Systematics of the family Plectopylidae in Vietnam with additional information on Chinese taxa (Gastropoda, Pulmonata, Stylommatophora)

Abstract Vietnamese species from the family Plectopylidae are revised based on the type specimens of all known taxa, more than 600 historical non-type museum lots, and almost 200 newly-collected samples. Altogether more than 7000 specimens were investigated. The revision has revealed that species diversity of the Vietnamese Plectopylidae was previously overestimated. Overall, thirteen species names (anterides Gude, 1909, bavayi Gude, 1901, congesta Gude, 1898, fallax Gude, 1909, gouldingi Gude, 1909, hirsuta Möllendorff, 1901, jovia Mabille, 1887, moellendorffi Gude, 1901, persimilis Gude, 1901, pilsbryana Gude, 1901, soror Gude, 1908, tenuis Gude, 1901, verecunda Gude, 1909) were synonymised with other species. In addition to these, Gudeodiscus hemmeni sp. n. and Gudeodiscus messageri raheemi ssp. n. are described from north-western Vietnam. Sixteen species and two subspecies are recognized from Vietnam. The reproductive anatomy of eight taxa is described. Based on anatomical information, Halongella gen. n. is erected to include Plectopylis schlumbergeri and Plectopylis fruhstorferi. Additionally, the genus Gudeodiscus is subdivided into two subgenera (Gudeodiscus and Veludiscus subgen. n.) on the basis of the morphology of the reproductive anatomy and the radula. The Chinese Gudeodiscus phlyarius werneri Páll-Gergely, 2013 is moved to synonymy of Gudeodiscus phlyarius. A spermatophore was found in the organ situated next to the gametolytic sac in one specimen. This suggests that this organ in the Plectopylidae is a diverticulum. Statistically significant evidence is presented for the presence of calcareous hook-like granules inside the penis being associated with the absence of embryos in the uterus in four genera. This suggests that these probably play a role in mating periods before disappearing when embryos develop. Sicradiscus mansuyi is reported from China for the first time.


Introduction
At present, 477 species and subspecies in 22 families of terrestrial pulmonates are known from Vietnam (Schileyko 2011). As in other southeast Asian gastropods, most of these (77%) were described between 1880 and 1920 with poor locality data and based on shell characters only. Several species were described by examining only a single shell. Internal anatomy and exact collecting locality have been documented only for a few taxa. Accordingly, the systematics of most Vietnamese land snails remains questionable. Without accurate knowledge on their distribution and taxonomy, the recognition of possible threats and the subsequent establishment of appropriate conservation measures of these populations are impossible.
The Plectopylidae are currently the fifth largest pulmonate family in Vietnam with 28 species, after the Camaenidae s.l. (=Camaenidae and Bradybaenidae: 127 sp., 9 ssp.), Clausiliidae (84 sp., 10 ssp.), Ariophantidae (68 sp., 3 ssp.) and Streptaxidae (47 sp., 2 ssp.) (Schileyko 2011). The Plectopylidae are a group of medium sized (5-35 mm), usually flat, sinistral or dextral species, which have internal lamellae and plicae on both the palatal and parietal walls. This family is currently included within the Plectopyloidea together with the south Asian family Corillidae Pilsbry, 1905 and the south-west family African Sculptariidae Degner, 1923 (Bouchet andRocroi 2005). However, Schileyko (1999) classified Sculptariidae in the superfamily Acavoidea. Plectopylidae differ from the probably closest group, which are the Corillidae with one or two vertical lamellae on the parietal wall. The mainly Sri Lankan Corillidae only have horizontal plicae. Plectopylidae have a wide distribution from northeastern India through the majority of southeast Asia to Peninsular Malaysia, northern Vietnam and southern Japan (Páll-Gergely and Hunyadi 2013 and references therein).
For the present revision of the Vietnamese Plectopylidae, we examined all the type specimens as well as many available non-type material deposited in public institutions. All samples deposited in HNHM, NHMSB, NHM, MNHN, NHMW, SMF and SNM were investigated. Some "problematic" samples were loaned and identified from RBINS and USNM. Material (usually with GPS data) obtained from the following private collections were investigated: András Hunyadi, Jozef Grego, Christa and Jens Hemmen, Kenji Ohara, Jamen Uiriamu Otani and Wim Maassen. Altogether approximately two hundred samples with exact locality data were examined. Fischer and Dautzenberg (1904) mentioned two names (Plectopylis anoplon and simulans) from Vietnam but presented no formal descriptions. Although listed by Thanh (2008), these nomen nuda cannot be assigned to species. Gude's material is deposited in NHM, and most samples from Lieutenant Colonel Messager are housed in MNHN. Messager probably sent only a few shells to Gude, who published on these in 1909. The six species described by Gude (1909) are problematic. Investigation of these specimens including Messager's original material allowed us to gain a better understanding of species boundaries based on morphological gaps in continuously varying shell characters.
Here we present the outcome of systematic revision of Vietnamese Plectopylidae (see summary in Table 1) with reproductive anatomy and radula morphology of eight species. Additionally, we publish information on the radula of fifteen Chinese species. The genus Gudeodiscus is divided into two subgenera based on anatomical and radula information of Chinese and Vietnamese species.
The palatal plicae can be observed from the interior and exterior view. This is indicated in the figure captions in all cases. If enough shell material was available, a shell fragment with the palatal plicae was broken out and the lamellae were observed directly (interior view). If shell material was limited, the plicae are figured as they were  Gude's (1899c) revision, in the original description (in case of species described after Gude 1899c), and in this study. Synonymies are also indicated. Valid taxa with bold italic.
(sub)species section in Gude 1899c
To demonstrate the continuous variation of shell heights and diameters across Plectopylis anterides/gouldingi, P. fallax, and P. fallax var. major specimens (synonyms of Gudeodiscus phlyarius; Figure 16), we randomly selected a few samples which can be assigned to those taxa. The buccal mass was removed and soaked in 2 molar KOH solution for 5 hours before extracting the radula, which was preserved in 70% ethanol. Radulae were directly observed without coating under a low vacuum SEM (Miniscope TM-1000, Hitachi High-Technologies, Tokyo).

Taxonomic treatment
This revision is based on morphology by examination of specimens and literature. Thus the present taxa are defined based on their morphological differences. The present species are hypothesized as species defined by the biological species concept (Mayr 1942), although evidence for differences in sympatry was not always available within the relevant species group. Table 2 shows sympatric species pairs. No specimens were found that show transitional characters between sympatric species. This suggests that these are biological species reproductively isolated from each other.
Previously recognized taxa are synonymized when their differences between traditionally recognized species (often present as only a few individuals) are considered to be very minor. Sometimes, these differences (mainly in the morphology of the plicae and lamellae) show a geographical pattern. If these minor differences fall within the range of the species' morphological diversity, the taxa are synonymized. leg collected by ex from the collection of D shell diameter H shell height

Radula information
Information on the radula morphology of Chinese Plectopylidae species has never been published. To provide a comprehensive basis of the radula morphology of Vietnamese species, we publish images of the radula of some Chinese species as well. The key characters of the radula (size of the central tooth in relation to the ectocone of the first lateral, the shape of the mesocone of the first lateral and the morphology of the marginals) are compiled in Table 6. The overall morphology of the radula was similar in all species. The lateral teeth are arranged along straight rows, whereas the marginals stand in oblique rows. The distinction between the last laterals and the first marginals is not easy, especially in those specimens in which their morphology (bi-or tricuspid) does not differ. Therefore, the data on the number of laterals and marginals are only guidelines. Type species. Plectopylis phlyaria Mabille, 1887, by original designation. Included taxa. Subgenus Gudeodiscus and subgenus Veludiscus subgen. n. Diagnosis. Shell rarely small, usually middle sized or large, dextral, body whorl rounded, without periostracal folds on the "upper keel" of the whorls. The whole protoconch is usually very finely, regularly ribbed (see Figure 10A). The only known exceptions are Gudeodiscus villedaryi (see Figure 10B) and G. dautzenbergi. Teleoconch usually has a reticulated sculpture; more prominent on the dorsal side; sometimes with very small periostracal filaments, but these are always arranged radially, never in spiral lines. A short apertural fold is present in the majority of the species. Palatal plicae usually 6, sometimes 5 or 7, they are usually free, very rarely connected by a ridge. Middle palatal plicae can be horizontal, oblique or almost vertical, they are usually depressed Table 6. (G.) villedaryi 9 10 as large as or slightly smaller than the ectocone of the first lateral rhomboid, pointed bicuspid or tricuspid with blunt inner cusp and shallow incision between the inner two cusps H. fruhstorferi 8 12 much smaller than the ectocone of the first lateral slender rhomboid mostly bicuspid, some of them tricuspid with blunt inner cusps H. schlumbergeri 10 14 smaller than the ectocone of the first lateral oval bicuspid or tricuspid with blunt inner cusp and shallow incision between the inner two cusps Sic. invius 7 8 as large as or larger than the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them Sic. mansuyi 8 10 as large as the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them, some of them quadricuspid Sic. schistoptychia Sin. fimbriosa 10 15 larger than the ectocone of the first lateral slender with concave inner line tricuspid with pointed cusps and deep incision between them Sin. jugatoria 9 12 as large as the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them Sin. murata 8 12 as large as the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them Sin. reserata azona 11 14 as large as the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them Sin. stenochila 8 13 as large as the ectocone of the first lateral slender with parallel, straight margins and pointed end tricuspid with pointed cusps and deep incision between them "Z" or "V"-shaped. The first plica is always straight and parallel with the suture, the last is slightly curved or oblique. On the parietal wall there are two vertical lamellae or the anterior one is missing or dissolved into small denticles or parallel horizontal plicae. Usually horizontal plicae are visible above and below the anterior lamella, near the sutures.
Penial caecum usually present (very rarely absent). Penis internally with longitudinal folds; the middle or proximal portion of the penis can have transverse or reticulated sculpture; the longitudinal folds are thickened on the apical part of the penis and form "pockets", each of which holds a calcareous, usually hook-or claw-like translucent granule; these granules are probably present seasonally when the snails are reproductively active and disappear when embryos develop in the uterus; the pockets stand in one row or rarely in two rows on the opened penis wall. Epiphallus with simple internal longitudinal folds.
Differential diagnosis. The body whorl of the species belonging to Sinicola is keeled or shouldered, often with flat, deciduous periostracal folds arranged in one row on the keel. In contrast, all Gudeodiscus species have rounded body whorl and never have periostracal folds arranged in a spiral line. Moreover, in Sinicola there are no "pockets" on the inner wall of the penis. The shells of Halongella gen. n. are indistinguishable from those of Gudeodiscus. Halongella gen. n. species have parallel, longitudinal folds on the inner wall of the penis with tiny, flat calcareous granules between the folds, all along the penis; there are no determined "pockets" for the granules at the apical part of the penis, which are so characteristic for Gudeodiscus. Additionally, the longitudinal folds inside the epiphallus of Halongella gen. n. species have characteristic transverse projections which overlap with those of neighbouring folds. In contrast, Gudeodiscus species have parallel folds on the inner wall of the epiphallus. Additionally, most anatomically examined Gudeodiscus specimens had a penial caecum, which is missing in both Halongella gen. n. species. See also under Sicradiscus.

