Molecular phylogeny and systematics of Australian ‘Iravadiidae’ (Caenogastropoda: Truncatelloidea)

The family Iravadiidae is found to be polyphyletic in a molecular phylogenetic analysis using a subset of Australian taxa. Taxa previously assigned to Iravadia form a monophyletic clade, but Nozeba topaziaca clusters with Auricorona queenslandica n. gen. and n. sp. in an unnamed family related to Tornidae. Aenigmula criscionei n. gen. and n. sp., an iravadiid-like species from the Northern Territory, belongs to another unnamed family related to Caecidae, Calopiidae and Clenchiellidae. A systematic revision of some Australian ‘iravadiids’ raises the subgenera Fluviocingula and Pseudomerelina to full generic rank and reinstates two former synonyms of Iravadia (Fairbankia), Pellamora and Wakauraia, as genera. The species formerly identified in Australia as Iravadia quadrasi is recognised as three allopatric species; Iravadia pilbara n. sp. and the reinstated species Iravadia goliath and Iravadia quadrina. Pellamora splendida n. sp., from Western Australia, is recognised as distinct from Pellamora australis, and Fluviocingula superficialis n. sp. from Fluviocingula resima. Wakauraia fukudai n. sp. is recorded from central Queensland. http://zoobank.org/urn:lsid:zoobank.org:pub:1B9917F6-48B2-4597-85C1-F90BA9093475


Introduction
The Iravadiidae are a diverse group of microgastropods from the tropical to temperate Indo-West Pacific region. They occur in environments ranging from the deep ocean to shallow marine and estuarine waters, but are frequently associated with mangrove forests and rocky intertidal to shallow subtidal habitats. Ponder (1984) produced a morphology-based generic revision of Iravadiidae that recognised eight Recent genera (plus five subgenera) and one fossil genus, and subsequently Ponder (1994) described an additional monotypic genus from Hong Kong. The prevailing systematic classification of Iravadiidae divides the family into two groups: Iravadia Blanford, 1867, defined by several morphological characters and constituting the majority of the species-level diversity, and a heterogeneous assemblage of the other iravadiid genera (Ponder 1984;Ponder and de Keyzer 1998). Starobogatov et al. (1989) subsequently removed all genera except Iravadia to either Hyalidae or an unnamed family group, although this classification is not generally used.
Most iravadiid species and several genera are known only from their shells, but others have been described in some anatomical detail (Ponder 1984;Fukuda 2000). The only apparent morphological feature that characterises the entire family is a flattened, unsculptured protoconch. * Email: rosemary.e.golding@gmail.com Supplementary data available online at www.tandfonline.com/10.1080/13235818.2014.888971 Ponder (1984) also noted that most (but not all) iravadiids lack a pallial or metapodial tentacle, but this situation is not unique in Truncatelloidea (sensu Criscione and Ponder 2012) so does not confirm a common origin for the group. Species in Iravadiidae are generally not well characterised and their taxonomy is sometimes inconsistently applied in the literature. There are several instances where iravadiids are recorded as undescribed or unidentifiable using the present resources (e.g. Robba et al. 2003; Japanese Association of Benthology 2012).
The paucity of available information for many iravadiids led Ponder (1984) to remark that his proposed classification was tentative and open to refinement with the benefit of additional evidence. Since that publication, several new iravadiid species have been recognised (Starobogatov et al. 1989;Ponder 1994;Fukuda 2000), but no further attempts have been made to address the higher classification of Iravadiidae. Criscione and Ponder (2012) generated the first molecular phylogenetic analysis to include multiple iravadiids, as part of a broader analysis of Rissooidea and Truncatelloidea. They included four iravadiids representing Nozeba Iredale, 1915 and three subgenera of Iravadia. Significantly, Iravadiidae sensu lato (s. l.) was found to be paraphyletic because Nozeba was separated from the other species. This preliminary evidence high-lighted the need for a more inclusive molecular treatment of Iravadiidae.
There are currently 13 recognised iravadiid species occurring in Australia, representing all five subgenera of Iravadia and at least three other genera: Rissopsis Garrett, 1873, Chevallieria Cossmann, 1888, Nozeba and Liroceratia Ponder, 1984. Several of these species, such as Iravadia quadrasi (Böttger, 1893), have a broad Indo-West Pacific distribution that includes northern Australia but others are endemic to a specific region, such as Chevallieria australis Ponder, 1984, which occurs only in southern Australia. Some Australian iravadiids are abundant and widely distributed, particularly in mangrove forests throughout the tropical and subtropical regions.
The primary aim of this study was to conduct a preliminary phylogenetic analysis of available material of Iravadiidae using molecular data, and to reassess the present generic classification using this new line of evidence. A secondary objective was to test the apparent paraphyly of Iravadiidae (Criscione and Ponder 2012). Because iravadiids occur in many specific locations across the Indo-West Pacific and are often rare or have never been collected alive, many taxa could not be included in this study so it does not completely reflect the diversity in the family. Of necessity, this study therefore focuses on resolving the relationships of those Australian iravadiids for which material was available, the systematic redescription of named taxa and the description of new taxa.

