New insights from museum specimens: a case of Viviparidae (Caenogastropoda: Mollusca) in Iwakawa’s collection preserved in the National Museum of Nature and Science, Tokyo

Abstract Background In this study, we clarify the classification of museum specimens of the family Viviparidae, which is composed of six species/subspecies in Japan, including three endangered species. We examined Viviparus sclateri specimens from the Tomotaro Iwakawa collection (1855-1933) in the National Museum of Nature and Science, Tokyo. The Iwakawa's collection was catalogued in 1919 and Viviparus sclateri, labelled with Naga-tanishi, the current Japanese name for H. longispira, which was, at the time, equivalent to Viviparus sclateri, was listed in this catalogue. The catalogue noted localities of Viviparus sclateri (Naga-tanishi) from outside Lake Biwa, including occurrences in Lake Kasumigaura and Lake Suwa. However, Heterogen longispira (Naga-tanishi) is currently considered to be endemic to Lake Biwa drainage. The actual status of Viviparus sclateri in Iwakawa (1919) has not been clarified until now. New information Our examination revealed that Viviparus sclateri from Iwakawa’s catalogue included H. japonica, H. longispira and Sinotaia quadrata histrica, based on current taxonomy. Specimens assigned to H. longispira occurred only in Lake Biwa drainage. Heterogen japonica was confirmed to be present in all lots and some H. japonica from Lake Suwa had a distinctive morphology. Sinotaia quadrata histrica was only confirmed to occur in Lake Suwa. Furthermore, some specimens from southern Lake Biwa and the Seta River had intermediate characteristics between H. japonica and H. longispira and their populations are currently almost extinct.