Subgenus Gudeodiscus Páll-Gergely, 2013
Diagnosis. Shell indistinguishable from Gudeodiscus (Veludiscus) subgen. n. Anatomy: The epiphallus has a somewhat thickened proximal part; retractor muscle simple, inserts on the distal end of the penial caecum, or if it is missing, than on the distal end of the penis (at the penis-epiphallus transition). Radula: central tooth usually as large as or slightly larger than the ectocone of the first lateral; mesocone of the first lateral is moderately wide, in most cases has parallel edges. Marginals usually tricuspid with rather pointed inner cusp and rather deep incision between the inner two cusps.
Differential diagnosis. Gudeodiscus messageri is larger than G. anceyi and lacks the apertural fold and spiral lines on the ventral surface of the shell. Gudeodiscus anceyi is smaller than typical G. phlyarius, has stronger spiral lines, and has no horizontal plica under the lamellae, which are present in most populations assigned to G. phlyarius. The G. phlyarius populations living near the Chinese border (typical anterides, gouldingi, fallax, verecunda) are usually larger than G. anceyi and they often lack the apertural fold and the spiral lines on the ventral side of the shell. For differences with G. hemmeni sp. n. and Sicradiscus mansuyi, see under those species.
Intraspecific diversity. Relatively low; shell characters, namely the size and general shell and aperture shape are rather stable. The morphology of the palatal plicae shows some diversity. The species is easily recognisable and can be separated from other plectopylid species without major problems.
Distribution (see Figure 40): We have newly-collected material only from Bắc Kạn Province. The species was previously recorded from That Khé (Lạng Sơn Province) and Nac Ri (Hà Giang Province) (Gude 1901a, see also  Types examined. Tonkin, Muong-Kong, leg. Messager, NHMUK 1922.8.29.59 (syntype, Figure 2F). Diagnosis. Shell very small to small, discoid, polished with very weak apertural rim, weak or missing callus and without apertural fold. Parietal wall with two lamellae and an upper and a lower horizontal plica; the plicae can be free from the anterior lamella or in contact with it; palatal plicae straight, parallel, horizontal, sometimes connected with a slight ridge ( Figures 15E-G).
Measurements ( Differential diagnosis. The Chinese Gudeodiscus yunnanensis has a similar shell shape but possesses only one vertical parietal lamella (the anterior one is absent). The two species can be separated only the basis of the presence or absence of the anterior lamella. In G. soosi and in most specimens of G. multispira, few denticles are present between the upper and lower plicae, at the place of the anterior lamella. Moreover, G. multispira has a greater number of whorls and the last whorl is wider in relation to the previous one than in G. cyrtochilus. Gudeodiscus infralevis is larger with a more elevated spire, stronger apertural lip and usually a weak apertural fold. See also under G. fischeri. Intraspecific diversity. Low; shell characters rather stable. The parietal plicae and lamellae and their respective position (reaching each other or not) show some diversity within the species. The palatal plicae are not variable, but in some shells they are connected to each other with a ridge, whereas in others they are free. It is possible that mature specimens tend to have a connection between the plicae. The species is easily recognisable and can be separated from other plectopylid species without major problems.
Distribution (see Figure 41): The species was described from "Muong-Kong" (=Mường Khương, Lào Cai Province; see Figure 39). Material is noted from northeast of this locality, from northern Hà Giang Province and eastern parts of Yunnan Province (China) (see Páll-Gergely and Hunyadi 2013).
Differential diagnosis. Gudeodiscus villedaryi, which is probably the closest relative, differs from G. dautzenbergi by the presence of an additional horizontal parietal plica under the vertical lamellae, near the suture. Distinguishing G. dautzenbergi from some similar looking populations of G. villedaryi is impossible without breaking the shell and observing the parietal plicae. Most populations of G. villedaryi however, have a sharp periumbilical keel, which always absent in G. dautzenbergi (see also Remarks under G. villedaryi). Gudeodiscus dautzenbergi is flatter and more widely umbilicated than G. giardi. The latter species has a domed shell, thinner shell wall and thicker peristome. For comparisons with Halongella schlumbergeri, see under that species. Distinguishing G. dautzenbergi from H. schlumbergeri requires experience, but is possible without breaking the shell on the basis of the formation of the peristome and the apertural fold ( Figures 9K-N).
Intraspecific diversity. Low; shell characters stable. Distribution (see Figure 40): This species as well as Plectopylis persimilis (synonym of Gudeodiscus dautzenbergi) were described from That-Khé (northern Lạng Sơn Province) (see Figure 39). Our newly-collected material is from the border region of the Thái Nguyên and Bắc Kạn provinces.
Remarks. The holotype of Plectopylis persimilis and that of Plectopylis dautzenbergi do not show significant differences in terms of shell shape, size, aperture shape and the formation of the plicae and lamellae; therefore we synonymise Plectopylis persimilis with P. dautzenbergi. These two species were described in the same publication , therefore the name introduced earlier (dautzenbergi, page 198) is considered a senior synonym.
Gudeodiscus dautzenbergi and G. villedaryi are separated here on the basis of the presence or absence of a lower plica, although the two species may be conspecific. More information is necessary to clarify the distinctness of G. dautzenbergi.

