Two new species of Atopobathynella (Parabathynellidae, Bathynellacea) from the Pilbara region, Australia

Fifteen species of the Gondwanan genus Atopobathynella have been described so far from four countries. The position of the genus within the family Parabathynellidae and its species relationships are controversial due to the different characters and species considered by different authors, however most of the phylogenetic reconstructions are based solely on morphological characters. In the past few decades, the arid zones of Western Australia, including the Pilbara region, have been recognized as a hotspot for subterranean fauna. Material is constantly collected to produce Environmental Impact Assessments to protect and manage the subterranean environment. In 2009 the Cleaverville Iron Formation, in the Pilbara region, northern Western Australia, was proposed to be mined, therefore subterranean fauna sampling was conducted as per legislation. Preliminary molecular studies of stygofauna identified two distinct Parabathynellidae lineages from two ridges, however no morphological descriptions were carried out at that time. In this study, we describe two new species from Yarrie and Callawa ridges, respectively: Atopobathynella yarriensis sp. nov. and A. degreyensis sp. nov.. The new species show differences in male antennal organ, ventral tooth of mandible, number of teeth on distal endite of maxillula, absence or presence of epipod of thoracopod I, number of setae on article 4 of thoracopod II, length of distal external seta of exopod of all thoracopods, outer lobe of male thoracopod VIII, number of spines on uropodal sympod and furcal rami, length of the external dorsal plumose seta of the furcal rami. Additionally, we integrated the morphological data with sequences of three genetic loci, Cytochrome Oxidase c Subunit I (COI), 12S rRNA (12S), and nuclear 18S rRNA (18S), calculating p-distances and constructing a multigene molecular phylogeny to support morphology, and explore the monophyletic status of the genus and the relationship amongst its species. The two new species were well supported in our phylogeny, however they appear distantly related. http://zoobank.org/urn:lsid:zoobank.org:act:2D8DA822-6C08-45B5-B707-6F63A1B69DE2


Introduction
The genus Atopobathynella Schminke, 1973(Parabathynellidae Noodt, 1965) has Gondwanan distribution, with 15 species described from four countries: one in Chile, one in New Zealand, eight in Australia (various states from Western Australia (WA) to Tasmania (Fig. 1)), and five in India (Bandari et al., 2017;Cho, Humphreys, & Lee, 2006;Noodt, 1964;Ranga Reddy et al., 2008;Ranga Reddy & Totakura, 2015;Schminke, 1973) (Supplemental Material S1).Schminke (1981b) highlighted the presence of the genus also in Madagascar, however no description was completed.The relative position of Atopobathynella species within the genus and the genus within the family are unclear.Schminke (1973) considered Atopobathynella closely related to Chilibathynella Noodt, 1963 based on antennal organ, exopod of thoracopods and furcal rami morphology.He also included Notobathynella Schminke, 1973 and Correspondence to: Giulia Perina.E-mail: giulialittlepear@ gmail.comHexabathynella Schminke, 1972 in this systematic relationship since species of these genera show complex antennal organs.The genus Kimberleybathynella Cho et al., 2005 is also morphologically closely related to Atopobathynella.This relationship was strengthened with the description of additional Atopobathynella species from India which showed intermediate characters between the two genera (Bandari et al., 2017).Phylogenetic reconstructions based on well-established Atopobathynella generic and species-specific characters resulted in distinct topologies, however A. wattsi-A.glenayleensis and A. compagana-A.chelifera-A.valdiviana clades are always stable in two analyses of the morphological data (Bandari et al., 2017;Cho, Humphreys, & Lee, 2006).The use of genetic data to corroborate morphology and the construction of multigene phylogenies, to both understand the relationships amongst taxa and capture the cryptic diversity in the subterranean realm (King et al., 2022), is becoming a standard approach (Abrams et al., 2013;Camacho et al., 2013Camacho et al., , 2018;;King et al., 2012;Perina et al., 2018Perina et al., , 2019a)).
In the past few decades, the north of Western Australia has been recognized as a hotspot for subterranean fauna mostly due to environmental surveys, often related to mining development (Eberhard et al., 2005;Guzik et al., 2010).Western Australian environmental protection legislation includes specific subterranean fauna guidelines (Environmental Protection Authority, 2013, 2016), therefore much material is collected during surveys and frequently sequenced to overcome difficulties in morphological study and to delimit species for Environmental Impact Assessment (EIA).However, formal descriptions and systematic studies are rarely undertaken.
In 1993 BHP Billiton opened the Yarrie mine, and in 2009 the company proposed to mine the nearby Callawa ridge (Fig. 2).Subterranean fauna surveys were conducted in the area, and preliminary molecular analysis on Parabathynellidae identified two lineages from Callawa and Yarrie ridges, respectively (Helix Molecular Solution, 2009).Additionally, Matthews et al. (2020) included supplemental material from these ridges to a more extensive Parabathynellidae phylogeny, however no formal descriptions were carried out at that time.