Introduction
Museum specimens provide valuable insights into the evolutionary and ecological history of living and extinct populations, their taxonomy and knowledge for conservation biology (Suarez and Tsutsui 2004). For example, morphological and molecular data on historical specimens assist in the clarification of the classification and various other biological issues (Wandeler et al. 2007). In this study, we focused on specimens of endangered freshwater molluscs in a collection of Tomotaro Iwakawa from the National Museum of Nature and Science, Tokyo. Tomotaro Iwakawa (1855Iwakawa ( -1933 was an early contributor to zoology in Japan and he researched freshwater molluscs in Japan with particular attention (Taki 1933). Furthermore, he established one of the earliest systematic malacological lists in Japan (Taki 1933) and specimens of the list were deposited at the National Museum.
Viviparidae Gray, 1847, belonging to Caenogastropoda, is a family of freshwater gastropods with 125-150 valid described species globally and have a wide geographic range in Asia (Hirano et al. 2019b, Van Bocxlaer and Strong 2019, Stelbrink et al. 2020. Historically, Japanese viviparid gastropods have been classified as four species, based primarily on conchological features (Habe 1990, Masuda andUchiyama 2004): Cipangopaludina chinensis laeta (Martens, 1860), Cipangopaludina japonica (Martens, 1861), Heterogen longispira (Smith, 1886) and Sinotaia quadrata histrica (Gould, 1859). Recently, some attributions were modified by Hirano et al. (2019b), namely, Cipangopaludina japonica was modified to Heterogen japonica, based on molecular phylogeny. Furthermore, molecular analyses have led to the recognition of two subspecies of C. chinensis in Japan: C. c. laeta and C. c. chinensis and the existence of an undescribed species of Heterogen sp was also suggested (Hirano et al. 2019a, Hirano et al. 2019b). Thus, based on all the above-mentioned work on Japanese Viviparidae, we consider six viviparid species/subspecies to exist in Japan. Of these six species/ subspecies, H. longispira is listed on the IUCN Red List as EN (Köhler and Rintelen 2011). Furthermore, C. chinensis is listed on the IUCN Red List as LC, although the subspecies is not designated (Köhler et al. 2012). These two species are also listed on the Red List issued by the Government of Japan as NT and VU, respectively (Ministry of the Environment, Government of Japan 2019). Furthermore, although H. japonica is invasive on some Continents (Van Bocxlaer and Strong 2016), it is also listed on the Japanese Red List as NT (Ministry of the Environment, Government of Japan 2019).
The highly-endangered H. longispira (Köhler and Rintelen 2011, Nakai 2016, Ministry of the Environment, Government of Japan 2019) is endemic to Lake Biwa drainage (Hirano et al. 2015, Hirano et al. 2019b) which consists of Lake Biwa, a river flowing from Lake Biwa, Lake Yogo (artificially drained from Lake Biwa) and Lake Biwa Canal (Fig. 1). Besides, Lake Biwa has an ancient origin with a remarkable biodiversity (Horie 1971, Yokoyama 1984, Kawanabe 1996, Rossiter 2000, Okuda et al. 2013. Prior to its description, H. longispira was included under Paludina ingallsiana Lea, 1856 (Kobelt 1879, Iwakawa 1895, Iwakawa 1897a, Iwakawa 1897b. Then, Pilsbry (1902) indicated that P. ingallsiana does not occur in Japan and H. longispira was included in Vivipara sclateri Frauenfeld, 1865. After this, Vivipara sclateri (or Viviparus sclateri) was adopted as the species name for specimens from Lake Biwa (Hirase 1909, Hirase 1910, Kobelt 1909, Lake Biwa fisheries experimental station 1915, Annandale 1916, Kawamura 1918, Iwakawa 1919. In addition, the name Vivipara sclateri was used for viviparid specimens from Japanese localities other than Lake Biwa (Kobelt 1909, Annandale 1916, Iwakawa 1919. Furthermore, Annandale (1921) pointed out the difference between Vivipara sclateri by Frauenfeld and the distinctive viviparid gastropods from Lake Biwa (= H. longispira), based on comparison with the type illustration of Vivipara sclateri and he described specimens from Lake Biwa as Heterogen turris Annandale, 1921. However, Smith (1886 had already described the endemic species in Lake Biwa as Viviparus longispira Smith, 1886, although this paper was not referred to by other malacologists at that time. Finally, Kuroda (1929) reclassified Viviparus longispira as H. turris and considered Viviparus sclateri to represent a regional subspecies of H. japonica, based on the type illustration of V. sclateri. Later, Vivipara sclateri was generally considered a junior synonym of H. japonica (Yagura 1935, Kuroda 1947b, Kuroda 1955, Kuroda 1963. As Annandale (1921) indicated, H. longispira is clearly different from typical Vivipara sclateri; however, the early malacologists have considered the species to occur in other regions beyond Lake Biwa (Kobelt 1879, Iwakawa 1895, Iwakawa 1897a, Iwakawa 1897b, Iwakawa 1919, Pilsbry 1902, Annandale 1916, although this conclusion remains uncertain. Iwakawa (1919) listed 12 museum lots from 10 localities labelled with 'Naga-tanishi', the current Japanese name for H. longispira, which was at the time equivalent to Viviparus sclateri and three of the 10 localities are outside Lake Biwa drainage. Although Kuroda (1929) said "these may be elongated H. japonica" without any examination, these "H. longispira" from areas outside Lake Biwa drainage in Iwakawa (1919) have not been sufficiently examined and have not been illustrated to date. In addition, there were two records of "H. longispira" from the southern part of Lake Biwa and the Seta River flowing from Lake Biwa; however, H. longispira is now very rare in this area (Nishino 1991, Kihira et al. 2003, Kihira et al. 2009, Nakai 2016. Furthermore, one locality of "H. longispira" was drained and is now terrestrial (Ozawa 2012) and the viviparid gastropods from these localities have not been examined. In this study, to clarify the actual status of "H. longispira" (=Viviparus sclateri) from areas outside the current distribution and to classify extinct populations, we examined the collection of Tomotaro Iwakawa from the National Museum of Nature and Science, Tokyo.  Figure 1.

Sample
A map of collected localities of Viviparus scalateri in Iwakawa's collection . Yellow parts show the estimated ranges, based on the label of the museum lots. Numbers indicate the sample numbers in this study (see also Table 1). Sample numbers 2964J and 2964L were collected from somewhere in Lake Biwa, but the entire Lake is not coloured to make it easier to see. Green letters and downstream of the Seta River are Lake Biwa drainage as defined in this paper. A map is created from digital national land information (Ministry of Land, Infrastructure and Transport of Japan: https://nlftp.mlit.go.jp/ksj/index.html).

