A new genus of air-breathing marine slugs from South-East Asia (Gastropoda, Pulmonata, Onchidiidae)

Abstract As part of an ongoing effort to revise the taxonomy of air-breathing, marine, onchidiid slugs, a new genus, Laspionchis Dayrat & Goulding, gen. nov., is described from the mangroves of South-East Asia. It includes two new species, Laspionchis boucheti Dayrat & Goulding, sp. nov., and Laspionchis bourkei Dayrat & Goulding, sp. nov., both distributed from the Malacca Strait to the Philippines and Australia. This study is based on extensive field work in South-East Asia, comparative anatomy, and both mitochondrial (COI and 16S) and nuclear (ITS2 and 28S) DNA sequences. The two new species are found in the same habitat (mud surface in mangrove forests) and are externally cryptic but are distinct anatomically. Both species are also strongly supported by DNA sequences. Three cryptic, least-inclusive, reciprocally-monophyletic units within Laspionchis bourkei are regarded as subspecies: L. bourkei bourkei Dayrat & Goulding, ssp. nov., L. bourkei lateriensis Dayrat & Goulding, ssp. nov., and L. bourkei matangensis Dayrat & Goulding, ssp. nov. The present contribution shows again that species delineation is greatly enhanced by considering comparative anatomy and nuclear DNA sequences in addition to mitochondrial DNA sequences, and that thorough taxonomic revisions are the best and most efficient path to accurate biodiversity knowledge.


Introduction
The diversity of invertebrate species in mangrove forests of South-East Asia is still largely unknown, mainly because mangroves have not been explored well enough, which likely has to do with the fact that mangroves are not the most inviting habitats, even for savvy field naturalists: mangroves are extremely muddy, infested with malariacarrying mosquitoes and pit vipers, and often located in remote areas. Our lack of biodiversity knowledge is a major issue not only because nobody knows exactly how many species live in the mangroves of South-East Asia, but also because mangroves are still being eradicated at a large scale across the entire region. Onchidiid slugs illustrate well this general situation: until recently, nobody knew exactly how many species of onchidiids lived in the mangroves of South-East Asia, even though they are some of the most common and diverse animals in mangroves (Dayrat 2009).
Onchidiids are marine, air-breathing, true slugs. Adult onchidiids live in the intertidal zone and their larvae develop in sea water, although a few species are adapted to high elevation tropical rainforest (Dayrat 2010). They breathe air through a lung and are related to land snails and slugs (Dayrat et al. 2011). They are called 'true' slugs because they lack an internal shell, even a vestigial shell. The only other group of marine, air-breathing, true slugs is the genus Smeagol Climo, 1980, which is not closely-related to onchidiids, but rather considered to belong to the Ellobiidae (Dayrat et al. 2011). The terrestrial, air-breathing, true veronicellid slugs are the most-closely related group to the onchidiids (Dayrat et al. 2011).
In the past ten years, our laboratory has worked on a global taxonomic revision of the Onchidiidae, one genus at a time (Dayrat et al. 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c), based on extensive collecting efforts worldwide, especially in South-East Asia where onchidiids have greatly diversified. The application of old generic names, such as Onchidium Buchannan, 1800, Onchidina Semper, 1882, Paraoncidium Labbé, 1934, and Peronina Plate, 1893, is now clear, and new genera are also being discovered: Alionchis Goulding & Dayrat in Goulding et al. 2018b, Marmaronchis Dayrat & Goulding in Dayrat et al. 2018, Melayonchis Dayrat & Goulding in Dayrat et al. 2017, Paromoionchis Dayrat & Goulding in Dayrat et al. 2019, and Wallaconchis Goulding & Dayrat in Goulding et al. 2018a In the present contribution, we describe a new genus, Laspionchis gen. nov., and two new species: Laspionchis boucheti sp. nov., distributed from the Malacca Strait eastwards to the Philippines and Queensland, Australia, and Laspionchis bourkei sp. nov., from the Malacca Strait eastwards to the Philippines and the Northern Territory, Australia. Three cryptic, least-inclusive, reciprocally-monophyletic units within Laspionchis bourkei are regarded as three subspecies: L. bourkei bourkei Dayrat & Goulding, ssp. nov., L. bourkei lateriensis Dayrat & Goulding, ssp. nov., and L. bourkei matangensis Dayrat & Goulding, ssp. nov. New taxon names are needed because no existing genus-group name applies to the clade described here and no existing species-group name applies to the species and subspecies described here.
The present study follows an integrative approach to taxonomy (Dayrat 2005), which is based on: (1) a comprehensive review of the nomenclature (all available types of onchidiid species were borrowed and re-examined); (2) the field observation of live animals in their natural habitats; (3) comparative anatomy; and (4) analyses of both mitochondrial (COI, 16S) and nuclear (ITS2, 28S) DNA sequences. The new genus described here is characterized by a combination of anatomical characters which is unique (not found in other onchidiid genera), and its monophyly is strongly supported in molecular phylogenetic analyses. Even though both species are cryptic externally, they are strongly supported by DNA sequences and internal anatomy.
Laspionchis slugs live on the surface of the mud in mangrove forests, where they cooccur with many other onchidiid species with a similar appearance. Laspionchis slugs are most especially difficult to distinguish externally from Paromoionchis slugs, which are found in the same habitats and the same geographical regions (Dayrat et al. 2019).