Materials and methods
The majority of the material examined during this study was obtained from the collections of the Australian Museum or collected during field work in Queensland (September 2011) and northern Western Australia (June 2012) as part of a larger study of Australian mangrove microgastropods (see also Golding 2014). Material collected alive during fieldwork was photographed crawling in a Petri dish of salt water, using a microscope-mounted camera. Pigmentation and foot/tentacle morphology were observed and are included in descriptions (where available). Specimens were preserved for either molecular (95%-100% ethanol) or morphological (10% saltwater formalin or 70% ethanol) examination.
Penis morphology was observed and photographed (where possible) on living animals and examined on preserved specimens using a stereomicroscope with camera lucida. Radulae were removed from dissected buccal masses by dissolution overnight in a warmed solution of sodium hydroxide. Shells, radulae and opercula were cleaned in a sonic water bath, mounted on stubs and thinly coated with gold for examination by scanning electron microscopy (performed by Sue Lindsay, Microscopy and Microanalysis Laboratory, Australian Museum) using a Zeiss Evo LS-15 machine. Shell measurements were made using a calibrated camera lucida, with whorl counts rounded to the nearest quarter whorl.
For molecular sampling, either entire animals or samples of foot tissue were processed using a DNeasy kit (Qiagen, Inc., Hilden, Germany) and QiaCube ® robot to extract genomic DNA. Gene fragments from cytochrome oxidase subunit I (COI) and 16S genes from the mitochondrial genome and 28S from the nuclear genome were amplified using polymerase chain reactions (PCR). The PCR were performed in 25-μl volumes containing 1 × PCR buffer, 200 mm each dNTP, 2.0 mm MgCl 2 , 0.5 mm forward and reverse primers, 1.25 units Taq polymerase, and approximately 50 ng DNA. Amplification followed a standard protocol with 35 cycles of 94 • C for 1 minute, and primer-specific annealing conditions (see Golding 2014 for cycle conditions). Post-PCR products were purified using ExoSAP-IT proteinase solution (GE Healthcare, Pittsburgh, PA, USA) and sequenced in both directions by Macrogen Inc. (Seoul, Korea).
Sequences were compared to their electropherogram to correct misreads and compiled into contigs using BioEdit v7.0.9 (Hall 1999). All sequences were deposited in GenBank and their accession numbers are provided in Table 2 (KC439750-KC439959). Ribosomal sequences were aligned using the online MAFFT v.6 server (Katoh et al. 2002) with the E-INS-i option (Katoh et al. 2005) implemented for the 16S data set and Q-INS-i (Katoh and Toh 2008) for the 28S data set. Sequences of COI translated using the invertebrate mitochondrial code were unambiguously aligned by amino acid sequence. Data sets were compiled using Mesquite v2.75 (Maddison and Maddison 2011). Sequence divergence was estimated by uncorrected pairwise distances computed within and between taxa using MEGA5 (Tamura et al. 2011).
A subset of the Australian iravadiid fauna was sampled, with a focus on mangrove-affiliated species. Truly marine and subtidal taxa were not included due to the lack of suitable material. Hence, the genera known to occur in Australia, Nozeba and all subgenera of Iravadia except Iravadia (Pseudonoba), were represented by one or more taxa in the molecular analysis, but Chevallieria, Rissopsis and Liroceratia were not sampled (Table 1) (Ponder 1984). Several new and unassigned taxa were collected and have been included in the molecular analysis (systematic descriptions are also provided). 16S and 28S sequence data from non-iravadiid taxa were sourced from Criscione and Ponder (2012) and additional COI sequences were produced during this study (see Golding 2014: Table 2). Details for the source of non-iravadiid sequences are provided in Table 1.
Two molecular data sets were constructed to test the monophyly of Iravadiidae and its relationship to other families in Truncatelloidea and also to test the internal relationships and molecular diversity of Australian Iravadiidae.  (Ronquist and Huelsenbeck 2003). Bayesian posterior probability support was estimated by running four Markov chains (10 million generations each, with trees sampled each thousand generations). The first 25% of trees were conservatively rejected as burn-in, and stationarity was confirmed by examination of the log likelihood plot using Tracer (Rambaut and Drummond 2007). A summary consensus tree with support indices was generated by MrBayes. Data sets were partitioned by gene, with the GTR + G + I model of sequence evolution selected for both data sets by MEGA5. Trees were visualised using FigTree v1.3.1 and rooted using the outgroup. Support for individual clades was considered high for Bayesian probabilities > 95% and bootstraps > 80%, and moderate for Bayesian probabilities 90%-95% and bootstraps 70%-80%. Nodes with lower support values were not considered significant. Systematic descriptions have been provided for the Australian taxa included in the molecular analysis. Many other taxa assigned to Iravadiidae are represented in the dry collections of the Australian Museum. However, since those taxa were addressed by Ponder (1984) and no new information has been gathered, they are not redescribed here. Unless otherwise stated, the location of type material and materials examined is Australia. Shell measurements were made using a calibrated camera lucida attached to a dissecting microscope or directly from scanning electron microscope images of entire shells. Materials examined have been summarised here, but full details are provided in a supplementary file.

Abbreviations
States of Australia: NSW-New South Wales; NT-Northern Territory; QLD-Queensland; VIC-Victoria; WA-Western Australia.
Shell dimensions: SL-shell length; SD-diameter of last whorl; AL-aperture length; AW-aperture width; PWC-protoconch whorl count; SpP-number of spiral rows of pits on last whorl; SpR-number of spiral ribs on last whorl; SpRN-number of nodules on strongest spiral rib on last whorl (excluding varix); TWC-teleoconch whorl count.

Sequence divergence
Mean uncorrected pairwise distances (p-distances) of COI between individuals of each species examined in this study (i.e., within-species distances) ranged from 0.10% (n = 3, Iravadia pilbara n. sp.) to 2.42% (n = 7, Pseudomerelina cf. mahimensis) ( Table 2). COI sequence divergence between pairs of species from different genera (as recognised in the Systematics section, below) was > 17.52% in every instance. The p-distance between the species pair in Pellamora was 12.46% and in Fluviocingula it was 15.62%. The three species comprising the taxon formerly recognised in Australia as Iravadia quadrasi (Iravadia goliath, Iravadia quadrina and I. pilbara n. sp.) were separated by p-distances of between 8.51% and 18.11%. Sequence divergence from the most closely related species based on molecular phylogenies is given in the Remarks for each species in the Systematics section below.