Gudeodiscus
Diagnosis. Shell small to medium-sized, with smooth basal and usually finely ribbed apical surface (in some populations also smooth and glossy, see Figure 2E); shell usually flat, or with very slightly elevated spire, or only the protoconch is elevated from the dorsal surface; callus and apertural fold (if present) weak (Figures 9P-Q). Parietal wall with two lamellae (the anterior is exceptionally dissolved into small denticles); middle palatal plicae oblique, depressed Z or L-shaped, they are free or sometimes connected to each other ( Figures 15H-N).
Differential diagnosis. Gudeodiscus cyrtochilus is smaller than G. fischeri, it has a narrower umbilicus, more regularly growing whorls (the last whorl is only slightly wider than the penultimate one), a shorter lower horizontal parietal plica and no apertural fold. The Chinese G. multispira and G. soosi are also smaller, have a greater number of densely-coiled whorls and at the position of the anterior lamella there are usually 2-4 clearly separated denticles (see also Remarks). In some populations of G. multispira the denticles are missing so that only the posterior lamella is present. Gudeodiscus yunnanensis has no anterior lamella, just a curved single lamella (homologous with the posterior lamella). Gudeodiscus eroessi never has an apertural fold and its anterior lamella is dissolved into small denticles, or missing. Gudeodiscus infralevis and G. suprafilaris have a more elevated spire, narrower umbilicus and rather straight, horizontal, parallel plicae.
Intraspecific diversity. The variability is quite large in terms of shell size and shape, sculpture, strength of the callus and apertural fold and the formation of parietal plicae and lamellae. The combination of weak callus and apertural fold and the "nautiliform" shape helps in the identification of the species. See also Table 7.
The penis is a cylindrical tube with several longitudinal, parallel folds on the inner wall; there are pockets formed by some of these folds; in the wall of the opened penis the series of pockets are arranged along a bell-shaped line ( Figure 28D); there were calcareous hooks within the pockets of "Specimen1"; the base of the hooks were elongated, they lay within the pockets, whereas the tip portion projects out of the pockets ( Figure 30E); epiphallus as long as the penis, with few parallel folds in the lumen (Figure 29D); distal portion of the penis and the proximal part of the epiphallus are connected with weak membrane; more closely to the genital opening these two organs are more stronger connected; penial caecum tapers toward the end, it is about a quarter as long as the penis; its inner wall with irregular folds arranged in longitudinal lines, with calcareous granules in between (mainly at the distal end); retractor muscle attaching on the apical part of the penial caecum is approximately as long as the caecum; there is an additional retractor muscle on the proximal part of the penis. Vagina is thickened and forms a "vaginal bulb", which is attached to the body wall with several thin ligaments; inner wall of the vaginal bulb and the distal part of the vagina with well-developed, longitudinal, serrulate folds ( Figure 31D); stem of the gametolytic sac is long and slim; it is attached hardly to the spermoviduct; diverticulum well-developed, free; the diverticulum of the specimen from the Ba Bể Nat. Park contained three long, slightly C-shaped spermatophores; the proximal side of the spermatophores were damaged, thus they might have been connected; spermoviductus slim and long.
Besides the presence or absence of embryos and calcareous penial hooks between the two specimens the only notable difference is the longer retractor muscle in "Speci-men2" than in "Specimen1", but the taxonomic value of this character is unknown.
Remarks. Some samples from the Ba Bể Nat. Park (Vn10-28A, 20090517A, 2011/91, 2011/96) are identical with the type specimen of Plectopylis tenuis described from Cho Ra (see Figure 39). This town is situated approximately 7 km from the locality of our recent material. Some 3 km north of our tenuis localities there is a population (20090515C) which agrees with tenuis in every shell character except that the anterior parietal lamella and the lower horizontal plica are connected (typical in fischeri). Since no other shell characters are known to be different between tenuis and fischeri, and other populations of fischeri show relatively large variability in terms of several shell characters, we synonymize Plectopylis tenuis with P. fischeri.
The shells collected 9.8 km north of Hà Giang are relatively large and thick-walled, have the anterior lamella dissolved into 3-4 denticles, and have strong apertural denticle and callus ( Figures 3A, 15L-M). The shells collected at Đồng Tiến are small and very glossy in appearance ( Figure 2E).  Diagnosis. Shell small to medium-sized, yellowish or mustard-coloured, glossy, with slowly increasing whorls, deep umbilicus, domed dorsal side; thin apertural lip and well-developed apertural fold. Parietal wall with two parietal lamellae; the anterior one is connected to the lower plica; middle palatal plicae oblique, depressed Z-shaped (Figures 13E-K).
Differential diagnosis. The glossy, dark yellow shell, the characteristic apertural fold and shell shape makes this species easily distinguishable from most congeners. Gudeodiscus francoisi has a smoother shell, weaker apertural lip and more regular whorls than G. giardi giardi. In the type locality of francoisi (Déo-Ma-Phuc, see Figure  39) the species lives together with G. giardi giardi. In some cases the two species can be hardly distinguished, especially in the case of subadult giardi specimens which cannot be easily distinguished from francoisi. The possibility of hybridisation in that locality cannot be excluded; however specimens from other localities are easily distinguishable.
Intraspecific diversity. The species shows little intraspecific variability in terms of shell characters. The "lepida-like" shells are considered to the results of abnormal growth.
Distribution (see Figure 42): Newly-collected material from Cao Bằng and Bắc Kạn Provinces was examined. There is a single shell which is identical to the holotype of Plectopylis lepida and is labelled as being collected from Hạ Long Bay, but this collection locality is probably incorrect.
Remarks. Gudeodiscus bavayi is a synonym of G. francoisi. The two holotypes are identical in shell shape and arrangement of the inner lamellae. The only difference is that the holotype of G. francoisi lacks an apertural fold because it is a subadult shell. Other shells collected from the type locality are identical with the holotype of Plectopylis bavayi. Plectopylis lepida was described on the basis of a single shell. During the revision of the Vietnamese Plectopylidae material in the MNHN, we found a single shell (Baie d' Along, coll. Staadt, 1969, MNHN-IM-2012-2311 which is identical in shell shape and plication with the holotype of lepida. These two shells differ from G. francoisi only by the absence of the posterior lamella and the weak apertural fold. The absence of the posterior lamella is probably the result of unusual development, which is also visible in a specimen of G. suprafilaris (see under that species). The weak apertural fold can be explained by subadult stages of these shells. Since no other shell characters distinguish Plectopylis lepida and G. francoisi, the former is treated as a junior synonym of Plectopylis francoisi.
Diagnosis. Shell small to large, brownish (some Chinese populations are small and yellow, translucent), usually finely reticulated (resulting in a matt surface), umbilicus deep, dorsal side domed; apertural lip, callus and apertural fold very well-developed ( Figure 9I). Parietal wall with two lamellae; the anterior one is usually connected to the lower plica; middle palatal plicae short, depressed Z-shaped, or almost vertical, sometimes connected to each other ( Figures 13L-U).
Measurements ( Differential diagnosis. This species is most similar to G. francoisi. For comparisons, see under that species. Gudeodiscus dautzenbergi is larger, flatter, has wider umbilicus, a weaker apertural lip and the lower end of the anterior lamella is very much elongated anteriorly. Gudeodiscus villedaryi is also flatter and most populations have a keel around the umbilicus and an additional long plica below the parietal lamellae. Gudeodiscus phlyarius is usually flatter, has a wider umbilicus, slimmer peristome and lower callus. Most specimens of G. phlyarius have separated anterior lamella and lower plica, whereas these are always connected in G. giardi giardi. Typical Plectopylis verecunda shells (synonym of G. phlyarius) also have an elevated spire, but their shell shape is rather conical, whereas it is usually domed (rounded) in G. giardi.
Penis very short, almost ball-like; penis wall conspicuously thickened, its inner surface is characterized by transversal lines at the proximal part and longitudinal pockets in the distal part, arranged in a straight row ( Figure 28B); there are some calcareous, claw-like objects in the pockets; the claws have a wide, rounded basal part which is found within the pockets, and the short, hook-like part hangs out of the pockets; the base had a granulated surface, probably to provide a better attachment to wall of the pockets, whereas the tip was smooth ( Figure 30D); epiphallus C-shaped, longer than the penis; its inner wall with three longitudinal parallel folds ( Figure 29E); penis and epiphallus connected with weak membrane; penial caecum approximately as long as the penis; it has low tubercles on the inner wall and small calcareous rounded granules on each tubercle; retractor muscle attaches on the distal part of the penial caecum; it is longer and wider than the caecum; vas deferens convoluted near the vagina. Vagina very thick and long, it is attached to the body wall with several thin ligaments; one side of the vaginal bulb with very much thickened wall, the other side with thin, almost translucent wall, internally with fine, irregular, reticulated sculpture; inner wall of the distal portion of the vagina with well-developed, rather irregular transversal folds (Figure 32C); gametolytic sac and diverticulum slender, they are nearly the same length.
Radula. See Table 6 and Figures 35A-C. Distribution (see Figure 42): Newly-collected material was examined from Cao Bằng Province and the northern part of Lạng Sơn Province. The localities of "Col de Nuages" and "Halong Bay" are probably erroneous. This species is also known from the western part of Guangxi, China (Páll-Gergely and Hunyadi 2013).
Remarks. Plectopylis congesta Gude, 1899 was described without exact locality data. Some shells from populations in southern Cao Bằng and northern Lạng Sơn prefectures (Vn10-69;2011/84, 2011/83, 2011/82, 2011) resemble the holotype of P. congesta on the basis of relatively weak peristome and callus, weak (low) posterior lamella and the anterior lamella which is fused to the upper parietal plica. These populations however, falls within the morphological range of the very variable Gudeodiscus giardi giardi, therefore P. congesta is here synonymised with G. giardi giardi.
The genital anatomy of a Chinese specimen of Gudeodiscus giardi giardi was described by Páll-Gergely and Asami (2014). The only notable difference between the Chinese and Vietnamese specimens is the much longer penis in the Chinese individual. It seems that the long, slender, proximal portion of the penis visible in the Chinese specimen is entirely missing in the Vietnamese one. Diagnosis. Shell small, with slightly elevated spire, characteristically shaped aperture having wide upper sinulus and small apertural fold ( Figure 9F); parietal wall with two lamellae and horizontal plicae above and below; palatal plicae depressed Z-shaped; free from each other, or connected to each other with a ridge (Figures 11G-J).
Description. Shell very small to small, light brown to chocolate brown, with slightly elevated spire, consists of 5.25-5.5 whorls; suture relatively shallow, especially at the first 3-4 whorls; protoconch (2.25-2.5 whorls) glossy, very finely, regularly ribbed, but the ribs are sometimes hardly visible, they are more prominent at the upper part of the whorls, close to the suture; teleoconch without notable spiral lines, very finely regularly ribbed; sculpture strength equal on ventral and dorsal side; umbilicus narrow and deep; aperture with widened upper part (sinulus), apertural lip whitish, thin, slightly expanded but not reflexed; apertural denticle (fold) always present, very small, free from the callus or connected to it.
Two specimens were opened. Parietal side with a stronger anterior lamella with anteriorly widened lower part, and a slimmer posterior lamella; shorter upper and longer lower horizontal plicae free from the anterior lamella, the lower one a bit extends beyond the anterior lamella in the anterior direction. Palatal side with six plicae; first and last are straight, the others are depressed Z-shaped and are connected with a ridge.
Differential diagnosis. Gudeodiscus hemmeni sp. n. differs from most G. phlyarius populations by the smaller shell, shorter denticle (fold) in the aperture, thinner apertural lip, the wider and reflexed apertural rim, the wide upper sinus of the aperture, lack of spiral lines in the sculpture and narrower umbilicus. Gudeodiscus anceyi is usually smaller, has a longer apertural fold, prominent spiral sculpture, a weaker callus and differently shaped aperture.
In all localities, Gudeodiscus hemmeni sp. n. lives sympatrically with G. messageri raheemi ssp. n., which is much larger, lacks the apertural fold, and usually has an anterior lamella which is dissolved into small denticles. Intraspecific diversity. Low; shell characters are stable, although only a few shells are known.
Etymology. The new species is dedicated to Jens Hemmen , malacologist and much-valued friend, who contributed to our revision by providing shell and ethanol-preserved material. Distribution (see Figure 43). The new species is known from few locations in south-eastern Sơn La province.  Figure 3D); Tonkin, Quang-Huyen, leg. Mansuy, MNHN 24585 (holotype of soror, Figure 3E). Diagnosis. Shell small, solid, discoid, with elevated spire, relatively deep umbilicus; relatively thin apertural lip and rather parallel, thick, straight palatal plicae. See also under remarks.
Differential diagnosis. Our knowledge on the intraspecific variety of the species is very limited (see Remarks). It seems that the thick, rather horizontal palatal plicae, the strong basal sculpture and the elevated spire distinguishes the species from the similar species (Gudeodiscus eroessi, G. multispira, G. soosi, G. yunnanensis, G. cyrtochilus and G. fischeri). The shell and aperture shape suggest that the closest relatives are G. fischeri and G. suprafilaris (see comparisons under those species).
Intraspecific diversity. Plectopylis infralevis and P. soror are considered as conspecific (see Remarks). Only the holotypes of these taxa are known, therefore our knowledge on the intraspecific variability is limited.
Remarks. Only the holotypes of Plectopylis infralevis and P. soror are known. The notable differences between these two shells are the stronger sculpture, slightly shouldered body whorl and small apertural fold in soror. Additionally, there are three lamellae in infralevis versus only one in soror. The three vertical lamellae in the holotype of infralevis is possibly the result of abnormal development. No other species of Plectopylidae has three lamellae. Similar abnormal shells have been reported in Gudeodiscus giardi (see Gude 1908). Consequently, we do not know what the characteristic type of parietal lamellae in this species is (=one or two). The differences between the two specimens suggest only intraspecific variance. Unfortunately we have no freshly-collected material of these two forms, but because of the high similarity between the two holotypes and same type locality we here synonymise soror with infralevis. These two names were published in the same paper (Gude 1908), but infralevis was described earlier in terms of page numbers. (Gude, 1909) Diagnosis. Shell small to medium-sized, with slightly elevated spire, dorsal surface somewhat domed; aperture almost circular, apertural fold missing; callus rather blunt and only slightly curved. Parietal wall with two lamellae (the anterior lamella may be dissolved into small denticles); lower parietal plica free or connected to the anterior lamella; palatal plicae oblique, or depressed Z-shaped, usually in contact with each other.
Measurements (in mm). D = 12.75-18.5 (according to the original description). Differential diagnosis. Gudeodiscus messageri messageri inhabits northern Vietnam and in many museum samples it is mixed with Plectopylis gouldingi or Plectopylis fallax (synonyms of G. phlyarius). These two forms have flat shells with a sharp and angled callus, and sometimes with an apertural denticle. Also, the aperture of G. messageri is rather rounded, whereas it is rather elongated in those populations of G. phlyarius (Figures 9D: phlyarius, Figure 9E: messageri). This allows G. messageri and G. phlyarius to be distinguished without breaking the shell. The lower parietal plica, which does not extend beyond the anterior lamella in the anterior direction, is characteristic of G. messageri messageri (see also Remarks), but almost always extends in "P. fallax" and "P. gouldingi". "Plectopylis verecunda" (synonym of G. phlyarius) and typical G. phlyarius always have a strong apertural fold. Moreover, the lower parietal plica of the latter usually extends beyond the anterior lamella in the anterior direction. For comparison with G. messageri raheemi ssp. n., see there.
Intrasubpecific diversity. Low; the shell size, and the relationship between the lower parietal plica and the anterior lamella show some variability (see remarks). The shell and aperture shape are stable characters.
Distribution (see Figure 43): Only museum material was available for study, which suggested that this species is located along the Chinese (Yunnan) border.