Collecting
All specimens were collected by the authors in the last few years. Collecting parties were led by Benoît Dayrat in Brunei Darussalam, Malaysia, Northern Territory (Australia), Philippines, and Singapore, by Tricia Goulding in Queensland (Australia) and Vietnam, and by Munawar Khalil in Indonesia. We often were accompanied by local villagers or fishermen. Sites were accessed by car or by boat. Each site was explored for an average of two hours, but the exact time spent at each site also depended on the time of the low tide, the weather, etc. At each site, photographs were taken to document the kind of mangrove being visited as well as the diverse microhabitats where specimens were collected.
In the field, specimens were individually numbered and photographed in their habitat. At each site, we tried our best to sample as much diversity as possible. In addition to numbering individually the specimens that looked different, we also numbered individually many specimens that looked similar so that we could test for the presence of cryptic diversity. Importantly, a piece of tissue was cut for all specimens individually numbered (for DNA extraction) and the rest of each specimen was relaxed (using magnesium chloride) and fixed (using 10% formalin or 70% ethanol) for comparative anatomy.

Specimens
All available types of Onchidiidae were examined. Many worldwide museum collections were visited (but no Laspionchis material was found). Sixty-one specimens of Laspionchis are included in this study: 23 specimens of L. boucheti and 38 specimens of L. bourkei. Each specimen was examined for comparative anatomy and sequenced for molecular phylogenetic analyses. Individual DNA extraction numbers used in the phylogenetic analyses are indicated in the lists of material examined (numbers are be-tween brackets, and a capitalized letter H indicates a holotype), and size (length/width) is indicated in millimeters (mm) for each specimen. All specimens were deposited as vouchers in institutions in the countries of origin.
Both the external morphology and the internal anatomy were studied. All anatomical observations were made under a dissecting microscope and drawn with a camera lucida. Radulae and male reproductive organs were prepared for scanning electron microscopy (Zeiss SIGMA Field Emission Scanning Electron Microscopy). Radulae were cleaned in 10% NaOH for a week, rinsed in distilled water, briefly cleaned in an ultrasonic water bath (less than a minute), sputter-coated with gold-palladium, and examined by SEM. Soft parts (penis, accessory penial gland, etc.) were dehydrated in ethanol and critical point dried before coating.
The anatomy of L. boucheti, the type species, is fully detailed. The written description of the many anatomical features that are virtually identical between species (nervous system, heart, etc.) is given only for the type species to avoid repetition. So, any feature that is only mentioned in L. boucheti is identical in the other species. The color of live animals is described in detail for both species in order to demonstrate the overlapping individual variation between species. As expected, differences between species are mostly found in the male copulatory apparatus, which is described and illustrated in detail for each species. Special attention has been paid to illustrating the holotype of each of the species and subspecies, and the plates illustrating habitats also include a picture from type localities.

Intestinal types
Now that the types of intestinal loops have been reported for every species in many genera of onchidiids (Dayrat et al. 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c), it is possible, and actually necessary, to clarify the differences between the various types of intestinal loops. Plate (1893: pl. 8, figs 29-32) first distinguished four types of intestinal loops (types I to IV) and Labbé (1934: 177-178, fig. 3) later added a type V. However, the pattern of intestinal loops varies, both intra-specifically and inter-specifically. The differences between intestinal types are not as sharp as Plate and Labbé assumed they were, and now they must be clarified.
Here we provide a new approach to help reliably determine intestinal types. Because the intestinal loops found in Laspionchis are between type I and type II, we focus here on types I and II. This new approach is based on recognizing three different sections in intestinal loops, each section being colored differently: a clockwise loop is colored in blue, a counterclockwise loop in yellow, and a transitional loop between them in green ( Fig. 1). For the sake of clarity, Plate's (1893: pl. 8, figs 29, 31) original illustrations of his types I and II are reproduced here (Fig. 1A, C).
The onchidiid types of intestinal loops are defined based on the dorsal pattern of the intestine. The intestine always first appears dorsally on the right side. In a type I (Fig. 1A, B), the intestine starts by forming a clockwise loop (blue loop). This clockwise loop, however, does not form a complete circle and soon transitions into a counterclockwise loop (yellow loop). As a result, the transitional loop (green loop) between the clockwise and counterclockwise loops is oriented to the right, at 3 o'clock (horizontal red arrow). In a type II (Fig. 1C, D), the clockwise loop (blue loop) is longer and rotates more than in a type I and, as a result, the transitional (green) loop is oriented to the left at 9 o'clock (horizontal red arrow). There is, as always, individual variation. Most usually, the orientation of the transitional loop varies within a range of ca. 90 degrees around a mean axis (i.e., 45 degrees on either side of that axis). So, for instance, in Paromoionchis tumidus (Semper, 1880), the species that illustrates a typical type II (Fig. 1D), the transitional loop is always oriented to the left but is not always perfectly horizontal (at 9 o'clock); it can be descending (down to approximately 7 o'clock) or ascending (up to approximately 11 o'clock) (Dayrat et al. 2019: fig. 12).
In Laspionchis (Fig. 1E-H), the average orientation of the transitional loop is exactly between a typical type I and a typical type II. Indeed, in most Laspionchis individuals, the transitional loop is descending vertically at 6 o'clock ( Fig. 1F, G). So, intestinal loops of Laspionchis slugs cannot be assigned to either type I or type II. Naturally, there is individual variation (Fig. 1E-H): every specimen listed in the material examined was dissected to check its intestinal loops. In some cases, the intestinal loops appear to be of type II, with the transitional loop oriented to the left and descending at approximately 7 o'clock (see the red arrow in Fig. 1E). In some other cases, the intestinal loops appear to be of type I, with the transitional loop oriented to the right and descending at approximately 5 o'clock (see the red arrow in Fig. 1H). In the individuals examined for the present study, the transitional loop is not higher than 5 o'clock on the right (i.e., it is not oriented at 4 or 3 o'clock) and not higher than 7 o'clock on the left (i.e., it is not oriented at 8 or 9 o'clock). So, the intestinal loops of the two known species of Laspionchis are exactly between types I and II. Instead of creating a new intestinal type (number VI), the intestinal loops of Laspionchis are simply and adequately referred to as "between types I and II."
In addition, another set of analyses was performed with only COI sequences: genetic distances between COI sequences were calculated in MEGA 6 as uncorrected p-distances. COI sequences were also translated into amino acid sequences in MEGA using the invertebrate mitochondrial genetic code to check for the presence of stop codons (no stop codon was found).