Molecular phylogenies
The consensus trees produced by BI and ML analysis were very similar, and the few differences that were encountered are outlined in the following section. Only the BI summary trees are shown here, labelled with Bayesian posterior probabilities and bootstrap support values generated by ML analysis (Figs 1, 2).
The topology of the trees from BI and ML analyses of relationships within Truncatelloidea (including single representatives of all species in this study that were provisionally assigned to Iravadiidae) differed slightly (Fig. 1). The sister clade to Fluviocingula (lacking significant support) was Pellamora in the BI analysis, but Pseudomerelina in the ML analysis. Both analyses found strong evidence of paraphyletic structure in Iravadiidae. Three distinct clades were resolved: (1) 'Clade A', comprising Nozeba topaziaca and another new species (Auricorona queenslandica n. gen and n. sp., described below); (2) the unclassified genus Aenigmula n. gen., with a single new species (Aenigmula criscionei n. gen. and n. sp., described below); (3) Iravadiidae s. str., containing the remaining iravadiid taxa. Each of these three groups was supported by very high support indices from both ML and BI analyses. 'Clade A' was found to be sister to Tornidae and Aenigmula was basal to a clade containing Clenchiellidae and Calopiidae, and this trio of families was sister to Caecidae. The clade comprising Iravadiidae, 'Clade A', Aenigmula, Tornidae, Caecidae, Clenchiellidae and Calopiidae received high support indices and was also characterised by the deletion of a single codon from the COI gene (Fig. 2). This deletion was found in all taxa in this group but was not observed in any other truncatelloid sequences. No clear sister taxon to this group was identified.
The phylogenetic analysis of a restricted subset of species, including multiple representatives of most 'iravadiids' and just two outgroup taxa, produced essentially the same topology (Fig. 2). Two differences were found between ML and BI analyses of this data set. (1) Three species formerly assigned to Iravadia quadrasi (I. quadrina, I. goliath and I. pilbara n. sp.) were sister to Wakauraia fukudai n. sp. and Iravadia cf. capitata in the ML analysis (lacking significant support), but were more closely related to Fluviocingula, Pellamora and Pseudomerelina in the BI analysis (also lacking significant support). (2) The sister clade to Fluviocingula (lacking significant support) was Pellamora in the BI analysis, but Pseudomerelina in the ML analysis. Each multi-taxon genus recognised in the Systematic section Notes: Values in bold are mean p-distance within each species. Taxa described or revised in this manuscript are named accordingly. Figure 1. Summary tree from Bayesian analysis of concatenated cytochrome oxidase subunit I (COI; excluding third codon position bases), 16S and 28S sequences (10 million generations, trees sampled every 1000 generations). Support indices are Bayesian inference posterior probability (above nodes, > 90%) and maximum likelihood bootstraps (below nodes, > 70%); asterisks indicate a support value of 100%. Family and higher level names are in bold. Taxa described in this manuscript are named accordingly. Black bar indicates codon deletion from COI gene.
below was highly supported and monophyletic, except Iravadia in which I . cf. capitata was not sister to the other three species assigned to Iravadia. Species represented by more than one individual (all except I . cf. capitata) were reciprocally monophyletic with strong support by either bootstrap or Bayesian posterior probability indices or both. Within Iravadiidae sensu stricto (s. str.), Wakauraia + I . cf. capitata formed the most basal clade. The three species formerly recognised in Australia as I. quadrasi clustered together, although the relationships between this species group and other iravadiids were poorly supported and inconsistent between analysis methodologies. Pseudomerelina, Fluviocingula and Pellamora formed the crown group of Iravadiidae s. str., but the relationships between these genera remain uncertain. Pseudomerelininae Starobogatov, 1989in Starobogatov et al. (1989: 36. Type genus (original designation): Pseudomerelina Ponder, 1984. Brandt (1968), as First Reviser, gave Iravadiinae precedence over Fairbankiinae. Ponder (1984) supported this decision, and synonymised the invalid name Hyalidae (a homonym of the amphipod family Hyalidae Bulycheva, 1957) with Iravadiidae. Starobogatov et al. (1989) reversed that decision, placing Hyala H. & A. Adams, 1852 and Nozeba in Hyalidae, and removing Rissopsis, Ceratia, Liroceratia and Acliceratia Ponder, 1984 to a separate unknown family. Hyala is characterised by a suite of characters (including a paucispiral operculum) that are more typical of 'Clade A' or Aenigmula, rather than Iravadiidae s. str., so Hyalidae is not here considered Figure 2. Summary tree from Bayesian analysis of concatenated cytochrome oxidase subunit I (COI; including third codon position bases), 16S and 28S sequences (10 million generations, trees sampled every 1000 generations). Support indices are Bayesian inference posterior probability (above nodes, > 90%) and maximum likelihood bootstraps (below nodes, > 70%); asterisks indicate a support value of 100%. Taxa described or revised in this manuscript are named accordingly, with locality data given for each sample. Family names are in bold and family-level groups examined here are shaded. Shell images are all to the same scale.

Remarks
to be a synonym of Iravadiidae. Pseudomerelininae was recognised as a synonym of Iravadiidae by Bouchet and Rocroi (2005). Ponder (1984) recognised eight Recent genera (Iravadia, Rissopsis, Chevallieria, Acliceratia, Hyala, Ceratia, Liroceratia and Nozeba) and one fossil genus (Rhombostoma Seguenza, 1876) in Iravadiidae. The genus Lantauia Ponder, 1994 was additionally described. The present study examines species of only two of these genera (Iravadia and Nozeba) and finds them to belong to different branches of the truncatelloid tree, preventing their unification in a single family. Nozeba has been removed from Iravadiidae (see 'Clade A', below). With the available material the affiliation of the unsampled genera to Iravadia cannot be tested using molecular data at this time. The inclusion of Lantauia in Iravadiidae is supported by opercular characters but the others must be regarded as tentative at best as their operculum is either paucispiral or unknown. On the basis of the information on these genera presented by Ponder (1984), they show more similarity to 'Clade A' or Aenigmula n. gen. than to Iravadia s. l., and therefore probably do not belong in Iravadiidae s. str. A molecular study including taxa from these ambiguous groups is sorely needed to resolve the true composition of Iravadiidae.
The taxon sampling in the molecular analysis and systematic sections of this study is too limited to provide a comprehensive diagnosis of Iravadiidae s. str. Based on the available information, it seems likely that Iravadiidae should be restricted to the group of taxa possessing an operculum with a nucleus on the umbilical margin (i.e., Iravadia s. l.) and an initially planar protoconch, excluding those with a paucispiral operculum and domed or tall protoconch. This separation has been validated for Nozeba, Auricorona n. gen. and Aenigmula n. gen. (see below), but not for Lantauia, Rissopsis, Chevallieria, Acliceratia, Hyala, Ceratia or Liroceratia.

Iravadia Blanford 1867
Iravadia Blanford 1867: 56-58. Type species (original designation): Iravadia ornata Blanford, 1867; Recent, India. Ponder (1984) recognised five subgenera within Iravadia s.l.; Iravadia, Pseudomerelina, Fluviocingula, Fairbankia and Pseudonoba. No species confirmed as belonging to Iravadia (Fairbankia) or Iravadia (Pseudonoba) was included in this study, so no definitive conclusions can be drawn on the status of those 'subgenera'. However, the concept of Iravadia has been substantially modified by the recognition in this study of Fluviocingula and Pseudomerelina as full genera, as well as the reinstatement of Pellamora Iredale, 1943 andWakauraia Kuroda &Habe, 1954 for some taxa previously assigned to Iravadia (Fairbankia). It therefore seems likely that Pseudonoba and Fairbankia will also prove to be distinct genera.

Remarks
Only four taxa in this study were retained in Iravadia s. str., three forming a group of species formerly identified in Australia as Iravadia quadrasi and one provisionally classified as I . cf. capitata. Other taxa currently assigned to Iravadia were not examined, including the type of the genus, I. ornata. Molecular phylogenetic analysis of Iravadiidae shows that the four sampled species of Iravadia do not form a monophyletic group, due to the nested position of Wakauraia fukudai n. sp. within the clade (sister to I . cf. capitata). On the basis of branch lengths and shell shape, it is likely that I . cf. capitata belongs to a distinct, unnamed genus (including similar taxa such as Iravadia carpentariensis, and perhaps some species currently assigned to Pseudonoba). At the present time, it is practical to retain Iravadia as a non-monophyletic group until further progress is made. No generic description is provided here for Iravadia, because it appears to be paraphyletic and because so few of its constituent taxa, including the type species, have been studied in detail. Other material. Twenty-three wet lots and 34 dry lots from Queensland between Moreton Bay and Cooktown. See supplementary data for full list of material examined.
Operculum (Fig. 4B). Oval; growth striae concentric; nucleus positioned above midpoint on rounded umbilical margin. Exterior surface with coarse growth striae; interior surface with groove and thickened ridge along umbilical margin, and two low radial ridges emerging from nucleus.
Radula (n = 3) (Figs 4B, 5A). Central tooth 6 − 9 + 1 + 6 − 9/2 + 2; central cusp large, secondary cusps diminishing outwardly; paired basal denticles small, positioned just below cutting edge; lateral margins straight. Lateral teeth 3 + 1 + 4 − 5. Marginal teeth with subequal cusps; inner marginal teeth with ∼20 cusps; outer marginal teeth with ∼15 cusps. (Fig. 4C). Head-foot cream or white; grey pigment on sides of neck and foot and dorsal surface of head, cream to yellow speckles between and behind eyes; snout either solid black (except for lips) or black on ventral and lateral surfaces with yellow and black patches on dorsal surface. Dorsal surface of propodium with diffuse, grey lateral stripes; posterior margin of foot pointed, lacking metapodial tentacle. Cephalic tentacles white to yellow with broad black bands at midpoint and 3 / 4 length. Pallial tentacles absent.