Gudeodiscus (Gudeodiscus) messageri raheemi
Description. Shell medium in size, light to dark brown or dark yellowish, sometimes almost flat but usually with slightly elevated spire, consists of 6.25-6.75 whorls; suture relatively shallow; protoconch (2.5-2.75 whorls) glossy, very finely, regularly ribbed; teleoconch very finely, rather irregularly ribbed, spiral lines visible mainly at the dorsal side where sometimes they are as strong as the ribs (resulting in a reticulated surface), in some specimens however hardly any spiral lines are visible; sculpture weaker on the ventral side but within the umbilicus are as strong as on the dorsal side; umbilicus relatively narrow and deep; aperture wide with whitish or light brown, thickened and reflexed apertural rim; callus slightly S-shaped, well-developed, with upper and with or without lower canal between the ends of callus and the apertural lip; apertural fold always missing.
More than ten specimens were opened belonging to different populations. Parietal side with two lamellae and upper and lower horizontal plicae above and below the anterior lamella; the lower plica usually extends beyond the anterior lamella in the anterior direction; in some populations the anterior lamella (or only the upper part of the lamella) is dissolved into several denticles. Palatal wall with six plicae; first and last are short and relatively straight, the four middle plicae are usually depressed Z-shaped and in many cases connected to each other with a ridge.
Measurements ( Differential diagnosis. The lower parietal plica extends beyond the anterior lamella in the anterior direction, which is extremely rarely the case in the nominotypical subspecies. The anterior lamella was dissolved into small denticles in many samples, which has never been observed in the nominotypical subspecies ( Figures 12N-Q: messageri, 12R-V: raheemi ssp. n.). The umbilicus of the new subspecies is narrower, it has more rounded whorls and a sharper, more angled callus, than in most samples of Gudeodiscus messageri messageri.
Gudeodiscus messageri raheemi ssp. n. lives sympatrically with an atypical form of G. phlyarius in Ninh Bình Province (see under G. phlyarius). Gudeodiscus phlyarius is flat and has an apertural fold, whereas G. messageri raheemi ssp. n. has somewhat elevated spire and always lacks the apertural fold. See also under G. hemmeni sp. n.
Intrasubspecific diversity. Relatively variable; the colour, spire height, size and morphology of the palatal and parietal lamellae and plicae show considerable variability (see Table 8).
Description of the genitalia. Two specimens were anatomically examined. Both specimens had embryos developing in their uterus.  (Figures 28E, 29G).
Penis relatively short and slim, attached to the slightly shorter epiphallus by weak fibres; penis internally with longitudinal folds; the folds are more elevated in the distal part of the penis and they from characteristic "pockets" ( Figure 28E); the pockets are arranged in two rows, the upper row (closer the distal end of the penis) is slightly curved on the opened penial wall, but the lower row follows a a wavy line with two peaks; epiphallus have longitudinal folds on the inner wall; penial caecum long; "Specimen1" had two times longer caecum than "Specimen2"; internally with small hollows arranged in longitudinal lines ( Figure 29F); "Specimen2" had a few elongated and globular calcareous granules within the hollows ( Figure  29G); retractor muscle very long and slim, attaches on the distal end of the penial caecum; vas deferens very long. Vagina extremely long, cylindrical in "Specimen1" and with well-developed vaginal bulb in "Specimen2"; inner wall of the vagina with 6-8 low, parallel or converging folds ( Figure 31B); gametolytic sac and diverticulum of the same length, both relatively slim, although the gametolytic sac is a bit swollen.
Radula. See Table 6 and Figures 35D-F. Etymology. The new subspecies is dedicated to and named after our colleague and much-valued friend, Dinarzarde Raheem.
Diagnosis. The species is very variable in terms of shell characters (spire height, presence/absence of the apertural fold, aperture shape, morphology of the parietal and palatal plicae and lamellae, fine morphology of the periostracum folds) between and within traditionally recognized species which are synonymized here. Therefore, it is impossible to give a general diagnosis.
Intrasubspecific diversity. Extremely large.  (Figures 21, 28A); Penis rather spindle-shaped, very much thickened in the middle; internally with a fine papillated/reticulated structure (proximal part) which gradually becomes a laterally folded structure with flat calcareous granules between the folds; pockets are ar-ranged in a rather straight line; epiphallus much shorter than penis, thickest at the penis-epiphallus transition, slowly becoming slimmer towards the vas deferens; penis and epiphallus connected with weak muscle fibres; penial caecum absent in one of the specimens and very small in the other; retractor muscle thick, short, inserts on the small penial caecum (or on the penis-epiphallus transition of the other specimen); vas deferens very long; the proximal section curves within a translucent, straight tube, most convolutions occurring proximally to the vaginal bulb, before becoming a solid, thick tube (until the sperm-oviduct). Vagina long, centrally with well-developed vaginal bulb; vaginal bulb thick-walled, internally with fine reticulated sculpture; distal part of the vagina internally with low, dense, transversal folds; gametolytic sac and diverticulum long, of equal length, extending in parallel; gametolytic sac spindle-shaped, diverticulum of equal thickness throughout.
typical phlyarius: Two specimens were anatomically examined, both contained a few embryos at an early developmental state. Localities: Lạng Sơn Province, ca.  Figure 31C).
Penis spindle-shaped with thickened middle section; internally with elongated folds of various thickness; this internal ribbed surface also continues in the small penial caecum; retractor muscle short, inserts on the penial caecum; epiphallus shorter and much slimmer than the penis; distally the penis and proximal part of epiphallus bound with connective tissue; vas deferens very long, proximally simple, slim, curved centrally and covered with a sheath distally simple and thickened. Vagina long with well-developed central vaginal bulb; internally the proximal part of the bulb is almost smooth; this sculpture changes to parallelly folded structure in distal direction ( Figure  31C); the distal part of the vagina is strongly folded; gametolytic sac and diverticulum of equal length, both being relatively short. Radula. See Table 6 and Figures 35J-L. Distribution (see Figure 43). The populations assigned to Gudeodiscus phlyarius inhabit several regions of northern Vietnam (Lạng Sơn, Cao Bằng, Ninh Bình, and along the border region with the Chinese Yunnan Province) and the Chinese Guangxi. A single shell of typical P. fallax Gude, 1909  Remarks. Gudeodiscus phlyarius and taxa of similar appearance are one of the most problematical groups in the Plectopylidae. Gude (1909) described six species (anterides, cyrtochila, fallax, gouldingi, messageri, verecunda) from the border region of northern Vietnam with the Chinese Yunnan Province. One species, Plectopylis cyrtochila differs from the rest of the species by the smooth, lenticular shell and week peristome and callus. Therefore, it is discussed separately, under the name G. cyrtochilus. In face of the obvious similarities between the remaining five species, Plectopylis messageri and P. fallax were only compared with P. moellendorffi, and P. verecunda was compared with P. messageri. The shell characters of P. anterides and P. gouldingi were only compared with each other. Shells having transitional characters were explained by hybrid origin. Gude (1909) mentions that a specimen of messageri from Pac-Kha might be a hybrid with moellendorffi, and another specimen from the same locality was believed to be a hybrid of anterides and gouldingi. The shell characters distinguishing G. messageri and the sympatric species referable to fallax, gouldingi and anterides are stable, therefore G. messageri is handled separately from the rest of the taxa.
In the recent revision of the Chinese members of the family (Páll-Gergely and Hunyadi 2013), Gudeodiscus phlyarius was reported from several localities in Guangxi. Plectopylis moellendorffi Gude, 1901 was synonymized with P. phlyarius. Gudeodiscus phlyarius werneri was described from two nearby localities near Duan city. All other Chinese G. phlyarius populations were assigned to the nominotypical subspecies. Gudeodiscus phlyarius phlyarius populations were listed in two separate groups based on their appearance, namely "phlyarius-like, mainly flat, small form" and "larger, stronglybuilt shell (transition to werneri)".
Here we include the following taxa as synonyms of Gudeodiscus phlyarius: anterides Gude, 1909, fallax Gude, 1909, fallax var. major Gude, 1909, gouldingi Gude, 1909, moellendorffi Gude, 1901, verecundus Gude, 1909, werneri Páll-Gergely, 2013. The last taxon was described on the basis of a keel with a light band around the umbilicus, the dissolved anterior lamella, the posteriorly elongated upper and lower ends of the posterior lamella and the parallel, horizontal palatal plicae. All other formerly recognized species (anterides, fallax, gouldingi, moellendorffi, verecundus) have two well-developed lamellae and oblique, usually depressed Z-shaped palatal plica, often with Y-like posterior ends. However, this study revealed that G. phlyarius is a widely distributed, very variable species and at this moment we see no good reason to maintain one of the morphologically distinct forms as a subspecies. Consequently, we synonymize G. phlyarius werneri with G. phlyarius.
According to the original description the anterior lamella of gouldingi is simple whereas that of anterides is "provided with buttresses". The upper parietal plica is in contact with the anterior lamella in gouldingi, but the lamella is shorter and free in anterides. Both the upper and lower plicae are shorter in anterides. The first palatal plica of anterides has a descending ridge; the same plica is straight in gouldingi. Additionally, the palatal plicae of anterides are not united by a vertical ridge and are more widely spaced than in gouldingi (the drawings in the original description show the reverse). All of the differences mentioned by Gude (1909) are unstable even within a single sample (assumed to be single population). For example, six shells were opened from a sample collected in Nat-Son (leg. Messager, MNHN-IM-2012-2153, containing 118 "gouldingi" shells). The length of the lower horizontal plica varies greatly, but extends beyond the anterior lamella in the anterior direction in every cases. One specimen had buttresses on the anterior lamella. Two specimens possessed an anterior lamella and the upper horizontal plica united, whereas in the case of four specimens this plica was free. Even among the few shells examined by Gude, he found that shells exhibited transitional character states between anterides and gouldingi. Therefore, these forms cannot be handled as separate species.
In the original description of Plectopylis fallax, Gude (1909) compared it only with P. moellendorffi. He did not compare P. fallax either with P. anterides, or with P. gouldingi. Based on the material housed in the NHM and the specimens mentioned in Gude's (1909) paper, Gude received very few shells from Messager. Examining the type specimens of the above-mentioned taxa revealed that besides the difference in size (typical fallax is larger than anterides and gouldingi), the only distinguishing feature is the simple and free palatal plicae in fallax and the bifurcated and usually connecting plicae of gouldingi (syn: anterides). The palatal plicae are very variable even within the same sample (see Figures 11) and certainly cannot be used to separate these taxa. Larger shells usually have separated palatal plicae and smaller shells tend to have joint palatal plicae. In addition, the characteristic "nautiliform" shape of typical fallax shells is also not a reliable distinguishing feature from Plectopylis gouldingi/anterides as this trait is also variable across gouldingi and fallax samples.
Based on shell size, most of Messager's samples in the MNHN can be assigned to three forms (approximately 11-13 mm: gouldingi, 14-16 mm: fallax, 19-21 mm: fallax var. major). However, the ranges of shell size overlaps within a few samples (see "mixed" samples under the material) and assigning some of these shells to one of the forms is impossible. The size range from typical gouldingi (11 mm) to fallax var. major (21 mm) shows a clinal variation without interruption (see Figure 16). On the other hand, we found one sample where the shells clearly differ from two separate forms, namely six typical "fallax var. major" (D: 18.9-20 mm) and gouldingi (D: 12.4-13.5) shells. Unfortunately, as in other samples, the collection locality is not exact enough to determine if these specimens were sympatric.
The apertural fold is always present on typical Gudeodiscus phlyarius shells, but can be rudimentary or missing in typical anterides/fallax/gouldingi shells. The edge of the periostracal folds has a pointed structure which seems to occur in a spiralling pattern on the shell of most Vietnamese phlyarius specimens, but these are always missing in fallax and gouldingi specimens (this trait is visible only in fresh shells) ( Figures  10C-F). Typical moellendorffi specimens (synonym of phlyarius) possess a somewhat elevated spire, whereas typical anterides/fallax/gouldingi shells are almost always entirely flat. The only shell character found to be stable within typical Vietnamese Plectopylis phlyarius shells and Plectopylis anterides/fallax/gouldingi shells, however, is the rounded aperture in the former and the elongated aperture in the latter (Figures 9C-D). Even this difference is found to be variable in Chinese populations. The populations listed as "transitions to werneri" in Páll-Gergely and Hunyadi (2013) have rather elongated aperture, similar to that of typical Vietnamese fallax shells, but have elevated spire and overall similar shell shape to typical Vietnamese phlyarius. Therefore, we refer to anterides, gouldingi and fallax as synonyms of G. phlyarius.
The genital structure of typical fallax and typical phlyarius differ considerably. Namely, the former lacks the penial caecum or has only a very small one, and has a reticulated inner surface of the penis, whereas the latter has a short penial caecum and its penis has parallel folds on the inner wall. The size of the penial caecum however, may not have a strong taxonomic value because it was found to vary largely within species (e.g. Gudeodiscus multispira, see Páll-Gergely and Asami 2014). The sculpture of the wall of the proximal portion of the penis may have a seasonal variability (see under G. villedaryi and in Discussion).
A sample (MNHN 2012-2177) labelled verecunda, which contained 9 shells from the type locality (Phony-Tho) supports the synonymy of the taxon in relation to gouldingi and fallax, and therefore to Gudeodiscus phlyarius. Seven of the shells were typical verecundus with an elevated spire, a strong apertural fold connected to the callus, and an anterior lamella fused to the lower plica; the plica does not extending beyond the lamella anteriorly (confirmed in 3 shells). The two other shells however, have somewhat lower spires, the apertural fold is not connected to the callus and the lower plica is free from the anterior lamella and extended beyond it anteriorly (one of the two shells was opened). These two shells can be interpreted as transitional forms between verecundus and fallax in terms of spire height, apertural fold and parietal plicae/lamellae morphology. Since transitional forms were found between typical verecunda and fallax shells, P. verecunda can be interpreted as a local form of fallax having elevated spire and fused anterior lamella and lower plica. Therefore, we synonymise Plectopylis verecunda with G. phlyarius.
There are two Vietnamese "forms" of Gudeodiscus phlyarius which differ from all other typical Vietnamese phlyarius shells. One of the morphologically distinct forms inhabits Ninh Bình Province, where we have knowledge of two populations (number 3 on Figure 43). These shells are smaller and comparatively flatter than the usual phlyarius, and have a characteristic "nautiliform" shape, wider umbilicus, with the last whorl leaving the larger part of the penultimate whorl visible. No differences in the lamellae were recognized. The other form is known from one locality in north-western Thái Nguyên Province (number 2 on Figure 43). This has an elevated spire and narrow umbilicus. Only three specimens are known, and two of them were opened. One of the opened specimens had three very weak parietal lamellae (possibly an abnormal character state, similar to that of the holotype of Plectopylis infralevis), and the second has the anterior lamella and the lower plica fused; the plica did not extends beyond the anterior lamella in the anterior direction.
Two Chinese populations (near Baxianyan, number 1 on Figure 43) have an oblique anterior lamella and an aperture more reflected downwards. Differential diagnosis. The shell shape of Gudeodiscus suprafilaris is similar to that of G. infralevis, but G. suprafilaris has more regular whorls, a more elevated spire and its sculpture changes suddenly from reticulated dorsally to smooth basally on the last whorl. The sudden change of the sculpture and the almost globular shell distinguishes the species from other species (G. eroessi, G. multispira, G. soosi, G. yunnanensis, G. cyrtochilus and G. fischeri). The Chinese G. eroessi hemisculptus Páll-Gergely & Hunyadi, 2013 and G. yanghaoi which have similar sculpture are larger, have a flatter shell and different lamellation.
Intraspecific diversity. The species is very variable in terms of spire height, the formation of parietal and palatal plicae and lamellae, and the extent of the sculptured portion on the dorsal side of the shell. The distinctive aperture shape, minute apertural fold and the unique sculpture render this species distinctive and easy to identify. See also Remarks and Table 10.
Remarks. The palatal and parietal plicae and lamellae exhibit extreme variability between populations. The holotype exhibits relatively long, horizontal palatal plicae connected with a ridge; the parietal side possesses a well-developed posterior lamella, upper and lower plica, and a reduced, short anterior lamella ( Figures 14S-T). The museum specimens we examined (probably from the same sample as the holotype) had similar palatal plicae and also a reduced anterior lamella. Two examples collected close to the type locality (2011/81, see Figures 14U-V and 2012/44) were examined. Shells belonging to both populations had identical palatal plicae to those of the holotype, but in contrast, had a much longer anterior lamella, free from the lower plica or almost united to it. Additionally, in the type series, the sculptured dorsal surface changes to a smooth surface at around the middle line of the body whorl. In contrast, in the two newly-collected samples the change between the two different sculptures occurs lower, closer to the umbilicus. In a shell from another population (Vn10-125, see Figures 14X-Y) the palatal plicae were greatly reduced in length so that when viewed through the semi-transparent shell, they appear as though only a single vertical plica was present. The parietal wall of the same shell was ornamented by a strong anterior lamella entirely fused with the lower plica; the posterior lamella was absent, its position was indicated only by a very slight elevation within the structure of the shell. Figures 8B-D  Diagnosis. Shell medium-sized to large, strongly-built, nearly smooth, with thick apertural lip and an oblique, strong apertural fold ( Figure 9J); umbilicus frequently keeled. The anterior parietal lamella is supported by an anteriorly elongated lower plica; an additional, long horizontal plica is present near the lower suture; middle palatal plicae oblique (Figures 13V-Y).
Differential diagnosis. See under Gudeodiscus dautzenbergi and Halongella schlumbergeri. Intraspecific diversity. The morphology of palatal and parietal plicae and lamellae do not show significant variation. Conversely, shell size, aperture shape, shape of the dorsal side of the shell, spire height and the presence or absence of the periumbilical keel show considerable variation across populations. See also Table 11.
Description of the genitalia. Three specimens were anatomically examined; they were collected at the same locality at different times of the year (20090520A: 20 May, two specimens; 2011/102: 12 November, one specimen). One of the specimens from the 20090520A sample had abnormally developed genitalia. Namely, the penis was "normally" connected to the genital opening, but the vagina was only attached to the atrium area with weak fibres. Nevertheless, the gametolytic sac was filled with fragments of a spermatophore which is an indication of successful mating. An epiphallus was absent and the vas deferens started from the base of the vagina. The other specimen from the 20090520A sample (collected in May) had 18 embryos developed in its uterus, and had no claws between the folds on the inner wall of the penis, whereas the one collected in November was not gravid, but had several claws within the folds inside the penis. The claws had a moderately long base inside the pockets, whereas their hook-like tip was hanging out of the pockets. The SEM images revealed that the base had a granulated surface, probably to provide a better attachment to wall of the pockets, whereas the tip was smooth. Additionally, the specimen from November had parallel, dense, wavy, horizontal folds on the inner wall of the proximal part of the penis, and longitudinal, parallel folds on the distal portion of the penis. The other specimen sampled in May had only a slightly waved proximal part of the longitudinal folds. Other parts of the genitalia did not differ between the two specimens.
The penis is short, pear-shaped internally with pockets standing in a straight row at the distal part of the penis; the epiphallus is much more slender, and is somewhat shorter than the penis; there is no penial caecum, the retractor muscle attaches on the apical part of the penis (at the penis-epiphallus transition); epiphallus approximately as long as the penis, it transforms to vas deferens without obvious boundary; epiphallus internally with parallel folds; vagina long with a well-developed vaginal bulb, it is attached to the body wall with several ligaments; vaginal bulb with thickened wall, internally almost smooth, only with hardly visible longitudinal folds; inner wall of the distal part of the vaginal with low, parallel or converging, serrulate folds ( Figure 32D); there is a shorter, thicker gametolytic sac and a longer, more slender diverticulum. Table 6 and Figures 35M-O. Distribution (see Figure 40). The species is known from Thái Nguyên and Lạng Sơn provinces.