In this study, we describe two new species of Atopobathynella from the arid zone of Western Australia discussing the molecular phylogeny and the genus distribution.

Study area and groundwater sampling methods
The study area is located 170 km east of Port Hedland, north of the De Grey River catchment (Fig. 2), a major catchment in the Pilbara bioregion (IBRA7: Interim Biogeographic Regionalisation for Australia (Department of the Environment, Australian Government, 2013), WA.The area comprises low hills, plateaus alternate with sandy plains and sporadic claypans.Ridges of about 260 m (Australian Height Datum) comprise the Cleaverville Iron Formation (such as Callawa and Yarrie ridges), and the fractured and weathered Archean-aged formation (Hickman et al., 1983;Williams, 2003).Many ephemeral creeks run through the various ridge gaps (for example Shay and Kimberley Gap) and between the ridges (Dames & Moore, 1992).Aquifers occur in the alluvial deposits, mainly in the De Grey River and Eel Creek, and in the Archean bedrock that constitutes the ridges in the area.At Yarrie Ridge there is evidence of a perched aquifer confined by granite and mudstone layers that underlie the banded ironore formation (BIF) (see figure 3.2 in Dames & Moore, 1992).The perched aquifer is recharged by rainfall through infiltration of the permeable BIF on the plateau, however the low hydraulic conductivity of the lower layers makes regional water discharge (the alluvial of the De Grey and Eel Creek) very slow (Dames & Moore, 1992;Williams, 2003).Granitic rocks abut the ridges of the Cleaverville Formation (Hickman et al., 1983) consequently groundwater contained in the BIF is in part isolated from regional groundwater.
The Yarrie mine opened in 1993, and in 2009 BHP Billiton proposed to mine Callawa and Cundaline ridges, hence subterranean fauna surveys were conducted there as part of the Environmental Impact Assessment (EIA) for the area to develop recommendations to the Minister of the Environment on the proposal (Environmental Protection Authority, 2009, 2021).
Parabathynellid material used in this study was collected on Callawa and Yarrie ridges during surveys conducted between 2007 and 2009 by Subterranean Ecology.Stygofauna were sampled using a plankton net lowered into pre-drilled bore holes according to the EPA guidelines (Environmental Protection Authority, 2003, 2021).Specimens were fixed and preserved in 100% ethanol and stored in a fridge or freezer for most of the time prior to this study.

DNA extraction, amplification, and sequencing
Animals were placed in a drop of propylene glycol on a concave slide to allow the dissection of the animal without compromising the DNA (Moreau et al., 2013).Body segments of the abdomen bearing no useful morphological characters were dissected under a stereo microscope using tungsten needles.Specimens were labelled individually with a Western Australian Museum registration number and measured using an eyepiece micrometer.DNA was extracted from the body segments using Qiagen DNeasy Blood and Tissue kit, following the manufacturer's protocol (Alda et al., 2007) with the only modifications being a 17 h incubation prior to extraction and eluting the DNA into a 60 ml AE buffer.
Two microlitres of DNA template were amplified in 25-mL reactions which included 1 Â Bioline MyTaq reaction buffer, 1 unit of MyTaq HS DNA polymerase, 0.2 mM of each primer.Thermal cycling was performed in a BioRad T100 system.Reactions for the COI gene fragment were run with the following conditions: 95 C for 3 min; followed by 7 cycles of 95 C for 30 s, 40 C for 30 s and 72 C for 45 s; then 35 cycles at 95 C for 30 s, 50 C for 30 s, and 72 C for 45 s.The final extension step was carried out at 72 C for 10 min.For 12S fragment reactions: 95 C for 5 min, followed by 40 cycles of 95 C for 30 s, annealing at 47 C for 30 s, and 72 C for 45 s; the final extension step was carried out at 72 C for 10 min.The same protocol was used for amplification of the 18S fragments, with the annealing temperature adjusted to 49 C. Three microlitres of PCR product were run through 2% agarose gels and visualized with a Bio-Rad GelDoc to confirm the correct PCR amplicon size was successfully amplified and sent to the Australian Genome Research Facility (AGRF, Perth) for bidirectional Sanger sequencing.
The raw chromatograms were imported into Geneious 10 software (Kearse et al., 2012).Forward and reverse reads were assembled, checked by eye, primer sequences removed and edited where required.The consensus sequences were extracted, and blasted against the nucleotide database in GenBank to detect possible contaminations.Each gene was aligned separately using the MAFFT algorithm (Multiple Alignment using Fast Fourier Transform) (Katoh et al., 2002) with default parameters.The alignments obtained were successively concatenated in Geneious.