Materials and methods
All samples were from the mollusc collection of the National Museum of Nature and Science, Tokyo (NSMT-Mo). Specimens and labels were photographed using a digital single-lens reflex camera with a macro lens. Specimens were compared and identified by T. Saito based on mainly the references included in the full list of each species/subspecies synonymy (Suppl. material 1). In particular, the original descriptions of Japanese viviparid species and following references provided principal criteria for comparison and identification: Annandale (1921), Okada and Kurasawa 1950, Habe 1973, Masuda and Uchiyama (2004, Kihira et al. (2009), Hirano et al. (2015, Hirano et al. (2019a) and Hirano et al. (2019b). In addition, some specimens seemed to have an intermediate morphology and were tentatively identified to the species with which they are morphologically most similar.
The synonymy sections in the following text list only the first references for each combination of generic and specific names. As there are many references to Japanese viviparid species, it was impossible to list them all and also to establish objective quantitative criteria for selection. However, all references having the specific names, sclateri and ingallsiana from Japan were listed in the synonymy because of the main focus of this paper. To enhance reproducibility, we provide all the references that we examined in the synonymy list of Suppl. material 1.
Shell width (SW) of each specimen was measured using a Vernier micrometer (instrumental error: ± 0.03 mm). Furthermore, to compare quantitatively the shell shape of Heterogen species in Iwakawa's collection, we conducted elliptic Fourier (EF) analysis (Kuhl and Giardina 1982). The analysis was performed by Momocs 1.2.9 (Bonhomme et al. 2014) under R. 3.5.1 (R Core Team 2018). The EF coefficients were obtained from the images of 83 adult specimens in Iwakawa's collection and ten specimens figured in the published references (Frauenfeld 1865, Pilsbry 1902, Kobelt 1909, Annandale 1921. In addition, 143 images of Heterogen spp. studied by Hirano et al. (2019b) were also provided by these authors for analysis. The images were binarised before tracing shell outlines and the subsequent geometric normalisation was performed with Momocs. The number of harmonics was set to 40. Then, to summarise the results, a principal component analysis (PCA) was conducted using the obtained EF coefficients. Finally, the results of PCA and SW were graphed using Momocs and ggplot2 (Wickham 2016).

Data resources
Eleven out of 12 lots listed in Iwakawa (1919) were preserved in the National Museum and each lot contained several specimens ( Fig. 1

Diagnosis
Adult shell large and thick; shell shape subconical to pyramidal. Adult shell colour dark brown or brown or dark olive, often covered with many environmental attachments, such as alga; shell surface glossy, sometimes having weak hollows, growth lines and spiral striae. Suture moderately deep. Young shell small, relatively thin and fragile; shell shape pyramidal. Young shell colour yellowish-olive; shell surface quite glossy with no spiral ridges on the upper part of the spire.

