Follow the Footsteps of Leonardo Fea: An Example of an Integrative Revision of Freshwater Mussel Taxa Described from the Former British Burma (Myanmar)

N. Laverov Federal Center for Integrated Arctic Research of the Ural Branch of the Russian Academy of Sciences, Northern Dvina Emb. 23, 163069 Arkhangelsk, Russia Northern Arctic Federal University, Northern Dvina Emb. 17, 163002 Arkhangelsk, Russia Department of Zoology, Dawei University, Dawei, Tanintharyi Region, Myanmar Fauna & Flora International-Myanmar Programme, 34 D/9 San Yae Twin Street, Kaba Aye Pagoda Road, Bahan Township, 11201 Yangon, Myanmar Biology Department, University of new Brunswick, PO Box 5050, 100 Tucker Park Road, Saint John, NB, Canada E2L 4L5 Institute of Molecular Biotechnology (IMBT), Technical University of Graz, Petersgasse 14, 8010 Graz, Austria Zoological Survey of India, Telangana, Bhagat Singh Nagar, Hyderabad Plot 613, India


Introduction
First studies on freshwater mussels of the former British Burma (now Myanmar) have appeared as early as 1840s [1]. In particular, the American malacologist Augustus Addison Gould published the three earliest works with descriptions of a few unionid species from Burma [1][2][3]. These pioneering samples were collected and sent to him by Rev. Francis Mason, a Baptist missionary and naturalist [4]. During the period of 1860-1890s, the body of taxonomic literature on freshwater mussels of the region was increasing dramatically. Most of these historical works represent descriptions of new taxa and local faunal summaries [5][6][7][8][9][10][11][12][13], although Blanford [14] tried to compile the first checklist of freshwater mussels described from India, Ceylon (now Sri Lanka), and Burma.
The prominent Italian malacologist Cesare Maria Tapparone-Canefri published an extensive overview of freshwater and terrestrial molluscs of the former British Burma based on new samples collected by the Italian explorer and traveler Leonardo Fea [15]. Among other animal taxa, Fea collected a comprehensive freshwater mussel sample from six localities during his travels in 1885-1887 that was largely processed by Tapparone-Canefri [15]. Interestingly, the well-known Indian zoologist Baini Prashad criticized this work as follows: "Apparently Tapparone-Canefri had to base his work to a very great extent, if not entirely, on the incomplete published descriptions of the earlier authors and on the illustrations in the Conchologia Indica of Hanley and Theobald, for most of his identifications are incorrect, this would not have happened if he had had authentically named material for comparison. He referred all his new species to the composite genus Unio, and gave elaborate descriptions but did not publish any figures, his work, therefore, has been a great stumbling block in the way of all later work" ( [16]: 91). We would not agree with this disapproving decision, because the morphology-based taxonomy of freshwater mussels is complicated by multiple convergences in conchological traits, and it is always difficult to develop a taxonomic system for any group of the order Unionida based on morphological features alone [17][18][19][20][21][22].
Currently, our knowledge of the freshwater mussel systematics of Myanmar was largely enriched by the range of research using an integrative taxonomic approach [17][18][19][20][22][23][24][25][26][27][28][29][30]. Multiple species and several genera new to science were discovered and described, as well as broad-scale revisions of already described taxa were conducted [17,24,29,30]. It was shown that Myanmar harbors a species-rich and largely endemic freshwater mussel fauna and that this region could be considered a separate freshwater biogeographic domain, i.e. the Western Indochina Subregion [17,18]. However, many old works and corresponding museum collections are yet to be checked critically (e.g., [14][15][16]), although the crucial importance of these historical data can be recognized in order to better understand the range shifts and population dynamics, and, hence, to estimate the conservation status of freshwater mussels [19,22,31,32].
Taking into account the concerns, outlined above, this study (1) revises all the freshwater mussel taxa listed by Tapparone-Canefri [15] based on the primary descriptions, available museum specimens, and newly collected samples from the historical localities; (2) proposes new synonymies for several conchologically variable species such as Leoparreysia tavoyensis (Gould, 1843), Trapezidens dolichorhynchus (Tapparone-Canefri, 1889), Lamellidens generosus (Gould, 1847), and L. savadiensis (Nevill, 1877); (3) describes one freshwater mussel species new to science; and (4) discusses historical records on the market trade and commercial harvesting of freshwater mussels in the former British Burma.