Molecular data analyses
COI sequences were translated into amino acid chains to ensure no stop codons were present.COI and 12S uncorrected pairwise distances (P-distance) were calculated within and between species haplotypes from the same and different bores of the study area using Molecular Evolutionary Genetics Analysis (MEGA) version 7.0 (Kumar et al., 2016) to compare with former analyses done on parabathynellids (Abrams et al., 2012;Matthews et al., 2020;Perina et al., 2023).COI and 12S substitution saturation was tested implementing the Xia's test (Xia et al., 2003) (estimating the proportion of invariant (P inv ) sites prior) in DAMBE7 (Xia, 2018).The p-distance of the nuclear rRNA 18S gene was not calculated and the substitution saturation was not tested since 18S is a conservative gene (not used to discriminate Bathynellacea species), however it is useful for older divergences (Boyko et al., 2013).
Phylogenetic reconstructions of the concatenated alignment were conducted using maximum likelihood (ML) and Bayesian inference (BI) methods.The IQ-TREE phylogenomic software online (Nguyen et al., 2015) was used to build a maximum likelihood tree partitioned by gene, choosing the automatic substitution model option (which uses ModelFinder; Kalyaanamoorthy et al., 2017) Ultrafast Bootstrapping (UFBoot; Hoang et al., 2018), and the recommended SH-aLRT test (Guindon et al., 2010).It should be noted that a node is considered supported when UFBoot > ¼ 95% and SH-aLRT > ¼ 80%.Randomized Axelerated Maximum Likelihood (RAxML) analysis was also conducted using RaxML_HPC_ BlackBox in CIPRES online server (Miller et al., 2010) with partitions by gene (with GTR model for each gene) and other default values, including the recommended automatic bootstrapping stop, which determines the number of replicates sufficient to get stable support value using the MRE-based bootstrapping criterion (Pattengale et al., 2009).The optimal models of evolution for partitions, including the COI codon positions, for the Bayesian analysis were explored using PartitionFinder1 (Lanfear et al., 2012).The analysis was conducted using the greedy search option, "MrBayes" models of evolution and the Akaike Information Criterion (AIC).The Bayesian analysis was conducted using MrBayes on XSEDE (3.2.6) (Ronquist et al., 2012) run with the BEAGLE (Broad-platform Evolutionary Analysis General Likelihood Evaluator) library (Suchard & Rambaut, 2009) in CIPRES online server (Miller et al., 2010), using the Markov chain Monte Carlo (MCMC) approach.All parameters were unlinked, and four chains were run simultaneously for 10 million generations with two independent runs, sampling every 1000th generation.Convergence was established using the program Tracer v1.7 (Rambaut et al., 2018) to determine whether estimated sample size (ESS) values were above 200.A burn-in of 25% was chosen and a maximum clade credibility tree was constructed using MrBayes 3.2.6.
Trees were rooted in FigTree  Perina and Camacho, 2019).Other Parabathynellidae from Australia and other countries were included in the analyses to test the monophyly of the genus and to examine the relationships amongst the taxa within the genus.Supplemental material S2 lists the taxa included and their GenBank accession numbers.

Morphological analysis
Specimens were dissected and mounted on permanent slides following the glycerine jelly methods described by Perina and Camacho (2016).Morphology was examined using an oil immersion objective (100Â) on a Leica DM2500 and Zeiss interference microscope with phase contrast.Drawings were done using a drawing tube, digitalized through a WACOM tablet and retouched using drawing software.The material is vouchered at the Western Australian Museum (see Supplemental material S2 for voucher numbers).
The substitution saturation test performed using DAMBE (Xia, 2018) showed that both COI and 12S alignments built using representatives of different genera and species of Parabathynellidae from Australia and other countries were saturated.Therefore, we decided to remove the third codon position from the COI alignment and did not include the 12S alignment in the concatenated analyses.
The best partition schemes selected by ModelFinder in IQ-TREE online were: The best partition schemes selected by PartitionFinder1 (Lanfear et al., 2012) were: The IQTree, RAxML, and MrBayes consensus concatenated tree for the two markers provided similar topologies.In Fig. 3 we show the IQ tree highlighting the nodes supported also in the MrBayes and RAxML analyses (some nodes were collapsed in FigTree for clarity; the full IQ, MrBayes, and RAxML trees are available in the Supplemental material S4, S5, S6).Major (well supported) clades for the Australian Parabathynellidae were observed to represent the described genera: Chilibathynella, Notobathynella, Lockyerenella Little & Camacho, 2017, Arkaroolabathynella Abrams & King, 2013 (excluding A. bispinosa) Antennule 6-segmented often showing sexual dimorphism on the inner margin of the second article, with female bearing one simple seta, and male an antennal organ with one or two setae that can be modified.Antenna one-segmented, with three smooth setae (one inner proximal and two apical) and an apical plumose seta.Labrum heterodont with numerous teeth, of which middle ones are smaller.Mandible with pars incisiva with three teeth (in one case four); lobe with row of five claws (Borstenlobus).Distal endite of maxillula with five or six claws.Maxilla four-segmented (in two cases threesegmented) with setal formula: 2-4-n-1/7 (the third article can have a variable number "n" of setae, while the fourth article can have up to seven setae).Seven pairs of thoracopods.Epipod present or absent in thoracopod I, always present in thoracopods II-VII.Exopod of thoracopods I-VII one-segmented.Exopod of thoracopods II-VII with two terminal setae and one ventral subterminal one (one species has two ventral setae); endopod of thoracopods Two new species of Atopobathynella from Australia I-VII four-segmented, fourth article tiny, thoracopods II-VII with one or two very long claws on the fourth article; setal formula of endopod II-VII articles 1-3: Male thoracopod VIII small, semicircular in lateral view; protopod massive, anteriorly with dentate lobe, and often with a small protrusion; outer lobe drawn out into a conical projection directed anteriorly; basipod triangular or round/paddle-like, as large as outer lobe, with one or two setae of uncertain interpretation that could represent endopod and/or exopod.Pleopod I consisting of one seta.Sympod of the uropod with homonomous or non-homonomous row of several spines; endopod with dagger-like extension and two or three setae at its base (only one seta in one species); exopod of uropod often with ventromedial seta and 2-4 terminal/subterminal setae.TYPE LOCALITY: bore YP1063 (20 36 0 11.74 00 S, 120 20 0 23.46 00 E), Yarrie Ridge, Pilbara Region (Western Australia).