Taxon discussion
Heterogen japonica can be distinguished from other Japanese Viviparidae by the following criteria which are based on the illustrations and information from literature (Suppl. material 1).
H. japonica can basically be distinguished from Cipangopaludina chinensis laeta by its subconical to pyramidal shell shape (Kihira et al. 2009; Fig. 2). In addition, the spire is higher, the spire angle is narrower and the shell apex is more pointed than those of C. c. laeta. Furthermore, C. c. laeta often has a lipped aperture, but the aperture of H. japonica does not have a pronounced lip (Masuda and Uchiyama 2004). Next, the differences with C. chinensis chinensis are not fully revealed, as the shell morphology of C. c. chinensis in Japan was not examined in depth. Nevertheless, based on the analysis and illustration of Hirano et al. (2019a), the only study that explicitly examined Japanese C. c. chinensis, the shell shape of C. c. chinensis is similar to that of C. c. laeta except for the pointed shell apex of C. c. chinensis. In any case, C. c. chinensis is distributed only around Kyushu Island, the south-western part of Japan (Hirano et al. 2019a) and no specimens of this subspecies were included in our study.
Adult Sinotaia quadrata histrica is generally smaller than H. japonica. There is a high morphological diversity in the shell shape of S. q. histrica (Kihira et al. 2009); however, the body whorl and the spire whorl are more rounded and arched than that of H. japonica. In addition, the shell apex of S. q. histrica is more rounded and the aperture is relatively smaller than those of specimens of H. japonica having the similar shell shape, as the spire is dense. Moreover, the shell colour of S. q. histrica seems to be brighter than that in H. japonica and often has a yellowish colour.
H. japonica differs in shell shape from H. longispira (Hirano et al. 2019b). Quantitatively, H. japonica has a lower spire and a broader spiral angle than H. longispira (Kihira et al. 2009). In addition, H. longispira has a strong basal angulation even on the adult shell (Okada and Kurasawa 1950). Furthermore, the body whorl and the spire whorl of H. longispira are linear (like vertical) in lateral view and the upper periphery of the whorls turns sharply to nearly horizontal. Accordingly, the suture is quite deep and whorls have a strong shoulder just below the suture. This feature is extremely pronounced in juveniles and young shells and this morphological difference is diagnostic between two species (Okada andKurasawa 1950, Hirano et al. 2019b). In addition, a difference that rarely appears in the shell shape is the distinct spiral ridges on the upper whorl of adult H. longspira (Hirano et al. 2019b). The early whorl of adult H. japonica does not have such spiral ridges. Moreover, Heterogen sp. has weak spiral ridges on the upper whorls and so this is a distinguishing feature from Heterogen sp., which is not distinguished by the shell shape (Hirano et al. 2019b). The spiral ridges of H. longispira are also pronounced on the body whorl of the adult shell. Furthermore, some H. japonica usually has greenish shells without reddish colour (e.g. Fig. 2h), whereas the shell colour of most H. longispira contains reddish colour and accordingly are often brown or dark brown in colour.
Besides, H. japonica has one morphotype, ver. iwakawa, which had been described and synonymised. This morphotype has a pyramidal shell shape, a broad spire angle and a strong basal angulation (e.g. Fig. 2d). Typical specimens of Viviparus japonicus var. iwakawa (Pilsbry 1902) is very easy to distinguish from any other Japanese Viviparidae; however, the morphology is continuous with H. japonica (Okada and Kurasawa 1950).

Notes
Heterogen japonica was identified in all 11 studied lots (Figs 2, 3). Some specimens (Fig. 2b- Fig. 3c and g-h). Therefore, in this study, we treated only specimens that were clearly distinguished from H. longispira (e.g. Fig. 3a-b and l) as H. japonica.

Diagnosis
Adult shell moderately large and very thick; shell shape subconical. Adult shell dark brown or greenish-brown sometimes covered with attachments; shell surface often having growth lines and several spiral ridges. Suture deeper than in any other viviparid from Japan; whorls strongly shouldered. Young shell small and relatively thin, but not fragile; shell shape pyramidal to subconical. Young shell bright olive; shell surface quite glossy having strong spiral ridges on upper part of spire. The shell size upon birth is larger than that of any other viviparid gastropod in Japan.

Taxon discussion
Heterogen longspira can be distinguished from other Japanese Viviparidae by criteria based on the illustrations and information from literature (Suppl. material 1). In particular, the juvenile and young shell have strong spiral ridges on the upper part of each whorl, this feature having crucial diagnostic value (Okada andKurasawa 1950, Hirano et al. 2019b). Furthermore, the shell size of the juvenile upon birth is larger than that of any other viviparid gastropods in Japan.
Firstly, H. longispira is easily distinguished from Cipangopaludina species/subspecies in Japan, based on the pyramidal shell shape, the higher spire, the narrower spire angle and the linear body whorl (Kihira et al. 2009; Fig. 4).
Besides, adult Sinotaia quadrata histrica is generally smaller than H. longispira. There is a high morphological diversity in the shell shape of S. q. histrica (Kihira et al. 2009); however, the body whorl and the spire whorl are much more rounded and arched than those of H. longispira. In addition, the suture of S. q. histrica is shallower than that of H. longispira. The upper whorl of adult S. q. histrica does not have the pronounced spiral ridges. Moreover, the shell colour of S. q. histrica is brighter than that in H. longispira and often has a yellowish colour.
Both adult H. longispira and Heterogen sp. have the spiral ridges on the upper whorl (Hirano et al. 2019b), but the ridges of H. longispira are stronger than Heterogen sp. In addition, the shell shape of Heterogen sp. is indistinguishable from that of H. japonica and then there are some differences in the shell shape between H. longispira and Heterogen sp. (Hirano et al. 2019b; refer to taxon discussion on the section of H. japonica for the difference).