Radula. See
Remarks. Gudeodiscus villedaryi is a very variable species in terms of shell characters. The species is recognised on the basis of the presence of an additional lower plica, which is absent in G. dautzenbergi. The latter species might be only a variety of G. villedaryi which has lost the lower plica. More information is needed to determine whether the populations assigned to G. villedaryi and G. dautzenbergi form monophyletic groups. See also under G. dautzenbergi.
Subgenus Veludiscus Páll-Gergely, subgen. n. Type species. Gudeodiscus eroessi Páll-Gergely & Hunyadi, 2013. Diagnosis. Shell indistinguishable from those of the subgenus Gudeodiscus (Gudeodiscus) and the genus Halongella gen. n. Anatomy: Epiphallus is slender, cylindrical; retractor muscle inserts on the distal end of the penial caecum, but the whole caecum is covered by additional, fine muscle fibres which insert on the distal end of the penis. Radula: central tooth smaller than the ectocone of the first lateral; mesocone of the first lateral is usually wide, rhomboid. Marginals bi-or tricuspid, with blunt inner cusp and shallow incision between the inner two cusps. See drawings and descriptions of the genital anatomy in Páll-Gergely and Hunyadi (2013) and Páll-Gergely and Asami (2014).
Etymology. The name Veludiscus is composed of two Latin words. Velum (=curtain, sail, covering) refers to the characteristic feature of the genitalia, namely the additional curtain-like muscle covering the penial caecum and the retractor muscle, and discus (=disc) refers to the shape of the shell. The genus is gender masculine.
Remarks. Some conchologically similar species may belong to this subgenus, especially those which inhabit similar geographic regions. Future investigations on the anatomy and radula morphology of Gudeodiscus species should clarify the subgeneric status of the taxa with unknown anatomy. (Möllendorff, 1901) Diagnosis. A medium-sized to large species with dense, fine riblets; shell flat, callus always, apertural fold usually present. Parietal wall with C-shaped posterior lamella; anterior lamella (if present) slightly S-shaped; if anterior lamella is missing; one lower plica or four parallel plicae are visible in front of the lamella; palatal wall with almost straight, slightly oblique, depressed Z-shaped or Y-shaped plicae ( Figures 13A-D).