HABITAT: groundwater.ETYMOLOGY: the name of the species refers to the name of the ridge where it was collected (Yarrie Ridge).
Material examined was collected by P. Bell and B. Vine (15/07/2020) (five females and six males).Eight specimens are part of the type series.See Supplemental material S7 for type series and additional material details and WAM registration numbers.
ANTENNULA (Fig. 5A): six-segmented.First four articles wider than last two, fourth and fifth articles slightly shorter than others.Antennal organ not protruded, represented by two ventral setae.Very little inner flagellum and almost square.Articles five and six with three terminal aesthetascs.Antennular setation as in Fig. 5A.ANTENNA (Fig. 5C): one-segmented; very short almost square article with three smooth terminal/subterminal setae and one terminal plumose seta.

LABRUM (Figs 5D
): slightly convex free edge with 19 main teeth, nine on each side and one central.PARAGNATHS: absent.MANDIBLE (Fig. 5E): pars incisiva with four teeth; pars molaris with five strong and denticulate claws, the most distal slightly modified, thicker and separated from the rest, the two most proximal ones joined together; tooth of ventral edge small and triangular.Mandibular Two new species of Atopobathynella from Australia  palp short, with one long distal seta not reaching beyond the pars molaris.
THORACOPODS I-VII (Fig. 6A-G): length slightly increasing from thoracopods I-III, last four similar in size.Small epipod present in all thoracopods, about half to one-third of the length of the corresponding basipod.All basipods with one smooth, distolateral seta as long as the first article of the endopod.Exopod one-segmented in all thoracopods; exopod of thoracopod I shorter than corresponding two first articles of the endopod, exopod of thoracopods II-VII similar in length to the first two articles of the endopod.Exopod of thoracopods I-VII bearing three barbed setae, two terminal (with the internal one very long) and one subterminal slightly longer than the outer terminal seta.Endopod four-segmented, first article short with one barbed seta on thoracopod I, and no seta on the rest of thoracopods; second and third articles long and similar in length; second article with one external plumose seta in all thoracopods and two internal smooth setae only on thoracopod I; third article with one internal seta on thoracopod I and one small external distal seta on the rest of thoracopods; fourth article very reduced with two strong claws of different length on thoracopods I and II and only one long strong claw on thoracopods III-VII.Setal formula of endopods as follows: Thoracopod VIII (Fig. 5H-J): compact, like a balloon.Penial region with massive protopod with frontal protrusion.Outer lobe triangular and not defined at base in latero-external view.Dentate lobe with three distal and two proximal teeth.Inner lobe shorter than outer lobe and drawn out into rounded projection.Basipod triangular with one seta that could represent endopod or exopod, and a small frontal crest over-reaching the penile projection.
PLEOPOD I (Fig. 6H): long plumose seta located on the anterior third of the first segment of the pleon.UROPOD (Fig. 6I): sympod almost four times the length of the endopod and almost five times as long as wide, with seven equal spines and one larger distal one.Endopod as long as exopod, with a row of setules, distally drawn out into a dagger-shaped structure with terminal setules, distolateral angle bearing two long barbed setae that exceed the tip of the dagger-shaped structure.Exopod with two terminal barbed setae, and one basal plumose seta.PLEOTELSON (Fig. 6J): with two lateral very small plumose setae; anal operculum not protruded.FURCA (Fig. 6J): rami almost square, with six barbed spines; distal spine three times longer than others.Two dorsal plumose setae, the inner one very short and the outer one reaches the tip of the most distal spine.
HABITAT AND DISTRIBUTION: groundwater.ETYMOLOGY: the name of the species refers to the name of the river that runs south of the study area (De Grey River).