External morphology and colouration in life
Penis (Fig. 6A). Cylindrical with broad tip; duct opening on inner edge below inflated, transverse, glandular bulge on tip. Speckled white, with black flecks on inner margin.

Distribution (Fig. 7A)
Common in mangrove forests on the east coast of Australia from Brisbane northward at least as far as Cooktown, but not occurring west of the tip of Cape York.
Notes: Only mature specimens with a varix were measured; note that many specimens were missing the protoconch, in which case shell lengths are underestimates and no protoconch whorl counts are provided. Ht, holotype; Pt, paratype.
Australian species. As there are no known morphological differences between I. quadrasi from the type locality in the Philippines and I. goliath from northeastern Australia, it is possible that they belong to the same taxon. The taxonomic divisions in this group are provisional pending further molecular testing to support the genetic separation of South East Asian I. quadrasi from the Australian taxa. Fortunately, previous authors have provided several names for Australian species in this group, two of which are reinstated here: I. goliath and I. quadrina (see Remarks for I. quadrina, below). Although Laseron (1956) believed that the distribution of 'Merelina' goliath extended from the type locality at Hervey Bay (Queensland) to Darwin (Northern Territory), the molecular evidence suggests that the species is in fact restricted to the coastline east of the tip of Cape York. Ponder (1984) synonymised Merelina reversa Laseron, 1956 (type locality Bowen, Queensland) with I. quadrasi. However, the holotype of M. reversa (SL = 2 mm) is much smaller than adult specimens of I. goliath and may not be the same taxon.
There are no obvious shell characters to differentiate I. goliath from the closely related species I. quadrina and I. pilbara n. sp. Although the shells of the holotypes are clearly distinguishable (Fig. 3), examination of specimens from several populations shows considerable variation within each species. No reliable morphological characters have been identified, but I. goliath is generally intermediate in size and whorl count between the largest species I. quadrina and the smallest I. pilbara n. sp (Table 3). Iravadia goliath usually has fewer axial ribs on the last whorl (14-18) than I. quadrina (16-20), although there is overlap between specimens from each species. The central radular teeth of I. goliath have a greater number of cusps (6-9) than I. quadrina and I. pilbara n. sp., and two pairs of basal denticles rather than one pair. Sequence divergence of the COI gene is 18.11% between I. goliath and I. quadrina, and 17.08% between I. goliath and I. pilbara n. sp. (Table 2). Where molecular or radular data are not available, I. goliath, I. quadrina and I. pilbara n. sp may be able to be reliably identified based on geographic location, as, so far as is known, their distributions do not overlap.
Other material. Four wet lots and two dry lots from Queensland between Weipa and the Wellesley Group of Islands in the Gulf of Carpentaria; eight wet lots and five dry lots from Northern Territory between Gove Peninsula and the Tiwi Islands; two wet lots from Broome, Western Australia. See supplementary data for full list of material examined.
External morphology, colouration in life (Fig. 4G) and penis (Fig. 4H). As for I. goliath. Distribution (Fig. 7A). Known from mangrove habitat between Weipa on the Gulf of Carpentaria coast of Queensland and Broome, in Western Australia. The species is not known west of Port Hedland or east of Cape York Peninsula. Laseron (1956) introduced Planapexia quadrina for a shell from Darwin, Northern Territory. Ponder (1984) placed 'P.'quadrina in Iravadia in synonymy with I. quadrasi. On the basis of molecular and phylogeographic evidence, I. quadrina is here reinstated with its type locality in Darwin, Northern Territory. Iravadia quadrina is slightly larger than either I. goliath or I. pilbara n. sp., based on limited shell measurements (Table 3). The radula has only a single pair of basal denticles, unlike the double pair present in I. goliath. Geographic information is useful in aiding the identification of I. quadrina, as morphological characters are mostly uninformative in separating species in this group. Molecular sequence divergence in the COI gene is 18.11% between I. quadrina and I. goliath, and 8.51% between I. quadrina and I. pilbara n. sp. (Table 2).
Other material. Three wet lots and one dry lot from Western Australia between the Port Hedland and Karratha. See supplementary data for full list of material examined.

Distribution (Fig. 7A)
Known only from two locations in the Pilbara, Western Australia; Port Hedland and the Dampier region (including Karratha).

Remarks
None of the several names introduced by earlier authors for species subsequently synonymised with I. quadrasi are available for this new Western Australian species. Molecular phylogenetic analysis shows that I. pilbara n. sp. is most closely related to I. quadrina, with sequence divergence in the COI gene of 8.51%, versus 17.08% between I. pilbara and I. goliath. The distributions of I. pilbara n. sp. and I. quadrina are separated by a relatively short distance (600 km) between Broome and Port Hedland, Western Australia. Species identification in this group is difficult, and should be informed by geographic information. Based on the small number of mature specimens available for shell measurement, it appears that I. pilbara n. sp. has fewer whorls and is smaller than other closely related species previously referred to I. quadrasi. It also seems to have fewer axial ribs on the last whorl than I. quadrina, although this is variable between individuals.

Etymology
Named after the Pilbara region of northwestern Australia, where the species occurs.
Operculum (Fig. 8F). Oval to D-shaped; growth striae concentric; nucleus positioned at midpoint on obtusely angled umbilical margin. Exterior surface with fine growth striae; interior surface with groove and thickened ridge along umbilical margin, and two low radial ridges emerging from nucleus. Figure 8. Iravadia cf. capitata (Laseron, 1956). A, B, Representative specimen; C, Pellamora capitata Laseron, 1956   External morphology and colouration in life (Fig. 8H). Head-foot transparent white with opaque white speckles on sides of neck and behind eyes. Posterior foot margin simple, lacking metapodial tentacle. Cephalic tentacles with scattered white speckles but otherwise unpigmented. Pallial tentacle absent.
Penis. Not observed.

Distribution
Based on specimens in the collections of the Australian Museum, the distribution of I. capitata extends along the eastern coast of Australia from southern to northern Queensland, terminating in the region of Cape York. However, as the identity of the examined material is unclear due to confusion between several Australian species of Iravadia (see below), this distribution is speculative.