Molecular phylogenetic analyses
DNA sequences were used to test species limits within Laspionchis. The monophyly of Laspionchis is strongly supported in all analyses (Figs 2-5). In the analyses based on mitochondrial COI and 16S concatenated sequences, there are four least-inclusive units that are all reciprocally monophyletic: L. boucheti, L. bourkei bourkei, L. bourkei lateriensis, and L. bourkei matangensis (Fig. 2). The monophyly of each unit is strongly supported by a bootstrap support of 100 and a posterior  Only numbers > 60% (ML) and > 0.9 (Bayesian) are indicated. Numbers for each individual correspond to unique identifiers for DNA extraction. All sequences of Laspionchis individuals are new. Sequences of the outgroups are from our previous studies (Dayrat et al. 2011(Dayrat et al. , 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Information on specimens can be found in the lists of material examined and in Table 1. The color used for each subspecies is the same as the color used in Figs 3-7.   (Dayrat et al. 2011(Dayrat et al. , 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Information on specimens can be found in the lists of material examined and in Table 1. The color used for each subspecies is the same as the color used in Figs 2, 4-7.   (Dayrat et al. 2011(Dayrat et al. , 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Information on specimens can be found in the lists of material examined and in Table 1. The color used for each subspecies is the same as the color used in   (Dayrat et al. 2011(Dayrat et al. , 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Information on specimens can be found in the lists of material examined and in Table 1. The color used for each subspecies is the same as the color used in Figs 2-4, 6, 7. . Diagram to help visualize the data on pairwise genetic distances between COI sequences within and between mitochondrial units in Laspionchis (see Table 3). Ranges of minimum to maximum distances are indicated (in percentages). For instance, within L. boucheti, individual sequences are between 0 and 2.5% divergent; individual sequences between L. boucheti and the other units are minimally 7.5% and maximally 10.4% divergent; overall, the distance gap between L. boucheti and L. bourkei is of 5% (i.e., between 2.5% and 7.5%). The colors are the same as those used in Figs 2-5, 7.

Comparative anatomy
In the field, slugs were numbered individually without being assigned to any particular species because onchidiid species are commonly cryptic externally. As anticipated, Laspionchis boucheti and L. bourkei are externally cryptic (Table 3). However, Laspionchis boucheti differs from L. bourkei in internal anatomy, and they cannot be confused: in L. boucheti, the long retractor muscle of the penis inserts at the posterior end of the visceral cavity, while the retractor muscle is absent, vestigial, or short (inserting in the first third of the visceral cavity) in L. bourkei. Also, additional, distal, retractor muscle fibers are present in L. boucheti but absent in L. bourkei. However, the three subspecies of L. bourkei are hardly distinguishable anatomically (Table 3). Table 2. Intra-and inter-unit pairwise genetic distances between the four mitochondrial units of Laspionchis based on our data set of 61 COI sequences (Table 1). Ranges of minimum to maximum distances are indicated (as percentages): e.g., individual sequences within L. boucheti are between 0 and 2.7% divergent, and individual sequences between L. boucheti and L. bourkei bourkei are minimally 8.6% and maximally 10.4% divergent. Overall, the distance gap between all four mitochondrial units is between 2.5% (the maximum intra-unit distance within L. boucheti and within L. bourkei matangensis) and 3.9% (the minimum distance between L. bourkei lateriensis and L. bourkei matangensis). Finally, the distance gap between the two species L. boucheti and L. bourkei is between 2.5% and 7.5%.

Species delineation
The new genus described here, Laspionchis, is a strongly-supported clade in all molecular analyses (Figs 2-5). It also is characterized by a unique combination of anatomical characters (see below, the Remarks on the genus diagnosis). Two species are recognized here, Laspionchis boucheti and L. bourkei, which are cryptic externally but distinct ana-tomically (Table 3). Their reciprocal monophyly is strongly supported by both nuclear and mitochondrial sequences (Figs 2-5) and they are separated by a clear barcode gap (from 2.5% to 7.5%) in genetic distances between COI sequences (Table 2, Fig. 6).
In addition, three subspecies are recognized within Laspionchis bourkei: L. bourkei bourkei, L. bourkei lateriensis, and L. bourkei matangensis, which are cryptic externally and hardly distinguishable internally (Table 3). Their reciprocal monophyly is strongly supported by mitochondrial sequences as well as by nuclear sequences (Figs 2-5) even though L. bourkei matangensis is unresolved in nuclear analyses. All three subspecies of L. bourkei are separated by a barcode gap in genetic distances between COI sequences (from 2.5% to 3.9%) which, as expected, is not as large as the gap found between L. boucheti and L. bourkei (Table 2, Fig. 6). The ranking of the three least-inclusive units within L. bourkei as subspecies is discussed in the general discussion. Etymology. Combination of láspi, a Greek word meaning mud, and onchis, a word derived from the Greek ὁ ὂγκος (mass, tumor) and used in the past for onchidiid slugs. Laspionchis conveniently refers to those onchidiid species that always live on mud and are covered with a thin layer of mud.