Material examined was collected by P. Bell, N. Stevens, E. Ridley, B. Barnet, S. Eberhard (see Supplemental material S7) (five females and 11 males).Nine specimens are part of the type series and are mounted on permanent slides; nine specimens have sequences available, and seven specimens are kept in 100% ethanol.See Supplemental material S7 for type series and additional material details and WAM registration numbers.DIAGNOSIS: Males 1.10-1.90mm, females 1.30-1.71mm; male antennal organ slightly protruded, with one long strong ventral seta; labrum with 17 teeth; female thoracopod VIII boomerang-shape with spinule; sympod of uropod with 5-6 spines.
Description male holotype (WAMC57180): Body elongated, almost cylindrical, segments slightly widening towards posterior end of body (Fig. 4B).Body length 9.6 times as long as maximum width.
ANTENNULA (Fig. 7A): six-segmented.First two articles wider and longer than the others, fourth and sixth article similar in length, but shorter than fifth one.Fifth article slightly shorter than the first three ones.Antennal organ protruded, represented by one long and strong seta on the inner distal corner of the second article.Very little rectangular inner flagellum.Article five with three terminal aesthetascs similar in length.Article six with three terminal aesthetascs unequal in length.Antennula setation as Fig. 7A.ANTENNA (Fig. 7C): one-segmented; with three smooth setae (one subterminal and two terminal) and one terminal plumose seta.LABRUM (Fig. 7D, E): flat, free edge with 17 main teeth, eight on each side and one central.PARAGNATHS: absent.MANDIBLE (Fig. 7F): pars incisiva with three teeth; pars molaris with five strong and denticulate claws, the most distal one not modified, and the two most proximal ones joined together; tooth of ventral edge absent.Mandibular palp with one long distal seta reaching beyond the pars molaris.
THORACOPODS I-VII (Fig. 9A-G): length slightly increasing from thoracopods I-IV, last four similar in size.Epipod present on thoracopods II-VII, about onethird between one-third and half of the length of the corresponding basipod.All basipod with one distolateral, smooth seta almost similar in length to first article.Exopod one-segmented in all thoracopods.Exopod of thoracopod I shorter than corresponding two first articles of the endopod.Rami similar in length to the first two articles of the endopod in thoracopod II, and slightly longer in thoracopods III-VII.Exopod of thoracopods I-VII bearing three barbed setae, two terminal (with the internal one very long) and one subterminal similar or slightly longer than outer terminal one.Endopod: foursegmented, first article short with one smooth seta on thoracopod I and without seta on the rest of thoracopods; second and third articles long and similar in length.Second article of thoracopod I with one external plumose seta and two internal smooth setae.Second article of thoracopods II-VII with only one external plumose seta.Third article with one internal seta on thoracopod I, and one small external distal seta on the rest of thoracopods.Fourth article very reduced with two strong claws of different length on thoracopod I, and only one long strong claw on thoracopods II-VII.Setal formula of endopods as follows: Thoracopod VIII (Fig. 7J, K): compact, like a balloon slightly elongated.Penial region with massive protopod without frontal protrusion.Outer lobe broad, triangular and defined at base in latero-external view.Dentate lobe club-shaped with three small distal teeth.Inner lobe shorter than outer lobe and distally bilobed.Basipod elongated over-reaching the outer and inner lobe, with two setae of uncertain interpretation on inner side.PLEOPOD I (Fig. 8H): long antero-lateral plumose seta located on the anterior third of the first segment of the pleon.UROPOD (Fig. 8I): sympod almost three times the length of the endopod and almost three times as long as wide, with four equal spines and one larger distal one.Endopod same length as exopod, with a row of setules, and drawn out distally into a dagger-shaped structure with terminal setules.Dagger-shaped structure bearing two long barbed setae on the distolateral angle that exceed the tip of the structure, with one seta twice as long as the other one.Exopod with two long terminal setae similar in length, one subterminal and one short basal one, all barbed.
Two new species of Atopobathynella from Australia PLEOTELSON (Fig. 8J): with two small plumose lateral setae (one on each side); anal operculum not protruded.
FURCA (Fig. 8J): rami almost square, with four barbed spines similar in length, and two dorsal plumose setae, with the internal seta short, 1/3 of the length of the external one.External seta twice the length of the spines.
Female paratypes differ from males in the second antennular article (Fig. 7B) that bears no antennal organ but one smooth seta instead, and boomerang-shape thoracopod VIII with spinule (Fig. 7I).