Gudeodiscus (Veludiscus) emigrans
Differential diagnosis. Gudeodiscus phlyarius has stronger apertural fold, a straight anterior parietal lamella (in the Chinese populations assigned to G. phlyarius werneri Páll-Gergely, 2013 = synonym of phlyarius, sometimes dissolved into small denticles) and usually a somewhat elevated spire. Gudeodiscus messageri, G. hemmeni sp. n. and G. anceyi have two parietal lamellae or several small denticles standing in a line at the position of the first lamella.
General distribution. The three subspecies of G. emigrans are known from northern Vietnam and northern Guangxi. (Möllendorff, 1901) (2013). Diagnosis. Spiral sculpture missing or not conspicuous, parietal wall with one lamella and a short lower parietal plica anterior to the lamella.
Intrasubspecific diversity. Very few shells are known from museum collections. The subspecies is easily recognisable, but more material is needed to understand the intrasubspecific diversity.
Distribution. Plectopylis (Sinicola) emigrans was described from the "Manson-Gebirge" = "Mau Son Mts, about 30 km E of Lang Son" (Schileyko 2011) (see Figure 39). Diagnosis. Shells do not differ from those of Gudeodiscus; small to very large, body whorl rounded, callus and apertural fold; Parietal wall with two lamellae or the ante- rior one is reduced or absent; parietal side with straight, slightly curved, or depressed Z-shaped plicae.

Gudeodiscus (Veludiscus) emigrans quadrilamellatus
Penial caecum absent. Penis internally with longitudinal, parallel folds, with tiny, flat, T-shaped calcareous granules between the folds, all along the penis; there are no determined "pockets" for the granules at the apical part of the penis. Epiphallus internally with longitudinal folds having several perpendicular projections which overlap with those of the neighbouring fold. Radula similar to Gudeodiscus (Veludiscus) subgen. n. by the smaller central tooth than the ectocone of the first laterals and the marginals which are bicuspid or tricuspid with blunt innermost cups and shallow incision between the two inner cusps.
Differential diagnosis. Sinicola species have a keeled body whorl, whereas it is rounded in Halongella gen. n. Moreover, all Sinicola species have a penial caecum, a central tooth which is as large as or larger than the ectocone of the first laterals and clearly tricuspid marginals with deep incision between the innermost two, sharp cusps. The same radular morphology has been observed in Sicradiscus species. Additionally, "eastern" Sicradiscus species possess keeled shells, whereas the rounded shelled "western" species of the genus have determined pockets on the inner penial wall, similar to that of Gudeodiscus. For comparison with Gudeodiscus, see there.
Included taxa. fruhstorferi Möllendorff, 1901 andschlumbergeri Morlet, 1886. Etymology. This generic name derives from the name of the Halong Bay, where both species occur. The genus is gender feminine.
Remarks. Calcareous granules of complicated shape have been found in the vagina of Halongella schlumbergeri, and some granules not having characteristic shapes have been found in the vaginal lumen of H. fruhstorferi. The taxonomic value of these granules are unknown. No granules of characteristic shape have been found in the vaginas of Gudeodiscus species, therefore this can be a synapomorphy of Halongella gen. n. (Möllendorff, 1901)   Diagnosis. Shell small, solid, thin-walled, almost flat and smooth, with weak apertural lip and sometimes a small apertural denticle ( Figure 9O). Parietal wall with one parietal lamella with two short horizontal plicae anteriorly, one above and one below; palatal plicae short, oblique, depressed Z-shaped ( Figures 14O-R).
Intraspecific diversity. The species is known from a very small area, and only few specimens are known. The intraspecific diversity is low.

Radula. See
Diagnosis. Shell medium-sized to very large, thick shelled, almost smooth or with very fine periostracal ribs; apertural lip well-developed; apertural fold long, more or less equally long in its total length, connected to the callus. Parietal wall with missing or short anterior lamella (always distant from the upper plica) and well-developed posterior lamella; palatal plicae depressed Z-shaped.
Differential diagnosis. Gudeodiscus dautzenbergi and some populations of G. villedaryi resemble Halongella schlumbergeri in terms of general, but the inner lamellae are entirely different, namely, G. dautzenbergi and G. villedaryi have strong, well-developed anterior lamella with an anteriorly elongated lower "leg", whereas most H. schlumbergeri Figure 24. Reproductive anatomy of Gudeodiscus (Gudeodiscus) villedaryi (Ancey, 1888), abnormal specimen. Locality information: Thái Nguyên Province, Võ Nhai District, Phú Thượng Commune, Phượng Hoàng Cave, Mỏ Gà Vill., ca 150 m, 21°46.836'N, 106°07.107'E, leg. Ohara, K., 20.05.2009. Scale represents 2 mm. shells lack the anterior lamella. It is possible to distinguish H. schlumbergeri from the other two species without breaking the shell, on the basis of the long apertural fold reaching the callus, which is short in G. dautzenbergi and G. villedaryi, and has an elevated "knob" part in some distance from the callus. See also under H. fruhstorferi.
Intraspecific diversity. The species is very variable in terms of shell size and the formation of plicae and lamellae on the parietal wall.
Radula. See Table 6 and Figures 36D-F. Distribution (see Figures 40 and 44). The species has only been recorded in the Hạ Long Bay area (Hải Phòng and Quảng Ninh provinces).
Remarks.  figured specimens of all three "species": schlumbergeri, jovia and villedaryi (later re-named pilsbryana). His observations were based on one specimen from each "species". He wrote the following: "A comparison of these three species has shown that that they are very closely allied, and that there is no difference of diagnostic value between the armature. They differ, however, in external aspect sufficiently to rank as separate forms. P. jovia is the largest of the three, while P. villedaryi is the smallest, P. schlumbergeri being intermediate in size." The additional differences mentioned by Gude, namely the strength of the callus, direction and small differences in the shape of the palatal and parietal lamellae and plicae are not sufficient to separate  (Gude, 1908). Locality information: Cao Bằng Province, southern edge of Pác Rải, Trùng Khánh 3 km towards Quảng Uyên, left side of the road, 570 m, 22°48.961'N, 106°30.533'E, leg. Hunyadi, A., 28.05.2012. Scale represents 2 mm. species. We had the possibility to observe and measure a number of shells collected in the Hạ Long Bay Area and provided with exact GPS data. The outer shell characters exhibit little variation other than in size. Therefore, we suggest syno nymising the three species under one name.
The shell differences between Plectopylis schlumbergeri (and its synonyms) and Plectopylis hirsuta, namely the short or missing anterior lamella in schlumbergeri and the relatively "normal" anterior lamella of hirsuta are considered to be very minor. This trait shows clinal variation across shells assigned to hirsuta and schlumbergeri (and its synonyms). We therefore synonymize Plectopylis hirsuta with Halongella schlumbergeri.

