. 4) may have had a pattern identical to that of D. ancoraria and D. constrictiva, implying that in the course of evolutionary history only a transposition of trnE occurred in D. japonica as well as transposition of trnN in D. ryukyuensis. However, the small number of species for which mitogenomic datasets are currently available, make it presently impossible to test this hypothesis.

A highly asymmetrical penis papilla with both a proximal as well as distal dorsal bumps is the most characteristic feature of Dugesia ancoraria. Similar bumps are known only from D. gibberosa Stocchino & Sluys, 2017. However, in D. gibberosa the penis papilla has a different, ventro-caudal orientation, while its ejaculatory duct opens terminally at the tip of the papilla, in contrast to the subterminal opening in D. ancoraria. Moreover, in D. gibberosa the bursal canal is surrounded by a very thick layer of circular muscle, while its ectal reinforcement extends more than halfway along the bursal canal. In contrast, the bursal canal musculature in D. ancoraria is thinner and the ectal reinforcement weakly developed. Furthermore, D. ancoraria and D. gibberosa are far removed from each other in the phylogenetic tree (Fig. 2), thus corroborating their separate taxonomic status.

Our molecular analyses, particularly the concatenated dataset, showed that D. ancoraria shares a sister-group relationship with Australian D. notogaea and Malaysian D. bengalensis. Morphologically, all three species have an asymmetrical penis papilla, with the dorsal lip being thicker than ventral lip, and a duct between seminal vesicle and ejaculatory duct. In addition, D. ancoraria and D. notogaea share the condition in which the oviducts open asymmetrically into the bursal canal. However, there are also clear differences between these three species. For example, in D. bengalensis and D. ancoraria, the ejaculatory duct has a subterminal opening at the tip of the penis papilla, whereas D. notogaea exhibits a terminal opening. In D. bengalensis and D. notogaea, the vasa deferentia open through the postero-lateral roof of the seminal vesicle, whereas in D. ancoraria the ducts open into the mid-lateral portion of the vesicle.

Although D. gibberosa is the only other species with two clear dorsal bumps on the penis papilla, there are a number of Dugesia species that deserve some comparison with D. ancoraria, viz., D. astrocheta Marcus, 1958, D. austroasiatica Kawakatsu, 1985, and D. tamilensis Kawakatsu, 1980. Dugesia astrocheta has a clear, proximal hunchback bump on its very asymmetrical penis papilla, while there is also some indication of a distal bump or bulge (cf. Sluys 2007, fig. 4A). But even when there is indeed such a distal bulge, the species differs in other details from D. ancoraria. For example, in D. ancoraria there is a relatively long duct interposed between the seminal vesicle and the diaphragm, whereas this duct is virtually absent in D. astrocheta. In D. austroasiatica there seems to be a rather flexible distal bulge on the dorsal lip of the penis papilla that receives the secretion of glands (cf. Kawakatsu et al. 1986, fig. 3). Apart from the absence of the hunchback, proximal penial bump in D. austroasiatica, there are also other differences which signal that it differs from D. ancoraria. For example, in the latter the oviducts open asymmetrically into the bursal canal, whereas D. austroasiatica has symmetrical oviducal openings. The penis papilla of D. tamilensis resembles that of D. ancoraria in that it is highly asymmetrical, with the ejaculatory duct opening also at the postero-ventral wall of the papilla, while its dorsal lip is provided also with a distal bulge. However, in this species the oviducts also open symmetrically into the bursal canal, in contrast to the asymmetrical oviducal openings in D. ancoraria.