Systematics and anatomical descriptions
Gender. Gender masculine of onchis (ICZN Art. 30.1.1), a word derived from the masculine Greek word ὁ ὂγκος.
Diagnosis. Body not flattened. No dorsal gills. Dorsal eyes present on notum. Retractable, central papilla (usually with four dorsal eyes) present, often raised above dorsal surface. Eyes at tip of short ocular tentacles. Male opening below right ocular tentacle (or below it and very slightly to its left). No transversal protuberance on oral lobes. Foot wide. Pneumostome median, on ventral hyponotum. Intestinal loops exactly between types I and II (with a transitional loop on average descending at 6 o'clock). Rectal gland absent. Accessory penial gland present with a hollow spine and a muscular sac. Penis with hooks: numerous, densely arranged next to each other, and pointed.
Remarks. No external diagnostic feature unambiguously distinguishes Laspionchis from other onchidiid genera. Externally, Laspionchis slugs are especially difficult to distinguish from Paromoionchis slugs, which live in the same habitat (mud surface) and are often found together at the exact same sites. Also, for a non-expert, Laspionchis slugs could easily be confused with Peronina or Onchidium slugs, although those are characterized by distinctive, external features. However, Laspionchis is characterized by a unique combination of internal and external characters: no dorsal gills, male opening below the right eye tentacle (or below it and very slightly to its left), no rectal gland, intestinal loops between types I and II (i.e., with a transitional loop on average oriented at 6 o'clock), accessory penial gland present with a muscular sac, penis with numerous, pointed hooks densely arranged next to each other. According to our data, any onchidiid slug with this combination of characters belongs to Laspionchis.
Intestinal loops between types I and II, with a transitional loop on average oriented at 6 o'clock, could almost be regarded as diagnostic of Laspionchis slugs, acknowledging the existence of variation (both intra-specific and inter-specific). Indeed, in Laspionchis slugs, the transitional loop is normally oriented at 6 o'clock, even though, strictly speaking, its orientation actually varies between 5 and 7 o'clock (Fig. 1). The intestinal loops of some individuals of other species can sometimes be characterized by a transitional loop oriented within that same range (between 5 and 7 o'clock), such as in species with intestinal loops of type I and a transitional loop oriented from 3 to 6 o'clock (as in Wallaconchis, see Goulding et al. 2018a), and in species with intestinal loops of type II and a transitional loop oriented from 6 to 9 o'clock (as in Paromoionchis, see Dayrat et al. 2019). However, the important difference here is that a transitional loop oriented at 6 o'clock is the norm in Laspionchis, while it is not the norm in those species from other genera.
A new generic name is needed because no existing name applies to the clade described here. Based on the examination of all the type specimens available in Onchidiidae (especially those of all the type species), a careful study of all the original descriptions (especially when no type specimens were available), and our ongoing taxonomic revision of every genus of the family (Dayrat et al. 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c), it appears that there is no generic name of which the type species matches the diagnosis of this genus. For a recent review of the application of all existing generic names of Onchidiidae, see Dayrat et al. ( : 1861.  Colored areas correspond to hypothetical geographical ranges based on known records. Habitat (Fig. 8). Laspionchis boucheti is found on mud, hard or soft, in open or dense mangrove forests. It is common across its entire distribution range.