Remarks
The specimens used in the molecular analysis and in this redescription have been provisionally referred to I. capitata, but they are not a perfect match to the holotype of that species. The holotypes of I. capitata and its synonym I. spiralis are both small but mature shells with three teleoconch whorls. In contrast, the shells of lot AMS C.470914 (used for this description) have 3 3 / 4 teleoconch whorls and have proportionally longer shells. Both the holotypes and lot C.470914 have 13 spiral ribs on the last whorl. Examination of the Australian Museum collections revealed a possible lot of paralectotypes for Pellamora laseroni Iredale, 1943. This lot was previously lost, and this species was not included in Ponder's (1984) revision of Iravadiidae. This material of P. laseroni have taller shells with 4 1 / 2 whorls and are a similar size and proportion to the shells of C.470914, but the last whorl only has 10 spiral ribs (Iredale 1943). There are many lots in the Australian Museum collections that could be assigned to either I. capitata or I. laseroni, and it appears that the size and number of ribs on the spiral whorl vary considerably. Without molecular material or intact specimens, it is not possible to further delineate these forms.
Operculum. Oval to D-shaped; growth striae concentric; nucleus positioned at midpoint on obtusely angled umbilical margin. Exterior surface with fine growth striae; interior surface with groove and thickened ridge along umbilical margin, and two low radial ridges emerging from nucleus, one parallel to umbilical margin.
Radula. Central tooth with subequal cusps diminishing in size outwardly; central cusp either slightly enlarged or equal; basal denticles absent; lateral margins of tooth straight. Marginal teeth with subequal cusps; inner marginal teeth with ∼20 cusps; outer marginal teeth either present or (possibly) absent.

External morphology and colouration in life.
Posterior foot margin simple, lacking metapodial tentacle. Cephalic tentacles with yellow pigmentation, lacking black bands. Pallial tentacles absent. Kuroda and Habe (1954) described the Japanese iravadiid Fairbankia (Wakauraia) sakaguchii and erected a new subgenus for the species. They tentatively placed Wakauraia as a subgenus of Fairbankia based on shell features, and the subgenus was subsequently synonymised with Fairbankia and further nested in Iravadia (Fairbankia) by Ponder (1984). The type of Iravadia (Fairbankia), Operculum (Fig. 9E). As for genus.

External morphology and colouration in life
Penis (Figs 5B, 9I). Cylindrical, rugose; duct opening at indentation on distal tip; one glandular swelling midway on outer edge. Bright yellow patch of pigment on inner edge near base.

Distribution (Fig. 7E)
Collected alive at several coastal localities in Queensland between Great Sandy Strait and Cooktown. Rare, found on damp or submerged leaf litter (in small, permanent pools) or under logs. The species seems to be associated with sandy (rather than muddy) substrate on the mid to high shore in estuarine mangrove forests.

Remarks
Wakauraia fukudai n. sp. has a finely sculptured shell that is distinct from any other Australian iravadiids. It closely resembles the Japanese species W. sakaguchii (originally placed in a subgenus of Fairbankia), but has a smaller, narrower shell and diffuse rather than banded yellow pigment on the tentacles (compared to W. sakaguchii figured by the Japanese Association of Benthology 2012). A peculiar feature of W. sakaguchii is the apparent absence of outer marginal teeth. There were few specimens available for morphological examination, so this may be a peculiarity of the two radulae mounted for scanning electron microscope examination.

Etymology
Named for my colleague Dr Hiroshi Fukuda, as a small note of appreciation for his efforts collecting the type specimens of this species and his many other contributions to the study of Australian mangrove microgastropods.
Operculum (Fig. 10E). Oval; growth striae concentric; nucleus positioned at midpoint on rounded umbilical margin. Exterior surface with fine growth striae; interior surface with groove and thickened ridge along umbilical margin and two low radial ridges emerging from nucleus.
External morphology and colouration in life (Fig. 10G-I). Head-foot cream or white with opaque black and white speckles on neck, snout, head and dorsal surface of foot. Posterior margin of foot slightly indented with a short, broad pallial tentacle on the dorsal surface. Cephalic tentacles with white speckles and four to six narrow, uneven, irregularly spaced black bands.
Penis (Fig. 6C). Flattened, L-shaped; duct opening at distal end; two glandular swellings, on outer edge at midpoint and near tip.

Distribution (Fig. 7B)
The species described here is found commonly in Australia in mangrove forest and muddy/rocky estuarine habitat along the northern coast of Australia between Hervey Bay, Queensland and Port Hedland, Western Australia. Pseudomerelina mahimensis is also found throughout Asia, including India (the type locality is in Bombay) and Thailand (Brandt 1974), but it has not yet been determined that the Asian and Australian animals are conspecific.
Pseudomerelina cf. mahimensis is sometimes found in very high densities, especially in mangrove forests where the substrate is rocky and subjected to a strong coastal influence.

Remarks
The species description is based exclusively on Australian specimens of Pseudomerelina cf. mahimensis. The tentative referral of this material to P. mahimensis is based on the lack of information in the literature about the Asian forms of this species, and the discovery in other iravadiids of significant geographical variation within supposedly homogeneous, wide-spread species (such as the I. quadrasi species complex). Molecular evidence shows that although the overall level of COI sequence divergence in P. cf. mahimensis across Australia is relatively low, samples from Western Australia are slightly divergent from those in other areas and the within-species divergence was higher than for other species (2.42%) ( Table 2). It is likely that if Australian specimens were compared with those from the type locality in Bombay, India, genetic divergence would be higher still. Without that evidence, the Australian material examined here has been provisionally referred to P. cf. mahimensis. Specimens in good condition are easily distinguished from other iravadiids  by their purple banding, but worn shells resemble those of I. goliath, I. quadrina and I. pilbara n. sp., which are of a similar size and shape and also have nodular sculpture. Unlike the observations made by Ponder (1984), the operculum of P. cf. mahimensis was found to have radial ridges on the interior surface of the operculum.
Operculum. Elongate oval, with flared anterior umbilical margin; growth striae concentric; nucleus positioned at midpoint on obtusely angled umbilical margin. Exterior surface with fine growth striae; interior surface with broad groove (abruptly terminating anteriorly), raised ridge on umbilical margin and two radial ridges emerging from nucleus, one parallel to umbilical margin.

External morphology and colouration in life.
Head-foot semi-transparent pinkish-white, with or without additional pigmentation. Posterior foot margin with shallow indentation, lacking metapodial tentacle. Cephalic tentacles with cream to bright yellow speckles, broad, black band at 3 / 4 length and black tips. Pallial tentacle absent.
Penis. Relatively short, tapering towards tip; duct opening not determined; outer base and middle of penis encircled by row of glandular lobes, inner surface with broad glandular swelling. Opaque white, without pigmentation.