REMARKS:
The two new species, occurring respectively on Yarrie and Callawa Ridges, display a unique combination of characters within the genus Atopobathynella (see Supplemental material S8).Both species are distinguishable from each other by: male antennal organ with two and one seta, respectively, on A. yarriensis and A. degreyensis; distal tooth of pars molaris modified (different from the other teeth of the pars molaris) in A. yarriensis and unmodified in A. degreyensis; six and five teeth on distal endite of maxillula in A. yarriensis and A. degreyensis, respectively; epipod of the thoracopod I present in A. yarriensis and absent in A. degreyensis; two setae on article 4 of the thoracopod II on A. yarriensis and only one in A. degreyensis; the distal external seta of exopod of all thoracopods is similar in A. yarriensis, while it is shorter on thoracopods I and II of A. degreyensis; outer lobe of male thoracopod VIII broad in A. yarriensis and triangular in A. degreyensis; eight non-homonomous spines on sympod of the uropod and six spines on furcal rami in A. yarriensis, while five spines on sympod and four spines on furcal rami in A. degreyensis; external dorsal plumose seta of the furcal rami longer in species A. degreyensis than A. yarriensis.
A comprehensive revision of 15 species of Atopobathynella and comparison with other genera occurring in India and Australia was completed by Bandari et al. (2017), therefore here we highlight only the most relevant differences and similarities of the two new Australian species compared with the rest of the Atopobathynella species.The two new species described here have an intermediate size just above 1 mm.Other Atopobathynella species range in size between 0.6-0.8(A.nelloreensis Bandari et al., 2017) and 3.0 mm (A. wattsi Cho, Humphreys & Lee, 2006).Bandari et al. (2017) also classified the antennal organ (AO) of Atopobathynella species in three types: type I: without protrusion, but with one or two setae more or less modified; type II: with protrusion and with or without setae; type III: with protrusion and with digitiform process, but no setae.According to this classification, A. yarriensis has AO type I with two unmodified setae, as do the majority of Australian Atopobathynella species (A.compagana Schminke, 1973, A. wattsi, A. hinzeae Cho, Humphreys & Lee, 2006, A. glenayleensis Cho, Humphreys & Lee, 2006, and A. schminkei Cho, Humphreys & Lee, 2006).A. degreyensis has AO type II (as does A. chelifera Schminke, 1973), with one thick seta. A. readi Cho, Humphreys &Lee, 2006 andA. gascoyneensis Cho, Humphreys &Lee, 2006 do not have AO.The male of A. hospitalis Schminke, 1973 is unknown, so we do not know the type of AO.The two new species have similar number of teeth on the labrum as the other Australian species A. chelifera, A. compagana, A. schminkei, the Chilean A. valdiviana (Noodt, 1965), and the Indian A. indica, while the other species have more (20 and over) or fewer (10-12).A. yarriensis has the distal claw of the pars molaris of the mandible transformed, as with 10 other species, while in the other six species, including A. degreyensis, it is not. A. yarriensis has six teeth on the distal endite of maxillula like 11 other species, while A. degreyensis has only five like A. schminkei and the three Indian species.Eleven species lack epipod on thoracopod I (including A. degreyensis) while six species have it, including A. yarriensis.The two new species have three setae on the exopod of all thoracopods, like most species, two distal ones of different length and a subdistal one, similar in length to the short distal one.Most species have one seta on the first article of the endopod of thoracopod I, except the Australian species A. wattsi that has two and two Indian species that have no setae (A.operculata Ranga Reddy et al., 2008 andA. paraoperculata Ranga Reddy &Totakura, 2015).Only the new species and two other Australian species (A.wattsi and A. glenayleensis) have two internal setae on the second article of the endopod of thoracopod I, while the third article of most species, including the two new ones, has one external distal seta.Six species have the fourth article of thoracopos I bearing two claws and one seta, including A. yarriensis, while eight species, including A. degreyensis, have only two claws.
On the rest of the thoracopods the endopodial setation of the three first articles is similar in all species, the only difference is in the fourth article, which may have one or two claws.Most of the species have only one claw, like A. degreyensis.A. readi has two claws on Two new species of Atopobathynella from Australia thoracopods II-VII, A. schminkei has two claws on thoracopods II and III, and A. yarriensis has two claws on thoracopod II.The ratio between the exopod and the first two endopodial articles of thoracopods.In the new species, in some Indian ones (A.inopinata Bandari et al., 2017, A. nelloreensis Bandari et al., 2017) and two Australian ones (A.hinzeae, A. glenayleensis) the exopod is shorter than the first two articles of the endopod of thoracopod I.The exopod is longer than the first two articles of the endopod of thoracopod I in A. readi and A. gascoyneensis.In the new species, the exopod is similar in length to the first two articles of the endopod in thoracopod II, and longer in thoracopods III-VII.Male ThVIII is similar in all species, small and square or balloon shaped and with massive semicircular protopod.The basipod is triangular in most of the species and fused with exopod with one seta.
Female thoracopod VIII is reduced, spine/denticlelike, triangular/pyriform in most of the species.A. yarriensis and A. degreyensis have female thoracopod VIII boomerang-shape.The pleopod I is present in all species, including the new ones.