Notes
Heterogen longispira was identified in four of the 11 lots examined, which were all from Lake Biwa drainage only (Figs 3, 4). Lots Mo2969 and Mo2970 contained non-typical H. longispira with an intermediate morphology to H. japonica and H. longispira (e.g. Fig. 3c and g-k). The shell shapes of these specimens tend to be slightly more similar to H. japonica; however, they have several distinctive morphological features of H. longispira, namely, the pronounced spiral ridges, the shouldered whorl, the deep suture, the strong basal angle and the high spire . Therefore these specimens were tentatively identified as H. longispira here (see also discussion).
Based on the previous taxonomy (Kuroda 1929), synonymy included the species-group name turris (and incorrect subsequent spellings) in addition to the species-group name longispira. For species-group names ingallsiana and sclateri, refer to the discussion.

Diagnosis
Adult shell small and thin, but not fragile; shell shape subconical to pyramidal. Adult shell bright brown or yellowish-or reddish-brown often covered with many attachments such as alga; shell surface slightly glossy, weak growth lines and spiral lines are usually present. Suture shallow; spire rounded. Young shell small, very thin and fragile; shell shape pyramidal with rounded angle. Young shell bright olive; shell surface quite glossy with no spiral ridges on upper part of the spire.

Taxon discussion
Sinotaia quadara histrica can be distinguished from other Japanese Viviparidae by the following features, based on the illustrations and information from literature (Suppl. material 1). In particular, the adult shell width tends to be relatively smaller than that of other Japanese Viviparidae (Masuda and Uchiyama 2004; for example, under 24 mm except for one exception in Kagawa et al. (2019)).
Firstly, S. q. histrica is easily distinguished from Cipangopaludina species/subspecies in Japan, based on the small shell width, pyramidal shell shape and the higher spire (Fig. 5).
The size of Heterogen sp. is larger and the apical whorls are also larger than those of S. q. histrica. In addition, Heterogen sp. has the weak spiral ridges on even the early whorls, whereas S. q. histrica does not have them. Diagnostic differences between S. q. histrica vs. H. japonica and H. longispira have been provided above, so we refer to the taxon discussions of these latter two species.

Notes
Sinotaia quadara histrica was identified in two of the 11 lots examined, which were from Lake Suwa (Fig. 5). Some specimens of H. japonica from Lake Suwa have a similar shell shape to S. q. histrica, but S. q. histrica has a more rounded shell apex, a broader spire angle, shallower sutures and a brighter colour than those of H. japonica.
We treated only the Japanese Sinotaia species for synonymy, as the taxonomic relationship between the continental Sinotaia and the Japanese Sinotaia species is not clear and not the point of this study. The species-group name, nitens was synonymised with a question mark by Pilsbry (1902) and we followed this treatment. Vivipara lacustris in Kawamura (1918) was documented as a small species that is distributed in the Kyushu Region. In the past, S. q. histrica was considered to be distributed in Kyushu Region (Pilsbry 1902) and so V. lacustris in Kawamura 1918 may be a junior synonym of S. q. histrica.

Analysis
The principal results of the PCA of EF analysis are shown in Fig. 6 and Fig. 7   Representative shells of Sinotaia quadrata histrica (Gould, 1859) from Japan, deposited in the collection of the National Museum of Nature and Science, Tokyo, which were treated as Viviparus sclateri (Frauenfeld, 1865) Hirano et al. (2019b), but some were located within the range of H. longisira from Hirano et al. (2019b) or in parts of the morphospace not covered by either. Some of H. longispira in Iwakawa's collection, which were identified qualitatively, were within the morphological range of H. longispira from Hirano et al. (2019b), but some were located within the morphological range of H. japonica from Hirano et al. (2019b) or in parts of the morphospace not covered by either.