Diagnosis.
A very small species with reticulated dorsal and glossy ventral surface, elevated spire, elevated, sharp callus and well-developed apertural fold connected to the callus ( Figure 9H). Parietal wall with two lamellae, the anterior one separated from both the lower and upper plicae; middle palatal plicae short, connected with a ridge and sometimes ornamented with small denticles posteriorly ( Figures 11A-B).
Measurements (in mm). D = 6.7-7, H = 3.4-3.9 (n=4, 20081116C). Differential diagnosis. All other similar congeners inhabit China. Sicradiscus feheri Páll-Gergely & Hunyadi, 2013 is larger, flatter with a wider umbilicus and a shinier dorsal surface, has a longer, horizontal palatal plicae without additional posterior denticles, and has a more elevated and longer apertural fold. Sicradiscus transitus Páll-Gergely & Hunyadi, 2013 has a lower spire and a wider umbilicus with slightly shouldered whorls, sometimes strong radial lines on the ventral surface, and a more elevated callus. Moreover, the anterior lamella of S. transitus is in contact with both the upper and the lower plicae, which are free from the lamella in S. mansuyi. Sicradiscus invius is flatter (has shallower umbilicus) with only the protoconch  (Möllendorff, 1901), empty triangle: Gudeodiscus (Gudeodiscus) dautzenbergi (Gude, 1901), filled triangle G. (G.) villedaryi (Ancey, 1888), empty square: G. (G.) anceyi (Gude, 1901), filled circle: Sicradiscus mansuyi (Gude, 1908). elevated from the dorsal surface; it has weaker dorsal sculpture resulting in a glossy surface (mansuyi is densely reticulated), and lacks the additional small denticles posterior to the palatal plicae, which are usually present in S. mansuyi. Gudeodiscus anceyi is larger and has a ribbed shell with spiral lines on the whole shell. Species possessing a glossy ventral surface (G. cyrtochilus, G. fischeri) are also larger and have weaker or no apertural fold. Intraspecific diversity. Low; shell characters stable. The species is easily recognisable and can be separated from other plectopylid species without difficulty.
Penis with a shorter, slimmer proximal section and a thinner, somewhat longer distal portion; internally with parallel folds which are more elevated in the thinner distal portion, forming pocket-like structures (similar to that of S. transitus, see Páll-Gergely and Asami 2014); these "pockets" did not contain granules; epiphallus approximately as long as the penis but much slimmer; internally penis and epiphallus wall with longi-tudinal, parallel folds; retractor muscle short, inserts on the penis-epiphallus transition; penial caecum absent. Vagina long, with distal vaginal bulb; vaginal bulb and other parts of the vagina with approximately 8, more or less parallel, serrulate folds ( Figure  31A); vas deferens long, thicker distally and more slender proximally; gametolytic sac and diverticulum are of equal length, in parallel.
Radula. See Table 6 and Figures 36J-L. Distribution. This species was described from Hạ Lang (eastern part of Cao Bằng Province, see Figure 39). We have seen newly collected material from northern Hà Giang and Cao Bằng provinces. The first occurrence of the species from China is reported. This locality is situated very close to the Vietnamese border.

Identification and species recognition
For this revision of the Vietnamese Plectopylidae, we examined the type specimens of all known taxa, 197 newly collected specimens with detailed locality data and 631 historical lots deposited in a variety of public collections. Altogether we examined  (Gude, 1908), triangle: G. (G.) fischeri (Gude, 1901), empty square: G. (G.?) cyrtochilus (Gude, 1909). more than 7000 shells (see Table 12). We found specimens of most species in European museum collections, probably because of intensive shell exchanges at the beginning of the 20 th Century. The present scale of specimen examination allowed us to understand species boundaries in the Vietnamese Plectopylidae better than the preceding studies.
Although the plicae and lamellae (especially on the parietal wall) are common characteristics of the family and useful for identification of some species, their value in species recognition has been somewhat overestimated. This appears to have led to descriptions of several species that differ only slightly in palatal and parietal plication. Our recognition of distinct species is primarily based on general shell and aperture shape, and secondarily on the morphology of plicae and lamellae.

Key characters for identification (see also identification key)
As a summary, below we present the most important shell characters for identification of each species from others within the Vietnamese Plectopylidae. In the case of Gudeodiscus emigrans emigrans and G. infralevis, however, available shell specimens were insufficient to provide help for "routine" identification.  (Gude, 1909) and "anterides", "fallax" and "gouldingi"like populations of G. (G.) phlyarius. Filled square indicates the position of Phong-Tho, the type locality of Plectopylis verecunda Gude, 1909 (synonym of G. phlyarius  Identification key to Vietnamese and Chinese plectopylid genera The "Eastern Plectopylidae" (see Páll-Gergely and Hunyadi 2013), namely taxa inhabiting China, Vietnam, Taiwan and Okinawa (Japan) are conchologically relatively diverse. Their common features are the ribbed protoconch and the absence of the long parietal horizontal plica. The genus Endothyrella, which mainly inhabits north-eastern India also shares these features with the genera of "Eastern Plectopylidae". Therefore Endothyrella is possibly a close relative to the genera Gudeodiscus, Halongella gen. n., Sicradiscus and Sinicola. The genera of "Western Plectopylidae" (Endoplon, Chersaecia and Plectopylis) have smooth but matt or "tuberculated" embryonic whorls and usually long horizontal parietal plicae (a main plica and a lower plica) which run to the peristome. Some Endothyrella species have long lower and main plicae, but these may not be homologous with those in the Chersaecia and Plectopylis. Some species which have been assigned to the genus Chersaecia (andersoni W. Blanford, 1869, laomontana L. Pfeiffer, 1863 also possess ribbed protoconchs. These probably do not belong to any of the genera mentioned herein, and their taxonomic status require revision. In the revision of the Chinese Plectopylidae (Páll-Gergely and Hunyadi 2013), three genera were recognized, namely Gudeodiscus, Sicradiscus and Sinicola. The most important shell characters for recognition of Sinicola are the following: body whorl keeled; periostracal folds usually present on the keel; apertural fold almost always absent; the anterior parietal lamella is absent or present only in some small, separate denticles. Gudeodiscus exhibits the following characters: body whorl rounded; periostracal folds absent; apertural fold often present; anterior parietal lamella often present. Both genera inhabit restricted geographical areas with minor overlaps; Sinicola ranges from Middle Sichuan to northern Guangxi, Guangdong and eastern Hunan, whereas Gudeodiscus ranges from northern Vietnam to southern Hunan and southern Guangdong. Reproductive anatomical investigations (Páll-Gergely and Hunyadi 2013, Páll-Gergely and Asami 2014) found that Sinicola species exhibit a ribbed inner penial wall with a few tiny calcareous granules. The ribs are more prominent in the distal part of the penis or continuous until the atrium but this varies between individuals. Examples of Gudeodiscus usually also have parallel folds, but they have characteristic small pockets arranged in one or two more or less straight transverse lines in the distal penis. These pockets contain calcareous granules, probably only during the mating period (see discussion on anatomy and biology). The genus Gudeodiscus is divided into two groups based on the morphology of the distal penis-penial caecum-retractor muscle complex. In one type, the epiphallus is slender, cylindrical, and addition to the retractor muscle, which attaches on the penile caecum, several muscle fibres attach to the penis itself. In the other type the epiphallus has a somewhat thickened proximal part, and has no additional muscle fibres attached to the penis (Páll-Gergely and Hunyadi 2013, Páll-Gergely and Asami 2014). It may not be legitimate to subdivide the genus on the basis of this anatomical difference, when the shell characters do not show clear distinction (Páll-Gergely and Asami 2014). For example, G. eroessi (first type) and G. multispira (second type) are conchologically very similar. However, we found that radula traits distinguish between them as well as the genital anatomy does. Therefore we find it well supported to separate these two groups into different subgenera (Gudeodiscus and Veludiscus subgen. n.).
The taxonomic position of the species classified within Sicradiscus is problematic. Sicradiscus was erected for several, small bodied species which inhabit a large area ranging from Sichuan to Okinawa, Japan. There is continuous variation across the genus Sicradiscus in terms of shell characters. Sicradiscus invius, S. securus, S. mansuyi and S. feheri have a rounded body whorl and possess a strong apertural fold. In contrast, Sicradiscus schistoptychia, S. diptychia, S. cutisculptus, S. ishizakii and S. hirasei have a shouldered body whorl and lack the apertural fold. The two groups are within the same genus because Sicradiscus transitus is similar to S. schistoptychia in possessing divided palatal plicae and a keeled body whorl, at the same time having a strong apertural fold similar to that of S. feheri. Moreover, S. transitus ranges between S. feheri and S. schistoptychia geographically. The present and a previous study (Páll-Gergely and Asami 2014) revealed that the inner morphology of the penis in S. schistoptychia is similar to that of Sinicola, whereas S. invius, S. mansuyi and S. transitus are similar to Gudeodiscus in that trait. Separating some Sicradiscus species into Gudeodiscus and others in Sinicola based on the penial morphology does not resolve their taxonomy because of the large conchological similarity among Sicradiscus species. An alternative classification might be to place all Gudeodiscus, Sicradiscus and Sinicola species into one genus because of the transitional features of Sicradiscus between Sinicola and Gudeodiscus. However, our study does not support this because both Sinicola and Gudeodiscus show clear synapomorphic characters and signs of their separate major radiations in different geographic areas. The most possible explanation is that Sicradiscus species represent basal lineages within the Gudeodiscus-Sicradiscus-Sinicola complex, in which others diverged into the two lineages, one with the keeled body whorl and folded penial wall and the other with the rounded body whorl and pocketed penial wall. Sicradiscus species may probably have undergone only slight conchological changes. This hypothesis is supported by the geographic distribution of most Sicradiscus species, roughly between the areas of Gudeodiscus and Sinicola.
Plectopylis schlumbergeri and P. fruhstorferi had parallel folds on the inner penial wall and calcareous granules were found between the parallel folds all along the penis. In both subgenera of the genus Gudeodiscus however, the pockets for calcareous granules are arranged in one or two rows, and they are absent elsewhere. Based on this morphological character, they are moved to a new genus, Halongella. Additionally, Halongella gen. n. species lack a penial caecum, which was found in the majority of Gudeodiscus species.