Remarks
Pellamora was introduced by Iredale (1943) to accommodate differences between Asian Iravadia and the Australian species Iravadia australis and Pellamora procera (the latter name is now regarded as a synonym of P. australis), but was synonymized with Iravadia (Fairbankia) by Ponder (1984). At present, there is little support for a close relationship between P. australis and the type species of Fairbankia, I. (Fairbankia) bombayana. The shell of I. (F.) bombayana has numerous weak spiral ridges and long, dense periostracal hairs, while P. australis and P. splendida n. sp. both have fewer strong spiral ridges and only minute fringes of periostracal hairs. Further examination of the type species and others belonging to Iravadia (Fairbankia) may support the synonymisation of Pellamora, but the latter name is here reinstated pending further discoveries. The species pair P. australis and P. splendida n. sp. have been removed from Iravadia s. str. (as recognised here), based on both molecular and morphological data but pending further examination of the type species of Iravadia. They form a sister clade to Fluviocingula, and are characterised by their distinct shell shape and sculpture, tentacle pigmentation and penial morphology. Presently only two species, both Australian, are recognised in Pellamora.

Type material
Holotype of Pellamora procera. Darwin, Northern Territory, coll. J. Laseron, in shell sand, could not be located in the Australian Museum collections.
Other material. Thirteen wet lots and six dry lots from Queensland between Gladstone region and Townsville; five wet lots and three dry lots from Northern Territory around Darwin. See supplementary data for full list of material examined.

Redescription
Shell (Figs 2, 11A-E). Up to seven whorls including protoconch (estimated due to protoconch loss), length 3.76-5.68 mm (excluding lost protoconch), diameter of last whorl 1.62-1.92 mm (Table 7). Brownish-orange; periostracum corrugated with numerous very fine spiral fringes between spiral ribs but worn away on ribs. Sculpture of evenly low spiral ribs (eight or nine on last whorl). Outer lip of aperture slightly prosocline; external varix moderately developed.

External morphology and colouration in life
Penis (Figs 6D, 12D). As for genus, with ∼10 glandular lobes on outer surface. Distribution (Fig. 7C) Found uncommonly in mangrove habitat on the coast of Queensland between Gladstone region and Townsville, and also known from Darwin in the Northern Territory. There are no records in the Australian Museum from the coast of the Gulf of Carpentaria.

Remarks
The apparently discontinuous distribution of P. australis between Cape York Peninsula and Darwin may be an artefact of low collecting effort in this relatively inaccessible region, or patchy habitat. Molecular data support the continuity of P. australis between the two regions, with 2.20% within-species COI sequence divergence across its range. In contrast, P. australis and P. splendida n. sp. are separated by 12.46% COI sequence divergence (Table 2).

Description
Shell (Figs 2, 11F-I). Up to 7 1 / 2 whorls including protoconch, length 7.93 mm, diameter of last whorl 3.28 mm (only one adult specimen available, Table 7). Cream; periostracum corrugated with numerous very fine spiral fringes between spiral ribs and longer fringes on spiral ribs. Sculpture of moderately tall spiral ribs (eight on last whorl), first and second spiral ribs spaced further apart and taller than other ribs. Outer lip of aperture moderately prosocline; external varix strongly developed.

External morphology and colouration in life
Opaque yellow speckles on sides of neck and dorsal surface of head; diffuse grey pigment only at base of tentacle near eye; snout entirely unpigmented.

Distribution (Fig. 7C)
Known only from the type locality in Broome, Western Australia.

Remarks
Despite the similar shells of P. australis and P. splendida, n. sp., these two closely related species can be readily distinguished by the larger and paler shell of P. splendida n. sp., which also has more prominent and widely spaced first and second spiral ribs and more conspicuous periostracal fringing. Pellamora splendida n. sp. and P. australis are separated by 12.46% COI sequence divergence ( Table 2). The distribution and adult anatomy of P. splendida n. sp. remain unknown, because only a single adult specimen (holotype) and three juveniles (paratypes) were collected from a single location in Broome, Western Australia.

Etymology
Named for its large, attractive shell.

Redescription
Shell. Ovate to elongate-conical, delicate; convex to almost straight-sided whorls with deep or shallow suture; either non-umbilicate or with narrow umbilicus; up to eight whorls including protoconch, length 2-7 mm. Transparent grey, brown or orange; periostracum thin, either simple or with hairs. Sculpture of multiple, indistinct, spiral rows of minute, oblong pits. Aperture elongate oval, anterior margin rounded, posterior margin angled; outer lip strongly prosocline; external varix absent.

Remarks
The Japanese type species of Fluviocingula, F. nipponica, was described in sufficient detail to be reasonably certain that the Australian species described here are congeneric. However, modern reports on micromolluscan fauna refer Japanese specimens to Fluviocingula elegantula (A. Adams, 1861) (Hasegawa 2000; Japanese Association of Benthology 2012). The most useful distinguishing feature of this group is the presence of spiral pitted sculpture on the shell. Fluviocingula was reduced to a subgenus of Iravadia by Ponder (1984), a decision that is reversed here on the basis of molecular evidence.
Other material. Three wet lots from Karumba, Queensland; six wet lots from Darwin and Tiwi Islands, Northern Territory. See supplementary data for full list of material examined.
Operculum (Fig. 14B). As for genus, except with groove and ridge on umbilical margin and broad, U-shaped ridge running longitudinally through centre of operculum.
Radula (n = 4) (Figs 5H, 14A). Central tooth 4 + 1 + 4; basal denticles absent. (Fig. 14C,  D). Head-foot grey; bright yellow speckles densely covering dorsal surface of head, neck and ventral surface of foot; dorsal surface of snout cream to yellow except for ring of black pigment behind white lips; ventral and lateral surfaces of snout black. Cephalic tentacles with cream speckles, broad black bands at 3 / 4 length and black tips. Posterior foot margin indented, lacking metapodial tentacle. Pallial tentacles absent.

External morphology and colouration in life
Penis (Figs 6F; 14E). Relatively short, broadest distally; duct opening not determined; outer middle of penis with lobate glandular swellings, tip of penis tapering to sharp, glandular point attached to broad, flat glandular appendage. Grey with bright yellow speckles.
Distribution (Fig. 7D) Known only from two regions of northern Australia, based on the collections of the Australian Museum and recent fieldwork; Darwin (Northern Territory) and Karumba (Gulf of Carpentaria, Queensland). The distribution of F. resima may extend further east and west and is presumably continuous between these two regions based on genetic similarity.