Material and Methods
2.1. Data Sampling. The Fea collection that was used by Tapparone-Canefri [15] was partly deposited in the Museo Civico di Storia Naturale di Genova (MSNG) in Italy. We studied the MSNG collection but were unable to locate many shell lots that were mentioned by Tapparone-Canefri [15]. These lots were apparently transferred to the Indian Museum (ZSI: Zoological Survey of India) in Calcutta (now Kolkata) on the request of Baini Prashad. In the revision of Burmese Unionidae, Prashad ([16]: 92) noted that "…I found that it was quite impossible adequately to work out the Burmese forms without an examination of Tapparone-Canefri's type-specimens, and I applied to Dr. R. Gestro of the Genoa Museum. He was not only kind enough sent me the whole of Fea's Burmese collection on loan, but also generously presented to the Indian Museum specimens of a number of the species, duplicates of which were still available. This kindness on Dr. Gestro's part, for which I am greatly indebted to him, has made it possible for me to assign T.-Canefri's species to their proper generic and specific position". We were able to find and check several lots from the Fea collection in ZSI under the framework of the present study.
Molecular diagnoses for the new species described here was designed based on available sequences for Trapezoideus taxa [17,20,22] using a Toggle Conserved Sites tool of MEGA7 at 50% level [41] as decribed in our previous works [17,24]. The uncorrected COI p-distances between Trapezoideus species were also calculated through MEGA7 [41].

Phylogenetic Analyses.
We reconstructed a multi-locus phylogeny (3 codons of COI+16S rRNA+28S rRNA) using 120 haplotypes of the Unionidae from Myanmar (Table S1 and Alignment S1). Two representatives of the  [42] using an automatic identification of the best evolutionary models [43] and an ultrafast bootstrap (UFBoot) algorithm with 5,000 replicates [44]. Models of sequence evolution for each partition were calculated through Model Finder [45] based on Bayesian Information Criterion (BIC) and were as follows: 1 st codon COI -F81; 2 nd codon COI -GTR + G; 3 rd codon COI -TN + G + I; 16S -TIM2 + G + I; 28S -TIM3 + G + I. Bayesian inference analysis (BI) was carried out in MrBayes v3.2.7 [46] with two runs, each with three heated (temperature =0.1) and one cold Markov chain, during 50 × 10 6 generations and sampling every 1,000th generation. The first 15% of trees were discarded as burn-in. The calculation was performed at San Diego Supercomputer Center through the CIPRES Science Gateway [47]. A trace analysis tool (Tracer v. 1.7) was used to check a convergence of the MCMC chains to a stationary distribution [48]. The effective sample size (ESS) for all the parameters was recorded as >5,000.
The mitochondrial COI trees were reconstructed separately for the Lamellidentini, Leoparreysiini, Indochinellini, and Contradentini tribes (Figures S2-S5, Table S1, Alignments S2-S5). Each COI dataset was analyzed using maximum likelihood (ML) and Bayesian inference (BI) methods mainly as stated above, but for BI analysis two independent runs of 25 × 10 6 generations were sampled at intervals of every 1,000th generation. The best-fit nucleotide substitution models and the partition scheme are listed in Table S2.
2.5. Nomenclatural Acts. The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature (ICZN), and hence the new name and the taxonomic opinions contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank (http://zoobank.org), the online registration system for the ICZN. The LSID for this publication is: http://zoobank.org/urn:lsid:zoobank.org:pub: 49FCF27F-8C9B-4235-9701-C0FDB623FD0D. The electronic edition of this paper was published in a journal with an ISSN, and has been archived and is available from PubMed Central.