Laspionchis boucheti
Etymology. Laspionchis boucheti is dedicated to Philippe Bouchet, professor of Malacology at the Muséum national d'Histoire naturelle, Paris, France, for the training that he generously provided to the first author as a graduate student at the MNHN, years ago, for kindly allowing us to study some material collected during expeditions that he organized (Kavieng, Madagascar, New Caledonia, Papua New Guinea, Vanuatu), and, more broadly, for his unconditional love of snails and slugs, biodiversity exploration, and alpha-taxonomy.
Diagnosis (Table 3). Externally, Laspionchis boucheti cannot be distinguished from L. bourkei. Internally, however, the insertion of the retractor muscle of the penis (at the posterior end of the visceral cavity) and the presence of additional, distal retractor muscle fibers can help distinguish L. boucheti from L. bourkei.
Color and morphology of live animals (Figs 9, 10). Live slugs are covered with mud and their dorsal color can hardly be seen. The background of the dorsal notum is brown, light to dark, homogenous or mottled with darker or lighter areas, and, occasionally, with red areas too. In some slugs, the tip of dorsal papillae (with and without dorsal eyes) can be yellow. The color of the foot is gray (light or dark), yellow, or orange. The hyponotum is light or dark grey, pale yellow or red, sometimes with a lighter whitish margin. The color of the foot and of the hyponotum of an individual can change rapidly, especially when disturbed. The ocular tentacles are brown (variable from light to dark) and short (a few millimeters).
Generally, the dorsal notum of any given slug can rapidly change from almost perfectly smooth to densely covered by many papillae. However, when slugs are not disturbed, the dorsum is usually covered by papillae of various sizes. In some slugs, larger papillae may be arranged in two longitudinal ridges on either side of the median line, but those ridges can appear and disappear rapidly. Some papillae bear dorsal eyes at their tip (most papillae bear three eyes). The number of papillae with dorsal eyes is variable (between 8 and 12, on average) and they mostly are on the central part of the notum. Their tip can be pale yellow, but not always. A central, much larger papilla, which also bears three dorsal eyes, is entirely retractable within the notum. In addition to the large papillae, the notum is covered by smaller, rounded papillae, which can make it look very granular.
External morphology (Fig. 11A, B). The body is not flattened. The notum is oval. Dorsal gills are absent. The large, central, retractable papilla at the center of the notum can usually only be seen in live animals. In preserved specimens, it is retracted inside the notum. The hyponotum is horizontal. The width of the hyponotum relative to the width of the pedal sole varies among individuals. The width of the hyponotum is approximately half of its total width. In the anterior region, the left and right ocular tentacles are superior to the mouth. Eyes are located at the tip of the ocular tentacles. Inferior to the ocular tentacles, superior to the mouth, the head bears a pair of oral lobes. The latter are smooth, with no transversal protuberance. The male aperture (opening of the copulatory complex) is below the right ocular tentacle (or very slightly to its left in dorsal view). The anus is posterior, medial, close to the edge of the pedal sole. On the right side (to the left in ventral view), a peripodial groove is present at the junction between the pedal sole and the hyponotum, running longitudinally from the buccal area to the posterior end, very close to the anus. The position of the female pore (at the posterior end of the peripodial groove) does not vary much among individuals. The pneumostome is medial. Its position on the hyponotum relative to the notum margin and the edge of the pedal sole varies among individuals but it tends to be closer to the notum margin.
Visceral cavity and pallial complex. The heart, enclosed in the pericardium, is on the right side of the visceral cavity, slightly posterior to the middle. From the anterior ventricle is an anterior vessel supporting several anterior organs such as the buccal mass, the nervous system, and the copulatory complex. The auricle is posterior. The kidney is more or less symmetrical, the right and left parts being equally developed. The kidney is intricately attached to the respiratory complex. The lung is in two left and right, more or less symmetrical, parts.
Digestive system (Figs 12, 13). There are no jaws. The left and right salivary glands, heavily branched, join the buccal mass dorsally, on either side of the esophagus. The radula is between two large postero-lateral muscular masses. Radulae measure up to 2 mm in length. Each radular row contains a rachidian tooth and two half rows of lateral teeth of similar size and shape. Examples of radular formulae are in Table 4. The rachidian teeth are unicuspid: the median cusp is always present; there are two inconspicuous lateral cusps (Fig. 13A). The length of the rachidian teeth (ca. 20 μm) tend to be approximately half the size of the lateral teeth (ca. 50 μm). The lateral aspect of the base of the rachidian teeth is straight, occasionally slightly convex. The half rows of lateral teeth form an angle of 45° with the rachidian axis. With the exception of the few innermost and outermost lateral teeth, the size and shape of the hook of the lateral teeth do not vary along the half row, nor do they vary among half rows. The hook of lateral teeth is extended posteriorly by a tail-like structure attaching to the radular membrane and making the hook look longer. The tail-like structure (posterior hook extension, Fig. 13D) is especially obvious in the outermost lateral teeth as its length gradually increases along each half row. The lateral teeth seem to be unicuspid with a flattened and curved hook with a rounded or pointed tip, but there also is a pointed spine on the outer lateral expansion of the base (basal lateral spine, Fig. 13A). In most cases, that spine cannot be observed because it is hidden below the hook of the next, outer lateral tooth. It can only be observed when the teeth are not too close (such as in the innermost and outermost regions) or when teeth are placed in an unusual position. The length of the spine decreases along the half row such that outermost teeth may be characterized by reduced or no lateral spine. The inner and outer lateral aspects of the hook of the lateral teeth are straight (i.e., not wavy and not with a protuberance).
The esophagus is narrow and straight, with thin internal folds. The esophagus enters the stomach anteriorly. Only a portion of the posterior aspect of the stomach can be seen in dorsal view because it is partly covered by the lobes of the digestive gland. The dorsal lobe is mainly on the right. The left, lateral lobe is mainly ventral. The posterior lobe covers the posterior aspect of the stomach. The stomach is a U-shaped sac divided into four chambers. The first chamber, which follows the esophagus, receives the ducts of the dorsal and lateral lobes of the digestive gland. The second chamber, posterior, receives the duct of the posterior lobe of the digestive gland. The third chamber is funnel-shaped and lined by ridges internally. The fourth chamber is continuous and externally similar to the third. The intestine is long, narrow, and the intestinal loops are exactly between types I and II, i.e., with a transitional loop on average oriented at 6 o'clock, though the orientation of the transitional loop ranges between 5 and 7 o'clock (Figs 1, 12). There is no rectal gland.
Nervous system (Fig. 11C). The circum-esophageal nerve ring is post-pharyngeal and pre-esophageal. The paired cerebral ganglia are close and the cerebral commissure is short (but its length does vary among individuals). Paired pleural and pedal ganglia are also all distinct. The visceral commissure is very short and the visceral ganglion is more or less median. Cerebro-pleural and pleuro-pedal connectives are short and pleural and cerebral ganglia touch each other on either side. Nerves from the cerebral ganglia innervate the buccal area and the ocular tentacles, and, on the right side, the penial complex. Nerves from the pedal ganglia innervate the foot. Nerves from the pleural ganglia innervate the lateral and dorsal regions of the mantle. Nerves from the visceral ganglia innervate the visceral organs. . A Posterior, hermaphroditic (female) reproductive system B posterior, hermaphroditic (female) reproductive system C anterior, male, copulatory apparatus. Scale bars: 3 mm (A) 1 mm (B) 2 mm (C). Abbreviations: ag accessory penial gland dd deferent duct dmf distal muscle fibers fgm female gland mass hg hermaphroditic gland ms muscular sac ov oviduct ps penial sheath rm retractor muscle rs receptaculum seminis sp spermatheca v vestibule.
Reproductive system (Fig. 14). Sexual maturity is correlated with animal length. Mature individuals have large female organs (with a large female gland mass) and fully-developed male parts. Immature individuals may have inconspicuous (or no) female organs and rudimentary anterior male parts. The hermaphroditic gland is a single mass, joining the spermoviduct through the hermaphroditic duct. There is a narrow receptaculum seminis (caecum) along the hermaphroditic duct. The female gland mass contains various glands (mucus and albumen) which can hardly be separated by dissection and of which the exact connections remain uncertain. The hermaphroditic duct becomes the spermoviduct. Proximally, the spermoviduct is not divided (at least externally) and is embedded within the female gland mass. Distally, the spermoviduct branches into the deferent duct and the oviduct. The free oviduct conveys the eggs up to the female opening and the exosperm from the female opening up to the fertilization chamber. The large, ovate-spherical spermatheca connects to the oviduct through a narrow and short duct. The oviduct is large (larger than the deferent duct) and straight. There is no vaginal gland.
Copulatory apparatus (Figs 14C, 15, 16). The male anterior organs consist of the penial complex (penis, penial sheath, deferent duct, retractor muscle) and the accessory penial gland (flagellum, muscular sac, hollow spine). The penial complex and the accessory penial gland share the same vestibule and the same anterior male opening.
The penial sheath is narrow and elongated. The penial sheath protects the penis for its entire length. The beginning of the retractor muscle marks the separation between the penial sheath (and the penis inside) and the deferent duct. The retractor muscle is strong, shorter than the penial sheath, and inserts at the posterior end of the visceral cavity. In addition, there is a cluster of retractor muscle fibers on the distal part of the penial sheath, near the vestibule. The deferent duct is highly convoluted with many loops. Inside the penial sheath, the penis is a narrow, thin, elongated, hollow tube, with numerous and densely-arranged (next to each other) hooks in its distal part. Penial hooks are pointed and measure from 50 to 100 μm. When the penis is retracted inside the penial sheath, the hooks are inside the tube-like penis; during copulation, the penis is everted like a glove and the hooks are then on the outside.
The accessory penial gland is a long, tube-like flagellum with a proximal dead end. The length of the flagellum of the penial gland varies among individuals but it is always heavily coiled. Near its distal part, the flagellum is enlarged into a muscular sac. Distally, the flagellum ends in a hard, hollow spine protected by a sheath which opens into the vestibule. The hollow spine is narrow, straight, elongated. Its base is conical. Its diameter is ca. 50 μm except at the base where it is larger (ca. 100 μm). The diameter of the opening at the tip measures ca. 30 μm. Its length ranges from 0.7 mm [1037] (BDMNH) to 1 mm [2693] (MTQ). There is no disc separating the spine of the penial gland and the vestibule.