Remarks
The holotype specimen is smaller than most adult specimens of F. resima and quite worn, but the shell shape  Notes: Ht, holotype; Pt, paratype; SpP, spiral rows of pits on last whorl. Juvenile specimens were not measured. and faint traces of spiral pitted sculpture confirm its identity. Fluviocingula resima can be distinguished from the other Australian species, F. superficialis n. sp., by its larger size, more convex whorls, greater number of spiral rows of pits, absence of basal denticles on the central tooth of the radula and presence of bright yellow pigmentation on the head-foot. COI genetic sequence divergence between F. resima and F. superficialis is 15.62%, but within-species divergence for F. resima is only 0.60% (Table 2). Paratypes. Same data (WAM S.82654, 10; AMS C.476013, 10). Other material. Five lots from Western Australia between Derby and Broome. See supplementary data for full list of material examined.
Operculum (Fig. 14G). Adult operculum not observed; juvenile operculum as for genus, but without internal ridge.
External morphology and colouration in life (Fig. 14H, I). Head-foot grey; white speckles behind eyes and on dorsal surface of head and snout; snout with black pigment on ventral and lateral surface and black ring behind lips. Cephalic tentacles with white speckles, broad black bands at 3 / 4 length and black tips. Posterior foot margin indented, lacking metapodial tentacle. Pallial tentacles absent.
Penis. Not observed.

Distribution (Fig. 7D)
Known from mangrove forests in northern Western Australia, between Port Hedland and the northern Kimberly region.

Remarks
Recent collections of this species from Port Hedland and Broome resulted in many live juveniles and several dead adult shells. The holotype was selected as a mediumsized specimen with an animal in the shell, but it may not be fully adult as there are larger dead shells among the paratypes. Fluviocingula superficialis can be distinguished from F. resima by its smaller shell with almost straight-sided whorls, absence of periostracal hairs, fewer rows of pitted spiral sculpture and white rather than yellow pigmentation on the head-foot. The central tooth of the radula also has basal denticles, unlike F. superficialis. COI genetic sequence divergence between F. superficialis and F. resima is 15.62%, but within-species divergence for F. superficialis is only 1.80%.

Etymology
Named for the superficial resemblance of this species to a juvenile of the related species Fluviocingula resima.

Uncertain classification -'Clade A' Remarks
Based on molecular phylogenetic analysis and morphological characterisation, Nozeba and Auricorona n. gen. do not belong in Iravadiidae s. str. However, these two genera are quite dissimilar to each other and also do not obviously belong in any other truncatelloid family. Ponder (1984) recognised a division within Iravadiidae s. l. between Iravadia and the other extant iravadiid genera with paucispiral opercula (Nozeba, Rissopsis, Chevallieria, Acliceratia, Hyala, Ceratia and Liroceratia) (see Iravadiidae, Remarks). Nozeba, and perhaps some or all of these other genera with a paucispiral operculum, belong in 'Clade A'. Although currently consisting of only two genera, it is anticipated that 'Clade A' will eventually comprise a large portion of the diversity previously assigned to Iravadiidae. A recently described Japanese species, Ceratia nagashima Fukuda, 2000, shares some anatomical and conchological features with Auricorona queenslandica n. gen. and n. sp. (see Remarks for Auricorona, below), suggesting that Ceratia (or at least C. nagashima) also belongs in 'Clade A' rather than Iravadiidae.
In the molecular analysis, 'Clade A' holds a highly supported sister relationship to Tornidae (although separated by moderately long branches), but the shell and soft-tissue anatomical characters of Nozeba and Auricorona do not closely resemble those of tornids. For example, tornids have short-spired to discoidal shells and a multispiral operculum (Ponder and de Keyzer 1998) while Nozeba and Auricorona have tall-spired shells and paucispiral opercula. Nozeba and Auricorona n. gen. both have a single right (posterior) pallial tentacle and a simple posterior foot, unlike the paired posterior and single anterior pallial tentacles and indented posterior foot with a metapodial tentacle recorded in tornids (Ponder 1994). While 'Clade A' and Tornidae probably share a common origin, there is currently no basis for including Nozeba and Auricorona n. gen. in the latter family.
Elachisinidae was not represented in the molecular analysis, but this group of truncatelloid marine microgastropods shows some similarities to both 'Clade A' and Aenigmula n. gen.. Elachisinids have a moderately tallspired (ovate-conic) shell, dome-shaped protoconch and a posterior tentacle (sometimes with an additional anterior pallial tentacle) (Ponder 1985). Either 'Clade A' or Aenigmula n. gen. (or neither) may be synonymous with Elachisinidae, but this hypothetical relationship is speculative and untested. One species of elachisinid, Elachisina ziczac Fukuda & Ekawa, 1997, has been placed in Nozeba in at least one publication (Japanese Association of Benthology 2012), reflecting the possibility of a relationship between these two groups. The Hong Kong iravadiid Lantauia taylori Ponder, 1994 has a squat, 'vitrinelliform' shell shape, but lacks a pallial tentacle.
'Clade A' is distinguished from Iravadiidae s. str. by its paucispiral operculum, posterior pallial tentacle and domed protoconch. Although Nozeba and Auricorona n. gen. are not very similar to each other, they share a few morphological characters. They are both minute (< 3 mm shell height) and have a radula with a distinctive central tooth bearing prominent lateral branches and well-developed basal denticles on the lower margin, unlike Iravadiidae s.

Description
Modified from Ponder (1984), including only characters known to be shared by all Recent taxa currently assigned to Nozeba (see Remarks).
Shell. Conical to elongate-conical; straight-sided to slightly convex whorl profile with smooth suture; umbilicate or non-umbilicate; up to 5 1 / 4 (?) whorls including protoconch, length to ∼3 mm. Smooth, or with raised spiral sculpture on initial whorls, entire shell or base of shell only. Aperture teardrop-shaped, anterior margin rounded or excavated, posterior margin sharply angled; outer lip orthocline; external varix absent.
Radula. Central tooth either with few cusps including prominent central cusp, or numerous small cusps.

Remarks
Recent species currently recognised in Nozeba are N. emarginata (New Zealand), N. mica Finlay, 1930 (New Zealand), N. topaziaca (Hedley, 1908), ?N. striata Ponder, 1984 (Philippines, deep water) and ?N. lignicola Hasegawa, 1997 (Japan, deep water). Also, Elachisina ziczac was recently transferred from Elachisina to Nozeba (Japanese Association of Benthology 2012). Nozeba? striata and ?N. lignicola were hesitantly included in Nozeba by their describing authors, because their shell and radular characters differ from those of N. topaziaca and N. emarginata (Ponder 1984;Hasegawa 1997). Hasegawa (1997) remarked that the radula and shell of N. lignicola closely resemble those of N. striata (as figured by Ponder 1984), with raised spiral ribs and a distinct central radular tooth with numerous equal cusps. It is not yet clear whether these species should be placed in a new genus (or perhaps in Auricorona n. gen., see Remarks for that genus), but they do not appear to be congeneric with N. topaziaca.
The type species of Nozeba, N. emarginata, has a simple, conical shell that is somewhat similar to that of both N. topaziaca and Aenigmula criscionei n. gen. and n. sp.. It is therefore possible that N. emarginata is in fact congeneric with Aenigmula criscionei n. gen. and n. sp.. This possibility cannot be explored using the available information, so it is preferable to provisionally retain the current species composition of Nozeba. See Ponder (1984) for a discussion of other possible synonyms and the fossil history of Nozeba.