Morphological Traits and Shell
Proportions of Tapparone-Canefri's Taxa. Among specimens from the MSNG collection and the taxa descriptions from Tapparone-Canefri's work [15], we identified Indochinella pugio pugio, I. pugio paradoxa, Indonaia andersoniana, Radiatula chaudhurii and R. mouhoti haungthayawensis (Figure 2 In the MSNG collection, the type series of only four nominal taxa described by Tapparone-Canefri [15] as new to science are still available, i.e. Unio rectangularis ( Figure 6(d)), U. pulcher (Figure 4(a)), U. protensus var. obtusatus (Figure 4(e)), and U. marginalis var. subflabellata ( Figure 5(e)). Accurate comparative conchological analyses of these shells supported the opinion [22] that Unio rectangularis is more likely representative of the Contradentini, specifically the genus Yaukthwa, mainly based on the rectangular shell shape and wrinkles on the posterior slope.        Indochinella pugio pugio (Benson, 1862) Mandalay [AY-01] (new sample) RMBH biv 441/2 n/a n/a n/a 11 Journal of Zoological Systematics and Evolutionary Research Morphological features of Unio pulcher, U. protensus var. obtusatus, and U. marginalis var. subflabellata allow to attribute them to the genus Lamellidens. The two first nominal taxa most probably belong to Lamellidens generosus, representing different morphological forms of this species. Unio pulcher has a shell shape, which is typical for L. generosus, with a rather high wing, broad posterior margin, and a truncated slope. Unio protensus var. obtusatus shares an elongated shell with smooth dorsal margin. However, Unio marginalis var. subflabellata has characters that are more specific to Lamellidens savadiensis such as an elongated, usually curved posterior margin, well-pronounced growth lines, and a shallow umbo cavity.
Other available lots of Lamellidentini from the MSNG collection such as Unio marginalis var. tricolor, U. corrianus, and U. marginalis according to morphological features belong to Lamellidens savadiensis (based on the shell shape and growth lines), L. generosus (shell shape and teeth According to the PCA based on Fourier coefficients, four principal components (PCs) were obtained (Figure 7). A Kruskal-Wallis test revealed two significant components PC1 and PC2 (P <0.05), which explained 76.9% and 12.7% of the total shell shape variance, respectively. The PC1 reflects height of the shell and shifts from an ovalelongated contour to a contour with well-developed wing and curved ventral margin. PC2 shows variation in the       Figure S3).
Topotypes of Unio protensus var. obtusatus from the Nga Wun River near the city of Pyay (Ayeyarwady Basin) and U. pulcher from the Mitan Stream (Haungthayaw Basin) fall into the Lamellidens generosus clade ( Figure S4). Among the specimens of Lamellidens from Mandalay we found only representatives of L. savadiensis ( Figure S4). Topotypes of Unio rectangularis collected from the Mole Stream (Ayeyarwady Basin) phylogenetically correspond to the Yaukthwa clade ( Figure S5).
A species that was identified by Tapparone-Canefri [15] as Unio exolescens from the Mitan Stream, a tributary of the Haungthayaw River, is described here as Trapezoideus mitanensis sp. nov. It represents a separate phylogenetic lineage that is more closely related to Trapezoideus lenya Bolotov   (Table S3). Prospective topotypes of other nominal Tapparone-Canefri's taxa were also phylogenetically divided among representatives of the Indochinellini, Leoparreysiini, Lamellidentini, and Contradentini (Figure 8, Figures S2-S5 (Figure 2(a)).
(  Figure S3. Journal of Zoological Systematics and Evolutionary Research (8) Comments. Bolotov et al. [24] applied the name Lamellidens generosus to a lineage endemic to the Inle Lake drainage, Salween River basin. This taxonomic hypothesis was based on their data on the single-drainage pattern of freshwater mussel endemism throughout Myanmar. However, the number of single-drainage endemic species in the country was overestimated due to the limited coverage of river basins by sampling efforts [21]. It was found that Lake Inle and surrounding waters are inhabited by other species, Lamellidens ferrugineus (Annandale, 1918) (=Physunio micropteroides Annandale, 1918) [56]. In turn, Unio pulcher var. lamellatiformis, U. pulcher var. ponderosulus, and U. generosus var. (1) Type. Not traced.
(2) Type Locality. At Sawady in the Tengleng Stream, also at Bhamo and at Shuaygoomyo.
(6) Comments. Bolotov et al. [27] re-described and illustrated this species in detail. It was transferred to the genus Indonaia Prashad, 1918 based on conchological features. Later, Pfeiffer et al. [22] argued that this species belongs to the Contradentini, and, more specifically, is a member of the genus Yaukthwa. This morphology-based taxonomic hypothesis was supported by our new DNA sequences of topotypes from the Mole Stream. Furthermore, a Yaukthwa species from mountain rivers and streams of the Upper Ayeyarwady Basin identified to be conchologically similar to Y. dalliana (see [20,23]) is in fact Y. rectangularis. Furthermore, sequenced topotypes of Yaukthwa dalliana revealed that it is a valid species having a restricted range in a section of the Lashio River near the town of Lashio (Figures 6(f) and 6(g)). (Preston, 1912). =Unio foliaceus Tapparone-Canefri, 1889: 345 [15] (identification error). =Trapezoideus foliaceus var. zayleymanensis Preston, 1912: 307 [50].