Remarks.
A new species name is needed because no existing name applies to the species described here, based on the examination of all the type specimens available in the Onchidiidae, a careful study of all the original descriptions, and our ongoing taxonomic revision of every genus of the family (Dayrat et al. 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Several problematic species names, already discussed in detail in our revision of Paromoionchis (Dayrat et al. 2019: 68-72), are regarded as nomina dubia for a variety of reasons (the type locality is too vague, the original description is not informative enough, the type material is destroyed or lost). One of those nomina dubia, Onchidium palaense Semper, 1880 (type locality in Aibukit, Palau Islands) could belong to Paromoionchis or Laspionchis but its generic placement cannot be determined. Onchidium palaense does not belong to Onchidium because several traits mentioned by Semper, such as the absence of a rectal gland and of an accessory penial gland, are incompatible with Onchidium (Dayrat et al. 2016). Onchidium palaense simply is a nomen dubium which was arbitrarily placed in the genus Onchidium and cannot reliably be placed in any of the onchidiid genera.   Additional material examined. See below for each subspecies. Distribution (Fig. 7). Australia (Northern Territory) for L. bourkei bourkei. Indonesia (Ambon) for L. bourkei lateriensis. Indonesia (Sulawesi, Sumatra), Malaysia (Peninsular Malaysia), Singapore, Philippines (Bohol), and Vietnam for L. bourkei matangensis.
Habitat (Figs 17, 26). Laspionchis bourkei is found on mud, hard or soft, in open or dense mangrove forests. It can be locally common across its entire distribution.
Etymology. Laspionchis bourkei is dedicated to Adam Bourke, from Darwin, Northern Territory, Australia, a very knowledgeable mangrove expert and great naturalist, who generously accompanied us in the field around Darwin and showed us good collecting sites.
Diagnosis (Table 3). Externally, Laspionchis bourkei cannot be distinguished from L. boucheti. Internally, however, the long retractor muscle of the penis inserts at the posterior end of the visceral cavity in L. boucheti while the retractor muscle is short (and inserting in the first third of the visceral cavity) in L. bourkei bourkei and vestigial or absent in L. bourkei lateriensis and L. bourkei matangensis. Also, additional, distal, retractor muscle fibers are present in L. boucheti but absent in L. bourkei.
Color and morphology of live animals (Figs 18, 27). Live slugs are covered with mud and their dorsal color can hardly be seen. The background of the dorsal notum is brown, light to dark, homogenous or mottled with darker or lighter areas. The color of the foot is a mix of gray (light or dark) and yellow, as is the color of the hyponotum. The color of the ventral surface (foot and hyponotum) can change rapidly, especially when slugs are disturbed. The ocular tentacles are brown (variable from light to dark) and short (a few millimeters). The number of papillae with dorsal eyes is variable (between five and ten, on average) and they mostly are on the central part of the notum.
Digestive system (Figs 19,20,25). Examples of radular formulae are in Table 4. Radulae measure up to 2.9 mm in length (see below for each subspecies). The intestine is long, narrow, and the intestinal loops are exactly between types I and II, i.e., with a transitional loop on average oriented at 6 o'clock, acknowledging minor individual variation (Figs 1, 19).
Reproductive system (Fig. 21). There is a narrow receptaculum seminis (caecum) along the hermaphroditic duct. The large, ovate-spherical spermatheca connects to the oviduct through a narrow and short duct. The oviduct is straight, slightly larger than the deferent duct or of a similar diameter.
Copulatory apparatus 28,29). The length of the flagellum of the accessory penial gland varies among individuals but it is always heavily coiled. The hollow spine of the penial gland is narrow, straight, elongated. Its base is conical. Its length varies from 0.35 mm to 1 mm (see below for each subspecies). The penial sheath is narrow and short. The penial sheath protects the penis for its entire length. The beginning of the retractor muscle marks the separation between the penial sheath (and the penis inside) and the deferent duct. The retractor muscle is short (as long as the penial sheath) and inserting in the first third of the visceral cavity, vestigial (and free with no attachment), or absent (see below for each subspecies). There is no additional, distal, retractor muscle fibers. The deferent duct is highly convoluted. Inside the penial sheath, the penis is a narrow, thin, elongated, hollow tube, with numerous and densely-arranged (next to each other) hooks in its distal part. Penial hooks are pointed and measure from 15 to 45 μm (see below for each subspecies). When the penis is retracted inside the penial sheath, the hooks are inside the tube-like penis; during copulation, the penis is everted like a glove and the hooks are then on the outside.