External morphology and colouration in life
Penis (Fig. 6G). Wide, short; enclosed duct opening at distal end; three glandular swellings on inner edge. Covered with grey speckles, densest around base of penis.

Distribution
Distributed throughout eastern Australia, from southern Tasmania (D'Entrecasteaux Channel) through Bass Strait (King Island), central and eastern Victoria (as far west as Port Phillip), throughout New South Wales and as far north as Gladstone in Queensland (Fig. 6F). Most specimens were collected from estuarine seagrass habitat (either by sweeping or dredging), but some specimens were collected Figure 15. Nozeba topaziaca (Hedley, 1908). A, B, Representative specimen, Broken Bay, NSW, AMS C.327063; C, holotype of Eulima topaziaca Hedley, 1908

Remarks
A full description of the genus is not given because the description of the only known species of Auricorona is given below and can be taken as a description of the genus. Although provisionally placed here in the same group as Nozeba ('Clade A'), the only species currently included in Auricorona (A. queenslandica n. gen. and n. sp.) is unlike any of the other 'iravadiids' examined in this study. The most relevant comparisons are to the Japanese iravadiids Ceratia nagashima and Nozeba lignicola. The shells of those Japanese species have a similar size, shape and spiral sculpture to A. queenslandica n. gen. and n. sp. They all have paucispiral opercula, but that of C. nagashima is more elongate and has a distinct pattern on the internal surface, unlike Auricorona n. gen. Ceratia nagashima also has a shallow indentation in the posterior margin of the foot, a unique configuration of the central tooth and a shorter protoconch. Despite these differences, C. nagashima, N. lignicola and A. criscionei n. gen. and n. sp. are likely to belong in the same family-level group.

Etymology
Named for the externally visible pigmented ring around the opercular lobe, which resembles a golden (auri-) ring or the halo of light (corona) visible during a solar eclipse. Protoconch (Fig. 16D). Tall, with initial whorl slightly descending, second whorl rapidly descending. Smooth, with ∼10 weak spiral threads on last whorl; transitional varix absent.

Distribution (Fig. 7G)
Known from only three locations in the Great Sandy Strait region of central Queensland. Auricorona queenslandica lives in the back of muddy mangroves and during recent field work was found living beneath mangrove leaves that had settled on the damp surface of the mud, rather than in pools.

Remarks
Auricorona queenslandica is a minute species that might easily be overlooked when sampling micromolluscs in mangrove habitat across its distribution in central Queensland. When examined alive it is instantly recognisable by the ring of vivid orange pigment that encircles the opercular lobe. This feature is visible externally through the delicate and transparent operculum.

Etymology
Named for the narrow distribution of this species on the coast of central Queensland.

Uncertain higher classification Remarks
Aenigmula criscionei n. gen. and n. sp. described below is the sole representative of an unknown family or higher ranked group. Aenigmula criscionei was included in this study because it closely resembles the 'iravadiid' Nozeba, with its minute, smooth, ovate-conic shell. Molecular analysis provides strong support for the position of Aenigmula n. gen. in a section of the truncatelloid phylogeny separate from any recognised iravadiids. The surprising discovery that Aenigmula n. gen. is, on the basis of the molecular analysis, not closely related to either Iravadiidae s. str. or Nozeba ('Clade A') presents a challenge for assigning this taxon to a particular truncatelloid family.
Aenigmula n. gen. is nested in a well-supported section of the molecular phylogeny between Caecidae and a clade composed of Calopiidae and Clenchiellidae. The position of Aenigmula n. gen. on a discrete branch of the tree precludes its placement in any of these related families, and there are no evident characters to contradict this conclusion. As discussed above (see Remarks for 'Clade A'), there may be a relationship between Elachisinidae and 'Clade A' or Aenigmula n. gen., based on morphological and conchological characters (but not yet tested by molecular data). While this identity is possible, it is speculative and requires testing using molecular data. This taxon is distinguished from Iravadiidae s. str. by its paucispiral operculum, configuration of the central radular tooth and domed (rather than flat) protoconch. Few morphological characters have been identified to distinguish Aenigmula n. gen. from 'Clade A', despite strong molecular support for the separation of these groups from each other and related taxa. As this taxon is monotypic, no description has been given for the unknown group of higher rank or Aenigmula n. gen. (see instead the description of A. criscionei n. gen. and n. sp.).

Remarks
Only one species is recognised in Aenigmula n. gen., and there is little morphological or ecological information available to characterise the genus. It is necessary to recognise a new genus for Aenigmula criscionei n. gen. and n. sp., based on its position in the molecular phylogeny of Truncatelloidea in which it is not closely related to any other iravadiid and shows little morphological similarity to other non-iravadiid taxa to which it is apparently related.

Etymology
Named for the 'little puzzle' (Latin, puzzle = aenigma) that the only known species brings to our understanding of truncatelloid evolution.

External morphology and colouration in life, and penis.
The bodies of specimens preserved in 95% ethanol are entirely white, but no other morphological features could be determined. Distribution (Fig. 7H) Known from a single location in mangrove forest at Rapid Creek, Darwin, living on mangrove leaf litter in pools.

Remarks
This species is an unanticipated discovery and requires further work. Not only is it currently unassignable to a truncatelloid family, nothing is known about its softtissue anatomy, and little about its ecology or mode of life. From the limited information available, it appears in most respects to be a 'typical' truncatelloid gastropod. Molecular evidence suggests that it is most closely related to Caecidae, Clenchiellidae and Calopiidae, with which it has little in common except its minute size (like the other  three families) and presence in a mangrove habitat (like most Clenchiellidae).

Etymology
Named for my colleague Dr Francesco Criscione, who collected the type (and only known) specimens of this species and kindly contributed the material to this study.

Discussion
This study has recorded six new species and two new genera from Australia. Two formerly synonymised species and two genera have also been reinstated. This newly recognised diversity is a significant addition to the known iravadiid fauna and presents an opportunity to re-evaluate the classification of the group. This is particularly necessary because while some of the new species are closely related to known iravadiid taxa, others are phylogenetically distinct. An interesting and unexpected outcome of this study was the discovery that some taxa that were believed to be widespread throughout Australia or the Indo-Pacific region actually have two or three molecularly distinct species represented in Australia. The taxon previously recognised as Iravadia quadrasi is now, with the benefit of molecular data, known to comprise three allopatric but morphologically similar species, each restricted to a different region of the northern Australian coastline. Likewise, the species formerly known as 'Iravadia' australis and 'Iravadia' resima are composed of species pairs with northern/eastern and western distributions. These findings increase the number of known species, but do not significantly alter our understanding of iravadiid evolution. In contrast, the discovery of Auricorona queenslandica and Aenigmula criscionei introduces a new element to our knowledge of Iravadiidae, and demands a critical re-evaluation of the classification and evolutionary interpretation of the group.