Yaukthwa zayleymanensis
=Yaukthwa zayleymanensis (Preston, 1912). (7) Comments. Although a reliable identification of Yaukthwa taxa using conchological features alone is complicated, Tapparone-Canefri's specimen from the Mole Stream clearly corresponds to Yaukthwa zayleymanensis conchologically. This genus contains a number of local endemic species, the range of which is restricted by a certain freshwater basin or even to a tributary of a river or a lake [20]. Our new sequenced sample from the Mole Stream was similar to the Tapparone-Canefri's specimen based on morphological features, and was found to belong to Yaukthwa zayleymanensis. Figures 6(l)-6(n), and 8, Tables 2 and 3, Figure S5.

Taxonomic Summary.
Our results show that the proportion of correct identifications of freshwater mussel taxa in historical works such as that of Tapparone-Canefri [15] could even be lower than it was previously suggested (e.g., [16]). This evidence indicates that preparing checklists of freshwater mussel species by means of a direct compilation of historical data (e.g., [31,57]) may lead to unreliable species richness estimates biased by multiple identification errors. In summary, collection of freshwater mussels from British Burma examined and published by Tapparone-Canefri [15] contains 13 taxa that were listed as 34 nominal species (  (Table 1). Additionally, Tapparone-Canefri identified samples of Gibbosula laosensis and Yaukthwa nesemanni from a tributary of the Sittaung River as Unio sella, although this name has never been introduced as such but was repeatedly published as a synonym [28]. Currently, only two species are considered valid among the new taxa described by Tapparoni-Canefri [15], i.e. Trapezidens dolichorhynchus and Yaukthwa rectangularis ( Table 2). Trapezoideus mitanensis sp. nov. from the Haungthayaw River represents a fourth species in this small genus belonging to the tribe Contradentini [17,22]. This lineage has a high level of genetic divergence from its congeners, sharing diagnostic substititions in all the studied gene fragments.