Remarks.
A new species name is needed because no existing name applies to the species described here, based on the examination of all the type specimens available  in the Onchidiidae, a careful study of all the original descriptions, and our ongoing taxonomic revision of each genus of the family (Dayrat et al. 2016(Dayrat et al. , 2018(Dayrat et al. , 2019Goulding et al. 2018a, b, c). Laspionchis bourkei is divided in three distinct units of which the reciprocal monophyly is highly-supported in both mitochondrial and nuclear analyses (except for L. bourkei matangensis, unresolved using nuclear data). The fact that the three units within L. bourkei are distinct taxa means that they should be recognized and named. Even though we could have ranked them as species, we decided to rank them as sub-species for three main reasons.
(1) The three units within L. bourkei are cryptic externally and internally. Some minor anatomical differences seem to exist but which can hardly be used for identification (Table 3).
(2) Ranking the three units within L. bourkei as subspecies rather than species is more in agreement with the genetic distances observed between L. boucheti and L. bourkei. Indeed, the distance gap between L. boucheti and L. bourkei is between 2.5% and 7.5%, while the distance gap between the three L. bourkei units is between 2.5% and 3.9%, clearly suggesting that the three L. bourkei units are much less divergent (their COI sequences) than L. boucheti and L. bourkei, supporting their ranking as subspecies. Distance values should not necessarily be compared from one genus to another, but they can be compared between very closely related species.
(3) As of today, the three units within L. bourkei are allopatric which means that a doubt remains as to whether the three units are reproductively isolated or not. Overall, the three units within L. bourkei probably are relatively young taxa which diverged recently, explaining that they are cryptic internally and that their COI sequences are less divergent than the COI sequences between L. boucheti and L. bourkei.  Distribution (Fig. 7). Australia (Northern Territory). Habitat (Fig. 17A-D). Same as the entire species Laspionchis bourkei (see above). Etymology. See above, the species L. bourkei. Diagnosis (Table 3). Externally, the three subspecies of L. bourkei cannot be distinguished. Internally, L. bourkei bourkei differs from both L. bourkei lateriensis and L. bourkei matangensis. Indeed, L. bourkei bourkei is characterized by a short retractor muscle of the penis which inserts in the anterior third of the visceral cavity while the retractor muscle is vestigial or absent in L. bourkei lateriensis and L. bourkei matangensis. Also, the spine of the accessory penial gland is on average slightly longer in L. bourkei bourkei than in L. bourkei lateriensis and L. bourkei matangensis. Color and morphology of live animals ( Fig. 18A-D, G, H). Identical to the species L. bourkei (see above).

Laspionchis bourkei bourkei
Digestive system . Identical to the species L. bourkei (see above). Examples of radular formulae are in Table 4. Radulae measure up to 2.9 mm in length.
Reproductive system (Fig. 21A). Identical to the species L. bourkei (see above). Copulatory apparatus (Figs 22A,23E,F,F). Similar to the species L. bourkei (see above) acknowledging some minor variations: the length of the spine of the accessory penial gland ranges from 0.75 mm [1657 H] (NTM P.57615) to 1 mm [1666] (NTM P.57618), the retractor muscle is short (as long as the penial sheath) and inserts in the first third of the visceral cavity, and penial hooks measure from 20 to 35 μm.
Remarks. See above, the remarks on the species Laspionchis bourkei.  Distribution (Fig. 7). Indonesia (Ambon). Habitat (Fig. 17E). Same as the entire species Laspionchis bourkei (see above). Etymology. The subspecies Laspionchis bourkei lateriensis is named after Lateri, in Ambon because the type locality is part of the preserved mangrove of Lateri. The name lateriensis is an adjective derived from Lateri and the suffix -ensis.
Diagnosis (Table 3). Externally, the three subspecies of L. bourkei cannot be distinguished. Internally, L. bourkei lateriensis differs from L. bourkei bourkei but cannot be distinguished from L. bourkei matangensis. All three subspecies, however, are clearly delineated using molecular DNA sequences.

Laspionchis bourkei matangensis
Habitat (Fig. 26). Same as the entire species Laspionchis bourkei (see above). Etymology. The subspecies L. bourkei matangensis is named after Matang, in Peninsular Malaysia. The type locality is part of the Matang mangrove forest. The name matangensis is an adjective derived from Matang and the suffix -ensis.
Diagnosis (Table 3). Externally, the three subspecies of L. bourkei cannot be distinguished. Internally, L. bourkei matangensis differs from L. bourkei bourkei but cannot be distinguished from L. bourkei lateriensis. All three subspecies, however, are clearly delineated using molecular DNA sequences.
Color and morphology of live animals (Fig. 27). Identical to the species L. bourkei (see above).
Remarks. See above the remarks on the species Laspionchis bourkei.