A. yarriensis and A. degreyensis are the only Australian species with non-homonomous spines on sympod of uropod, like all Indian species except A. indica.The number of spines is eight and six in A. yarriensis and A. degreyensis, respectively, while most Australian species have more than eight, A. operculata from India has five, and the Chilean species, A. valdiviana, has only four.Only three Indian species (A.nelloreensis, A. operculata, and A. paraoperculata) have a subdistal seta on the sympod of the uropod.The Australian species have between two and four setae on the exopod of the uropod (three and four respectively in A. yarriensis and A. degreyensis), while all Indian species have only two setae.A ventro-medial seta of the exopod is present in most Australian species including the new ones, while the Indian species do not have it.The endopod of the uropod has two setae in most species, including A. yarriensis and A. degreyensis; A. hospitalis, A. hinzeae, and A. gascoyneensis have three setae, and A. operculata only one.
In most of the species the furca has 3-5 spines, including A. degreyensis that has 4-5. A. yarriensis has six, and A. glenayleensis and A. wattsi have eight and nine, respectively.Fourteen species of Atopobathynella genus do not have protruded anal operculum, while A. compagana, A. schminkei, A. operculata, A. paraoperculata, and A. nelloreensis have anal operculum protruded.Supplemental material S8 shows the main similarities and differences amongst Atopobathynella species.

Discussion
Molecular and morphological data confirm the presence of two species A. yarrieensis and A. degreyensis, respectively on Yarrie and Callawa ridges in the De Grey Catchment (Pilbara region).They present the typical morphological features of the genus Atopobathynella, including six and one-segmented antennula and antenna, respectively, exopod of thoracopods I-VII one-segmented, balloon-shaped male thoracopod VIII, pleopod I as one seta, endopod of uropod with dagger-like extension.Several morphological characters can be used to distinguish the two new species (see Remarks).
Molecular divergences (COI and 12S uncorrected pdistance) between the two species were 14.5%, and 28.3%, respectively, which is within the COI/12S genetic distance range previously observed amongst Parabathynellidae (Abrams et al., 2012;Camacho et al., 2022;Matthews et al., 2020;Perina et al., 2023).The intra-specific COI and 12S genetic distances amongst A. degreyensis individuals (2.7% and 1.57%, respectively) was higher than that among A. yarrieensis individuals (COI ¼ 0.34% and 12S ¼ 0.15%).The difference between the species is probably due to their sampling distribution: A. yarrieensis was collected at a single borehole, despite other bores on the ridge being sampled (Subterranean Ecology, unpublished data), while A. degreyensis was collected in several bores on the northeast side of the Callawa Ridge (Fig. 2).Interestingly, two specimens of A. degreyensis sequenced from the same bore (CA0010) showed COI and 12S p-distance respectively of 1.15% and 0.87%, suggesting that even within the same bore the variability of A. degreyensis tended to be higher than in A. yarriensis.The wider distribution of A. degreyensis and genetic population structure may reflect the perched aquifer groundwater connection amongst bores.Nevertheless, both species occur in restricted areas and are considered short range endemic (Harvey, 2002).
Despite the two new species being found at geographically close locations (about 5 km between the two ridges, see Fig. 2) they are genetically distantly related to each other and to any other Atopobathynella taxa sequenced to date.Such observations indicate distinct ancestral populations and historical geographic separation of these species irrespective of likely hydrological connectivity during the Late Cretaceous when the climate was much wetter and the average annual rainfall of the Pilbara region was over 200 cm (Gonz alez-Alvarez et al., 2016).In contrast, the three Bathynellidae species occurring on Callawa (one species) and Yarrie (two species) ridges form a well-supported monophyletic clade (Perina et al., 2019b).Therefore, the two Bathynellacea families do not share consistent phylogenetic patterns in the same area.
Relationships among Atopobathynella were not well supported by the phylogeny presented here (Fig. 3).It is likely that incomplete sequences and missing data for most of the known and described Atopobathynella species (Supplemental material S1) have impacted the complete resolution of lineages in the genus.However, Atopobathynella revealed four additional well-supported clades (as indicated by node markers (green) in Fig. 3) including those from the Tanami Region (Parabathynellidae sp.(Browns Range) EM-2019, Clade 1), and the Yilgarn.Clades 2 and 3 (Fig. 3) represent taxa from an area of about 500 km 2 south of Meekatharra, Fig. 9 (Clade 2: A. sp. 2 and A. sp.PBAT015; Clade 3: A. sp.PBAT009/010/ 016/025).Lineages represented by the codes "PBAT#", may reflect new species and they were assigned by an environmental consultant based on COI distances, however further molecular and morphological work is needed to confirm the status of these lineages.The fourth well-supported clade represents Atopobathynella species sampled geographically more than 200 km apart (Clade 4: A. wattsi, A. glenayleensis, and A. sp.7).The relationship between A. wattsi and A. glenayleensis is also corroborated by the morphological analyses done by Bandari et al. (2017) and Cho, Humphreys, and Lee (2006).