Anatomy and biology
Stoliczka (1871) described the organ proximal to the gametolytic sac of Plectopylis as "a shorter, more muscular gland which appears to represent the arrow or amatorial gland". Pilsbry (1894) noted this as "an organ of unknown homology, either a dart sack, a diverticulum of the spermatheca or an appendicula". A spermatophore was found inside this organ of Gudeodiscus fischeri. This suggests that the organ is a diverticulum, starting from the wall of the distal end of the vagina/beginning of pedunculus. In most stylommatophoran land snails the diverticulum derives from the stalk of the gametolytic sac. The only exception known before this study was the subfamily Garniierinae (family Clausiliidae), in which the diverticulum derives from the pedunculus (Szekeres 1998).
The inner walls of the male genital organs, especially the penis, show a large diversity across the genera Gudeodiscus, Halongella gen. n., Sicradiscus and Sinicola. Sinicola and Halongella gen. n. have parallel folds on the inner penial wall, occasionally with tiny, usually flat calcareous granules, often without characteristic shapes. The penial wall of Gudeodiscus species is usually also characterized by folds, but also pockets arranged in one or two rows in the distal part of the penis. The rows can be straight (e.g. G. giardi and G. villedaryi), can follow a bell-shaped line (G. fischeri), or waves (G. messageri raheemi ssp. n.) on the opened penal wall. Sicradiscus species have both types of penial sculpture (with and without pockets) (Páll-Gergely and Asami 2014, and this study). In most Gudeodiscus specimens the granules are hook or claw-like, and each of them is placed within a pocket on the wall of the head of the penis. Two dissected specimens of Gudeodiscus phlyarius (typical fallax specimens), however, had flat, oval granules within the penial pockets. It is not clear whether this shape of granules is stable throughout the life span or dependent on season or age. In the revision of the Chinese species (Páll-Gergely and Hunyadi 2013) we described that calcareous hooks are easily removable from the folds in the penial internal wall. In the case of Vietnamese specimens (G. giardi, G. fischeri and G. villedaryi), however, the claws were attached into the wall inside the pocket and were difficult to remove. The SEM images of removed claws revealed that the base of each claw, which was buried into the pocket wall, is granulated in the surface, whereas the exposed tip of each claw was smooth. The hooks from the penis lumen of Chinese Gudeodiscus phlyarius (figured specimen in Páll-Gergely and Hunyadi 2013) dissolved with no remains in 90% lactic acid. Thus, these granules may consist of calcium carbonate.
The penial claws or hooks known in other stylommatophoran families (e.g. Zonitidae s.l., Streptaxidae, Cryptazeca) do not seasonally disappear and are fixed to the internal wall, because to our knowledge, hook-less specimens have not been reported in contrast to those in Plectopylidae (see also Páll-Gergely and Hunyadi 2013). Those of Cryptazeca and Streptaxidae are not calcareous (Visser 1973, Verdcourt 1979, 1985, Gómez 1991, whereas Zonitidae have calcareous claws (Schileyko 2003). The hooklike granules of Gudeodiscus and the minute, flat, or sometimes elongated or globular granules of other plectopylid genera may have similar roles but a different origin from the fixed claws of other Stylommatophora.
In some Gudeodiscus specimens the proximal (lower) part of the penial wall is ornamented with longitudinal folds only, but in others it has transverse and dense wrinkles (e.g. in G. giardi giardi and in one specimen of G. villedaryi). The transverse and longitudinal arrangement may result in a reticulated surface of the inner penial wall, such as those in G. phlyarius (fallax-like specimens). These traits need to be used for taxonomy with careful attention to collection dates and instead may provide opportunities for studies of functional roles for reproductive success for the following reason: two specimens of G. villedaryi collected in different periods of the year (20 May and 12 November) from the same locality greatly differed in these traits. The one collected in May was gravid, and its penis had only longitudinal folds on its inner wall, with slightly waved proximal portions of the folds. In contrast, a specimen collected in November was not gravid and had conspicuous, dense and transversal folds on the proximal portion of the inner wall of the penis. This transversal folded structure turned suddenly to a longitudinal folded area with calcareous claws between the pockets. This result suggests that the morphology of fine sculpture of the inner penial wall (at least inside the proximal half of the penis) may be seasonally variable. The gravid individual may have lost hooks in a mating period before collected in May. The latter individual with no embryo may have been in a period for copulation. Our observation suggests that the penial internal wall may be restructured to regenerate the hook-like calcareous claws for copulation. Further studies are necessary to test this hypothesis.
The other organs of male genitalia, penial caecum and epiphallus have generally a simpler inner surface, usually with parallel and longitudinal folds, than the penis. In smaller species it is difficult to open these very slim organs, especially the epiphallus. The longitudinal folds on the inner wall of the epiphallus of Halongella gen. n. species have perpendicular projections which overlap with those of the neighbouring fold. Besides this, all other species have an epiphallus with simple internal longitudinal folds. The inner wall of the penial caecum is also ornamented by longitudinal folds, which are sometimes wavy, and form hollows with the neighbouring fold. This structure is similar to the penial sculpture of Sinicola species. A function of these hollows would probably be to hold the small calcareous granules. In some species the sculpture of the penial caecum is more complex; Gudeodiscus messageri raheemi has deep sinuses with the calcareous granules. Gudeodiscus giardi giardi has pockets formed by two neighbouring papillae (Páll-Gergely and Asami 2014). The calcareous granule within the caecum can be elongated or globular without any characteristic shape, such as in one of the dissected G. messageri raheemi specimens, or the granules can be hook-like, similar to, but smaller than those found in the penial lumen, such as in a specimen of G. pulvinaris pulvinaris (see Páll-Gergely and Asami 2014).
Specimens that were fixed in 70% ethanol were used for this investigation. Thus, at this stage of study, we are not able to rule out a possibility that some of the granules appeared as observed because of the process of preservation. However, hook structure corresponds to pocket structure in the penial internal surface. Each hook is regularly located in a pocket in a determined orientation. Further, they exhibit a taxonomically characteristic and sophisticated shape. For these reasons, the presence of hooks and granules in the present family cannot be ascribed to an artefact during preservation.
The absence of embryos in the uterus was statistically significantly associated with the presence of calcareous granules inside the penis, within Gudeodiscus (p = 0.0001) and also across all the four genera (p = 0.0006) (Tables 3-5). This strongly suggests that these granules may function as a disposable male mating apparatus. These granules disappear perhaps through repeated copulation in a mating season. It could require some time to gain the granules again if they lose granules and bear offspring. Thus, for some time during the mating season, they might remain with no granules before embryos develop. If so, these would exhibit no granules or embryos. However, this was the case only in three of 34 specimens examined in this study. Our results illuminate the importance of further studies on their reproductive life history and the ecological function of these granules.
The function of the calcareous hooks and granules inside the penis are unknown, although they probably play some role as a mating apparatus as well as the non-calcareous hooks in other groups. It has been classically postulated that these may function for mechanical stimulation for mating success like other penial structures or darts (Tompa 1984;Atkinson and Atkinson 1987). However, later studies have shown that love darts are not for physical stimulation but to inject mucus which includes a substance that increases paternity by inducing reconfiguration of partner's organs for spermatophore digestion (Koene and Chase 1998;Chase and Blanchard 2006;Kimura et al. 2014). Separately, De Winter et al. (1999) proposed that the spines on the penial wall play a role in the process of spermatophore formation in the streptaxid genus Sinistrexcisa. This is probably not the case in Plectopylidae, because they have the structurally distinguishable epiphallus. Their spermatophores are formed in this organ instead of the penis, and thus the structure of parallel inner folds in the epiphallus matches the morphology of spermatophore. Tompa (1984) also suggested that the penial hooks may function as mechanical holdfasts during mating. The present study provides a systematic ground for further studies on the evolution of mating apparatus inside the penis.
The function of the characteristic vaginal granules in one of the Halongella schlumbergeri specimens are also unknown. To our knowledge, no disposable granules have been reported in land snails which are attached to the vagina wall. The presence of vaginal granules in a non-gravid specimen and the presence of "vaginal sand" in a gravid specimen indicate that these granules are present only seasonally, probably related to the mating period. The characteristic shape of the granules, namely the flat base portion and the needle-bearing apical part does not support the hypothesis that they are artefacts formed during preservation.
To our knowledge, information on plectopylid radulae was published by Stoliczka (1871; Plectopylis achatina, P. cyclaspis and Endothyrella pinacis), Solem (1966;Chersaecia simplex) and Chang and Ookubo (1999;Sicradiscus ishizakii). Here we publish the radula morphology of 23 Chinese and Vietnamese species. Our limited information suggests that the relative size of the central tooth and the shape of the marginal teeth may be used in the systematics of the family. The genera Sicradiscus, Sinicola and the subgenus Gudeodiscus have relatively large central tooth (as large as or larger than the ectocone of the first laterals), and their marginal teeth are tricuspid with pointed cusps and deep incision between the cusps. In contrast, Plectopylis, Halongella gen. n., and Gudeodiscus (Veludiscus) subgen. n. possess smaller central tooth than the ectocone of the first lateral, and their marginals are bicuspid, or even if they are tricuspid, the innermost cusp is blunt and small, and there is a shallow incision between the inner two cusps. Stoliczka (1871) mentioned that Endothyrella pinacis (that time Plectopylis pinacis) has a larger central tooth than the two Plectopylis species, but did not provide a description or drawing of the marginal teeth. The description of the radula of Chersaecia simplex by Solem (1966) is accurate but he did not publish drawings. In that species, the central tooth is "tiny", supposedly smaller than the ectocones of the first laterals. The ectocones of the outer marginals are "reduced and split" (= marginals are tricuspid). This information on the marginals, however, is insufficient to allow comparison with our data.

Habitat
Plectopylid species seem to be associated with calcareous areas. Living specimens occur at the base of large limestone rocks surrounded by leaf litter and humus. Thus, they are not rock-dwelling but ground-dwelling. Most living species have reticulated sculpture on the dorsal shell side, which is often covered with soil and this may be of value in providing camouflage.

Geographical coverage of the Vietnamese plectopylid fauna
At the beginning of the 20 th Century all the available information on the distribution and taxonomy of Plectopylidae came with specimens from northern and eastern part of northern Vietnam (Tonkin) (Figure 39). We were able to examine only a few newly collected northern Tonkinese samples. Therefore, our knowledge on those species reported from the northern border region of Vietnam is mainly based on museum specimens. On the other hand, we examined several newly collected samples from the eastern part of northern Vietnam (Tonkin). Almost all of these specimens were identified to hitherto known species. Most of these species were found in several localities. Thus, this study covered the taxonomic diversity of plectopylids in the eastern Tonkinese area relatively well. Plectopylid specimens from western Tonkin have been examined for the first time. This resulted in the present description of a new species and a new subspecies.
Little information on plectopylid diversity has been obtained in the lowlands of the Red River, although these areas may not provide suitable habitats for land snails that prefer limestone outcrops or mountainous areas. Molluscan fauna in the border region of Sơn La and Yên Bái Provinces (Phan Xi Păng= "Farsipan" Mountain and its vicinity) is nearly unknown, maybe due to their high abundance in the limestone-free bedrock. Humid mountain forests there, however, may provide suitable habitats for plectopylids.
The southernmost Vietnamese county where plectopylids have been recorded is Nghệ An. The southern part of Vietnam may have been less intensively studied than the northern area (Tonkin). Accordingly the southernmost distribution of the family remains undetermined.