Discussion
A few preliminary remarks can be made here regarding the types of intestinal loops, even though a more detailed discussion will be provided after our revisions of Peronia and Platevindex are published (in preparation).
(1) Nearly all onchidiid species are characterized by only one intestinal type. Some intraspecific variation exists, which can be evaluated based on the orientation of the transitional loop. However, the presence of more than one intestinal type in an onchidiid species remains exceptional, such as in Alionchis jailoloensis (see Goulding et al. 2018b).
(2) Nearly all onchidiid genera are characterized by only one or two intestinal types.
For instance, Wallaconchis and Marmaronchis are characterized by intestinal loops of type I; Onchidina, Paromoionchis, and Peronina by type II; Laspionchis by loops between types I and II; Alionchis, Melayonchis, and Onchidium by types II and III.
Platevindex is the only genus characterized by intestinal loops of more than two types (I, II, and III). (3) Intestinal loops are quite useful to identify genera. For instance, all known species of Wallaconchis are characterized by intestinal loops of type I. Therefore, a slug with intestinal loops of type II cannot belong to Wallaconchis, unless intestinal loops of type II are found in the future in a new species of Wallaconchis. Also, Laspionchis slugs are the only ones known so far with an intestinal type between types I and II, with a transitional loop oriented at 6 o'clock (acknowledging individual variation). Therefore, slugs with intestinal loops between types I and II likely belong to Laspionchis. (4) There must be some reasons explaining why intestinal types are not randomly distributed across onchidiid species and genera; however, the exact reasons are still unclear at this stage. Evolutionary history is possibly involved. For instance, the fact that all Wallaconchis species are characterized by intestinal loops of type I may be due to the presence of a type I in their common ancestor. Adaptation to different habitats is likely involved as well and will be discussed after our revisions of Peronia and Platevindex are published (in preparation).
Onchidiids are notoriously difficult to identify, both at the genus and species levels. Laspionchis slugs are no exception. They are most readily identified at the genus level using DNA sequences. Externally, they are practically impossible to distinguish from Paromoionchis slugs which live in the same habitat (mangrove mud surface) and are often found at exactly the same sites (see Dayrat et al. 2019). The male opening is clearly to the left of the right ocular tentacle in Paromoionchis, while it is just below the right ocular tentacle or only slightly to its left in Laspionchis, but this character is nearly impossible to check in the field when slugs are alive (because they retract as soon as they are being touched). Animal size can help distinguish Laspionchis slugs from Paromoionchis slugs in the field. Indeed, the maximum size of Paromoionchis slugs -55 mm in P. tumidus, 65 mm in P. daemelii, 47 mm in P. boholensis Dayrat & Goulding in Dayrat et al. 2019, 48 mm in P. penangensis Dayrat & Goulding in Dayrat et al. 2019, and 35 mm in P. goslineri Dayrat & Goulding in Dayrat et al. 2019 -is much higher than the maximum size of Laspionchis slugs, 31 mm in L. boucheti and 32 mm in L. bourkei. That being said, animal length needs to be used with caution and it obviously is useless for all individuals less than 30 mm long.
Laspionchis is characterized by a unique combination of external and internal traits: no dorsal gills, male opening below the right eye tentacle (or slightly to its left), no rectal gland, intestinal loops between types I and II, accessory penial gland present with a muscular sac, penis with numerous, pointed hooks densely arranged next to each other. This unique combination of characters of Laspionchis is close to that of Paromoionchis (Dayrat et al. 2019: 19). However, there are three important differences: in Paromoionchis, the male opening is clearly to the left of the right ocular tentacle, the intestinal loops are clearly of type II, and penial hooks (which are present in P. tumidus but are absent in the four other Paromoionchis species) are sparse (i.e., not densely arranged, next to each other) and not pointed (Dayrat et al. 2019 : figs 21, 22).
The two known species of Laspionchis are cryptic externally but distinct internally. They are found in exactly the same habitats and cannot be distinguished in the field. However, they can be identified successfully with both DNA sequences and internal anatomy. Species externally cryptic but internally distinct have also been observed in Paromoionchis, Peronina, and Wallaconchis (Dayrat et al. 2019;Goulding et al. 2018a, c). In Marmaronchis, species are cryptic externally and internally (Dayrat et al. 2018).
In Onchidium and Melayonchis, species are distinct both externally and internally (Dayrat et al. 2016. Finally, Alionchis and Onchidina are monotypic, at least as of today Dayrat et al. 2018).
The two new species described here are widespread and can be locally common. That they are discovered only now is not so surprising. Laspionchis is restricted to mangroves and mangroves of South-East Asia have been poorly explored and the biodiversity they host remains poorly known. Also, onchidiid taxonomy has been confused for a long time (Dayrat 2009). Now that onchidiid systematics is finally being revised, new taxa are being discovered (e.g., Goulding et al. 2018a, b, c), contributing to a better knowledge of the diversity of mangrove invertebrates in South-East Asia. It is very possible that additional species of Laspionchis will be discovered in the future, within or outside its current distribution. However, the study of the biodiversity of Laspionchis remains challenging, mostly because Laspionchis slugs are very hard to recognize in the field (they are not distinguishable from Paromoionchis slugs) and because Laspionchis species are externally cryptic (which means that many individuals looking similar need to be collected and individually numbered).
Hwai Tan (Malaysia), Marivene Manuel (Philippines), Munawar Khalil (Indonesia), and Quảng Ngô Xuân (Vietnam). Collecting in Brunei, New South Wales, Queensland, and the Northern Territory was done with permits from local institutions. A research permit was issued to Benoît Dayrat in Singapore (#NP/RP10-020). We thank the Ministry of Research, Technology and Higher Education, Republic of Indonesia (Ristek-Dikti) that issued a research permit to Benoît Dayrat (Ristek #134/SIP/FRP/ E5/Dit.KI/VI/2017). We also wish to thank the Universitas Malikussaleh for being our home base institution in Indonesia. This work was supported by the Eberly College of Science at the Pennsylvania State University and by a REVSYS (Revisionary Syntheses in Systematics) award from the US National Science Foundation (DEB 1419394).