Most of the taxa collected in WA come from groundwater accessed through pre-drilled boreholes, while few Atopobathynella taxa occur interstitially (such as A. schminkei, A. compagana, A. chelifera, A. hospitalis, A. spp.PBAT006 and PBAT031; see Fig. 1 and Fig. 10, interstitial/hyporheic species are indicated with an asterisk).The Gondwanan genus in Australia, hence, successfully colonized both environments.However, based on the current sequence data, species do not form a reciprocally monophyletic clade.We could therefore assume that only a small percentage of the diversity of the continent has been discovered, and we are in the presence of complex evolutionary patterns of diversification, where some ancestors were probably widespread throughout different regions during wetter periods and got gradually isolated with the increase of aridity in the Miocene (Byrne et al., 2008;Gonz alez-Alvarez et al., 2016;Van de Graaff et al., 1977).Molecular data from additional taxa from different areas could help in the understanding of the relationships amongst Atopobathynella species, confirm the monophyly of the genus, and shed some light on their history.
The phylogenetic relationship between Atopobathynella and Chilibathynella postulated by Schminke (1973) was not supported here by molecular data.Kimberleybathynella and Habrobathynella could instead be related, however not enough data are currently available to corroborate this hypothesis.Other genera with low support in our analysis were: Kimberleybathynella (however only COI was available for the species considered), and Hexabathynella, an allegedly cosmopolitan genus, nevertheless for most of the species described from different countries molecular data are not available (Perina et al., 2023).We note that Arkaroolabathynella bispinosa does not fall within the Arkaroolabathynella clade.In Abrams et al. (2013) and Perina et al. (2023), the node also showed low support (respectively PP ¼ 0.95; BS ¼ 70% and BS ¼ 29%), while in Matthews et al. (2020) the node was not supported.A. bispinosa also presents distinctive morphological characters compared with other Arkaroolabathynella species (Camacho, personal observation), therefore it is possible that A. bispinosa belongs to a new genus.The morphological study of additional material could clarify its systematic position.
It is well known that Parabathynellidae (and the sister family Bathynellidae) in Australia are an important part of stygofauna and therefore, together with other groundwater organisms, contribute to fulfil ecosystem services through water purification and nutrient cycling (Boulton et al., 2008;Griebler & Avramov, 2015;Korbel & Hose, 2011).Their environment is increasingly being exploited by human activities and the loss or alteration of their habitat could result in species extinction (Hose et al., 2015).Most members of the Bathynellacea have restricted distribution and limited dispersal abilities (Humphreys, 2008;Schminke, 1974Schminke, , 1981a)), with consequent endemisms (Asmyhr et al., 2014) often associated with single aquifers, like in the case of A. yarrieensis and A. degreyensis; hence disruptions, such as mining or dewatering, can have serious consequences for these subterranean crustaceans.A better understanding of their diversity and distribution can help their conservation and habitat management.We also want to highlight the importance of collection and preservation: the material used for this study was collected in 2008/2009, fixed and preserved in 100% ethanol, and refrigerated most of the time.These ideal conditions have allowed us to be able to obtain sequences 13 years later, while specimens kept at room temperature have very low success rate even after 12 months (Perina et al., 2019a).Storage of specimens in 100% ethanol in freezers is ideal as molecular data can be extracted many years later, saving time and avoiding further expensive field trips.
We would like to acknowledge the Western Australian Museum (in particular Dr Andrew Hosie and Ana Hara) and Museo Nacional de Ciencias Naturales de Madrid (CSIC) (Spain) for their support.We thank the Molecular Systematics Unit at the Western Australian Museum for assisting with the generation of these data.Many thanks to the reviewers who improved the manuscript with their comments.Totakura, 2015;15. A. indica Bandari et al., 2017;16. A. nelloreensis Bandari et al., 2017;17. A. inopinata Bandari et al., 2017. Abbreviations

Fig. 3 .
Fig. 3. Maximum likelihood IQ Tree constructed using the concatenate COI (third codon position stripped)-18S alignment of selected Parabathynellidae from Australia and other countries.Number near nodes represent the SH-aLRT/UFB (supported nodes > 80/95%).Green nodes are also supported in the RAxML and MrBayes analyses.
Park & Ranga Reddy, 2006from the Yilgarn region.There are many lineages, mainly from the Pilbara region, that need further studies and a systematic revision, however, these are beyond the scope of the current study and will not be discussed further.Hexabathynella and Kimberleybathynella clades were not supported by the current analysis (the support for these nodes is low in the Bayesian and RAxML analyses too, see Supplementary material S3, S4).Additionally, the Atopobathynella clade was not very well supported (PP ¼ 0.95, SH-aLRT/UFBoots ¼ 91.9/90, BS ¼ 17).However, the clades representing species were well supported, including A. degreyensis and A. yarriensis (respectively PP ¼ 1, SH-aLRT/UFBoots ¼ 99.8/100, BS ¼ 100 and PP ¼ 1, SH-aLRT/UFBoots ¼ 99.8/100, BS ¼ 99).