Conchological
Variability and Synonymy of Leoparreysia tavoyensis. Among the studied taxa, Leoparreysia tavoyensis was found to be the most conchologically variable species. Six nominal species such as Unio luteus, U. bhamoensis, U. mandelayensis, U. feae, U. houngdarauicus, and Parreysia choprae are synonymized here with L. tavoyensis based on the study of topotypes by means of morphological, molecular, and phylogenetic analyses. Each of these morpho-species differs from others by having a more or less specific shell shape, sculpture, color and surface of periostracum or even structure of the teeth (Figure 3). Until today these taxa were not subjected by detailed taxonomic revision. Unio luteus was synonymised with Parreysia corrugata by Simpson [58] and since then its taxonomic position has not been changed by other malacologists [54,57,[59][60][61]. While revising the Tapparone-Canefri's [15] work, Prashad [16] recognized Unio bhamoensis, U. mandelayensis, U. feae, and U. houngdarauicus as separate species belonging to the genus Parreysia. Later Prashad [53] described the new species Parreysia choprae from Indawgyi Lake which till now had a status of valid species but within the genus Leoparreysia [24,26]. Haas [60] synonymized Unio mandelayensis and U. bhamoensis leaving only one valid species Parreysia bhamoensis. According to the range of subsequent works, this taxon was transferred to the genus Leoparreysia [21,24,26,62]. In return, Unio feae and U. houngdarauicus until recently had saved their status of separate taxa [24,26]. Though Haas [60] synonymized Unio houngdarauicus with Parreysia tavoyensis, P. feae from the same river basin was considered a valid species. Regarding Unio parma Sowerby, 1868, this taxon was transferred to P. tavoyensis by Prashad [16] and Haas [60]. However, the subsequent revisions were not conducted.
This fact that so many researchers recognised the validity of taxa, discussed above, confirmed the high conchological variability of Leoparreysia tavoyensis and difficulty in distinguishing it using morphological methods alone. This variable species is widespread throughout Myanmar, and it was recorded in almost all large and medium-sized river 29 Journal of Zoological Systematics and Evolutionary Research basins of the country. It is clear that a large amount of intraspecific conchological forms in Leoparreysia tavoyensis may reflect environmental gradients such as habitat (rivers, streams or lakes) and substrate type.

Historical Market
Trade of Freshwater Mussels in the Former British Burma. Another interesting issue of this study is that the samples of several species, i.e. Indonaia andersoniana, Indochinella pugio pugio, and Lamellidens generosus, were obtained by Leonardo Fea from a market (del mercato) in the city of Mandalay [15]. This evidence indicates that freshwater mussels were historically used in the market trade in Myanmar at least since the 1880s. One might think that the most appropriate usage of these mussels was for consumption, as food and cooking. However, it is strange that the contemporaries of that historical period, e.g. F. Mason and W. Theobald in their work ( [63]: p. 129-131), did not mention that the Unionidae were used for eating in Burma as opposed to marine and estuarine bivalves.
Currently, freshwater bivalves are actively harvested for food and market trade throughout Southeast and South Asian countries, i.e. Myanmar, Laos, Vietnam, Indonesia, and India [18,57,[64][65][66]. A range of species is used for local sale, production of decorative, art goods, and jewellery, as well as producing artificial freshwater pearls [67]. Allen et al. [68] noted that freshwater biological resources such as molluscs play a vital role in the everyday life of local communities. We could assume that freshwater mussels in the local markets in the former British Burma were used with similar purposes and were actively involved in trading.
Nowadays an international export of freshwater mussels for aquaria is also widespread. According to the recent work of Ng et al. [69], several freshwater mussel species endemic to Myanmar such as Leoparreysia olivacea and L. tavoyensis were recorded among molluscs involved in ornamental pet trade in Singapore during the period of 2008 to 2014.
Harvesting and over-exploitation are one of the biggest threats to freshwater mussel populations throughout Southeast Asia, especially to local endemic and rare species [65,67]. Our present study is one more part of the research helping to understand unionoid systematics in Myanmar and the Oriental tropics generally and to create a scientific basis for environmental management and conservation planning.

Data Availability
The morphometric, sequence, molecular and phylogenetic data used to support the findings of this study are included within the supplementary information files.

Conflicts of Interest
The authors declare no conflict of interest.  Table S1. List of sequences of the Unionidae from Southeast Asia used in this study. Table S2. The bestfit models of nucleotide substitution and partition scheme. Table S3. Genetic divergences (mean uncorrected p-distances, %) from Trapezoideus mitanensis sp. nov. and its congeners based on the mitochondrial COI gene sequences.