Discussion
All eleven lots in the Iwakawa collection in the NSMT were previously thought to consist of a single species, Viviparus sclateri in Iwakawa (1919), but now have been found to consist of at least three species/subspecies, H. japonica, H. longispira and Sinotaia quadrata histrica. H. japonica was found in all lots studied during our investigation; H. longispira was found amongst four of the lots, which were all from Lake Biwa drainage (Table 1 and Fig.  1). This result reflects the taxonomic understanding of the period, prior to H. longispira being recognised as a distinct species (as H. turris in Annandale (1921)). Furthermore, all specimens of Viviparus sclateri in Iwakawa (1919) from outside of Lake Biwa drainage differed from H. longispira (Fig. 2a-h and Fig. 5), because these specimens do not have spiral ridges on the upper part of each spire which is a distinctive character of H. longispira (Figs 2, 4;Annandale 1921, Kihira et al. 2003. Nevertheless, almost all of these Plots of principal component axis one versus shell width for specimens from Iwakawa and Hirano et al. (2019b). The morphospace occupation by specimens from Hirano et al. (2019b) is also shown as coloured polygons.
specimens seem to have a higher spire and narrower spire angle than H. japonica as illustrated and documented in literature (Suppl. material 1; e.g. Figure 2 in Hirano et al. 2015; Figure 1 in Hirano et al. 2019b). The higher spire and the narrower spire angle of the H. japonica that was listed as Viviparus sclateri in Iwakawa (1919) from outside of Lake Biwa basin are noteworthy. Our morphological analysis showed that some specimens of Viviparus sclateri in Iwakawa's collection differ in shell shape from specimens of H. japonica in Hirano et al. (2019b), which were collected from the entire native geographic range of the species (Fig. 6). The shell shapes of these specimens seemed to be more elongated (i.e. having the higher spire and the narrower spire angle) and some could not be distinguished from the shell shape of H. longispira in Hirano et al. (2019b), based on PC1 and PC2 of the EF analysis. The similarities of the shell shapes shown by the EF analysis may suggest the complexity of the taxonomic relationship between the two species. Nevertheless, most of them had the shallower suture and this seemingly differed from H. longispira. Furthermore, most specimens of H. japonica in Iwakawa's collection had a smaller shell width than H. japonica in Hirano et al. (2019b) ( Fig. 7). This characteristic morphology seemed to be particularly abundant in H. japonica from Lake Suwa (2960J-2962J; e.g. Fig. 2c). The distinctive (see diagnosis and taxon discussion) shell morphology of H. longispira is considered to be a consequence of adaptation to its ancient lake habitat and this evolutionary change in morphology may have originated multiple times, based on examination of fossil specimens (Hirano et al. 2019b). Furthermore, plastic and/or adaptive morphological changes to the environment seem to occur easily both within and amongst species in viviparid gastropods (Hirano et al. 2019b, Kagawa et al. 2019, Stelbrink et al. 2020. Such morphological changes might have occurred in those H. japonica that have the elongated shell shape, as two of the three localities from where these specimens were collected are large lakes. In particular, Lake Suwa, where many distinctive specimens were found, was formed around 0.10 Ma (Anma et al. 1990) and H. japonica from Lake Suwa in Iwakawa's collection may represent an evolutionary distinct population.
On the other hand, Sinotaia quadrata histrica specimens were identified only from Lake Suwa (Fig. 5). These specimens were included in Viviparus sclateri of Iwakawa (1919). Fig.  6 of Iwakawa (1897a) and Figs. 5-6 of Iwakawa (1897b) shown as young specimens of Paludina ingallsiana may also be S. q. histrica from Lake Suwa. The existence of S. q. histrica in Lake Suwa might have resulted in further taxonomic confusion. Sinotaia q. histrica was considered to be only distributed in the southern part of Japan during that period (Pilsbry 1902, Kuroda 1929. Perhaps, S. q. histrica in the Iwakawa's collection may be the oldest record of the species from the eastern part of Japan. Now, S. q. histrica is a common viviparid gastropod throughout Japan, except for the Ryukyu Islands (Masuda and Uchiyama 2004). The species was considered to have been introduced after the prehistoric era, based on molecular phylogenetic studies and records of shell middens' records (Kurozumi 2001, Hirano et al. 2015, Kurozumi 2019. In a recent study, further complicated history of colonisation of S. q. histrica was estimated by genetic markers and multiple colonisations from the continent at different times were revealed (Ye et al. 2020).
The time when population established in the eastern part of Japan was estimated around 7910 years ago in this paper. However, the distribution of S. q. histrica has been considered to have recently expanded to the eastern part of Japan because there were few old records including from shell midden (Matsuoka 2001). Owing to thesemissing records, the history of the expansion of distribution have not been sufficiently clarified. Our probable oldest record from the eastern part of Japan may have implications for this, although further molecular and bibliographical studies are needed.
All samples from Lake Biwa drainage include H. japonica, though some lots contain both H. japonica and H. longispira (Table 1 and Fig. 1). The historically-documented coexistence of the two species differs from the current distribution of the two viviparid gastropods around Lake Biwa; H. japonica is rarely found within Lake Biwa at present (Horie 1971, Hirano et al. 2019b). Previous studies have stated that H. longispira is relatively common in the shallow area of southern Lake Biwa (Kuroda 1947a, Oyama and Kajiyama 1959, Kihira and Matsuda 1990. Although there is a possibility that each specimen was collected from several populations (including outside of the lake population), the co-existence of the two species in the same lot may be indicative of the past sympatric distribution of these species in Lake Biwa. Furthermore, some specimens from southern Lake Biwa and the Seta River flowing from Lake Biwa showed an intermediate morphology between H. japonica and H. longispira (Figs 3, 6; see also notes in taxon treatments of H. longispira). Some specimens, especially young and juvenile specimens, are easily assigned to one of the two species, based on the existence of spiral ridges and shell shapes (H. japonica: Fig. 3a, f and l; H. longispira: Fig. 3e, k and m). However, some specimens are difficult to distinguish. Namely, they have several distinctive morphological features of H. longispira (i.e. the pronounced spiral ridges, the shouldered whorls, the deep suture, the strong basal angle and the high spire), but the shell shapes of these specimens tend to be more similar to H. japonica (Fig. 6). In particular, many specimens from the Seta River have a distinctive morphology with a linear shape and smaller shell size, which is possibly indicative of a distinctive population (Figs 3, 6, 7). In addition, some specimens illustrated by past malacologists are presumed to be from southern Lake Biwa or the Seta River. For example, pl. 9, fig. 4 in Pilsbry (1902) (= pl. 16, fig. 7 in Kobelt 1909) from near Kyoto resembles specimens from the Seta River ( Fig. 3i and j), which flows to Kyoto City, and pl. 10, fig. 18 in Kobelt (1879) (= pl. 16, fig. 9 in Kobelt 1909) looks similar to specimens from southern Lake Biwa (Fig. 3b). In fact, our analysis of literature illustrations and other specimens showed that the shell shapes of these specimens were similar to southern Lake Biwa or the Seta River specimens, whereas the shell shapes of illustrations of Vivipara sclateri from the original description were far from the shape of H. longispira (Fig. 6b). A molecular study using genome-wide genetic data suggested introgressive hybridisation between the two species and some genetic populations seemingly had been generated by hybridisation (Hirano et al. 2019b). As such, hybridisation may have been the cause of the formation of the populations with an intermediate morphology. To reveal the traits and origin of these unique specimens, we need to examine current specimens carefully and include molecular analyses whenever possible. However, H. longispira of southern Lake Biwa and the Seta River are extinct or almost extinct (Nishino 1991, Kihira et al. 2003, Kihira et al. 2009, Nakai 2016. Similarly, populations of H. japonica of Lake Kasumigaura and Lake Suwa are critically endangered or extinct (Kurasawa 1988,