Three New Species of the Freshwater Shrimp Genus Caridina from Australia †

: Three new species of the genus Caridina are described from the northernmost part of Australia. Caridina darwin n. sp. resembles Caridina temasek Choy and Ng, 1991 but differs in the armature of the rostrum, the development of epipods on the pereiopods and the absence of an appendix interna on the male first pleopods. Caridina magnovis n. sp. resembles Caridina serratirostris de Man, 1892 but differs in the armature of the ventral margin of the rostrum, a shorter stylocerite, a stouter carpus of the first pereiopod, the number and size of spiniform setae on the third and fifth pereiopods, the shape of the preanal carina and the size of the embryos (referred to as “eggs” in most previous publications). Caridina wilsoni n. sp. resembles Caridina gracilirostris de Man, 1892 but differs in the size of the embryos and in some length to width ratios of the segments of the pereiopods. Detailed morphological descriptions of all three new species are given. A molecular phylogeny (mt DNA 16S) supports the morphospecies hypothesis and illustrates the phylogenetic relationship with morphologically similar species from outside Australia.


Introduction
The species-rich freshwater shrimp family Atyidae De Haan, 1849 (Crustacea: Decapoda: Caridea) currently comprises 46 genera with 544 species and the number of taxa described is constantly increasing [1].A significant number of species of the largest genus within the Atyidae, Caridina H. Milne Edwards, 1837 are known from Australia but are still awaiting formal species descriptions [2].Some new species were recently described from the northern part of Australia like Caridina malanda Choy, Page, de Mazancourt and Mos, 2019, a species belonging to the Caridina zebra species complex [3], Caridina biyiga Short, Page and Humphrey, 2019, a member of the C. thermophila species group [4] or Caridina pagei Christodoulou and De Grave, 2023, a morphological cryptic species related to Caridina sobrina Riek, 1953 [5].Christodoulou and De Grave [5] also provided an overview on the species of genus Caridina from Australia.Short et al. summarized the information on Caridina originating from the Northern Territory and gave a morphological key for their identification [4].Page et al. [6] already suggested the existence of species complexes of problematic, cryptic Australian species within the genus Caridina.In 2007, Page et al. studied the phylogenetic and biogeographic relationships of Australian atyid freshwater shrimps, including Caridina, and identified potential species radiations.They found several genetically distinct and putatively new species of the genus Caridina.This material was subsequently included in a paper on the molecular and conservation biogeography of Caridina shrimps from northwestern Australia [7] and, with additional samples, to discuss the relationship within shrimp of the genus Caridina from the Northern Territory [4].
This study presents a detailed morphological examination of some of the material from Page et al. [7], supported by molecular data.It confirms that the genetically distinct clades from Page et al. [7] are distinct species.Finally, we describe three new species from Australia.

Materials and Methods
All samples were obtained using dip and hand nets and preserved using 96% ethanol.Specimens were subsequently transferred into 70-80% ethanol for long-term preservation.A proportion of the preserved specimens were dissected and morphometric data were taken using a BMS 143 Trino Zoom dissecting microscope with an ocular grid.Details on setae and mouthparts were observed using a Reichert Biovar compound microscope.Rostral characters were taken from all examined specimens.Drawings were made from microphotographs using Adobe Illustrator following Coleman [8,9].The setae terminology used mostly followed that of Short et al. [10].
All materials examined were deposited in museum wet collections at the Museum and Art Gallery of the Northern Territory, Darwin (NTM); the Western Australian Museum, Perth (WAM); Queensland Museum, Brisbane (QM); and the Museum für Naturkunde (Museum of Natural History), Berlin, Germany (ZMB).The abbreviations used were as follows: cl = carapace length (measured from the postorbital margin to the posterior margin of the carapace); ov.= ovigerous; and juv.= juveniles.
DNA was extracted from abdominal tissue using a Qiagen Blood and Tissue Kit (Qiagen N.V., Hilden, Germany) according to the manufacturer's instructions.A fragment of the mitochondrial 16S rRNA (16S, ~590 bp) gene was amplified by polymerase chain reaction (PCR) and sequenced using primers 16S-F-Car and 16S-R-Car1 [11].Amplifications were conducted in 25 µL volumes containing 50-100 ng DNA, 1× PCR buffer, 200 µM of each dNTP, 0.5 µM of each primer, 2 mM MgCl 2 and 1 U of Taq polymerase.After an initial denaturation step of 3 min at 94 • C, 35 cycles of 30 s at 94 • C, 60 s at 50 • C and 60 s at 72 • C were performed, followed by a final extension step of 5 min at 72 • C. PCR products were sent to Macrogen Europe for purification and sequencing of both strands of the amplified gene fragments using the primers as given above.Contigs of forward and reverse strands were assembled using Geneious Prime (v.2019.2.1) and corrected by eye.In addition, published sequences, notably twelve sequences of the three new species published by von Page et al. [7], Cook et al. [12], Short et al. [4] and de Mazancourt et al. [13], have been included in the analysis (Table 1).Sequences were aligned with MAFFT [14].To determine the best substitution model for Bayesian inference analyses (see below), hierarchical likelihood ratio tests were carried out with jModelTest [15].Based on the Akaike Inference Criterion (AIC), the GTR + I + G model was selected.The new sequences (n = 22) have been deposited in GenBank (for accession numbers and museum voucher numbers, see Table 1).Phylogenetic trees were reconstructed by Bayesian inference (BI) [16] using MrBayes 3.2.6 [17].The MCMCMC algorithm was run with four independent chains for 10,000,000 generations, with samplefreq = 500 and burnin = 25%.Maximum likelihood analyses were run with IQ-TREE [18] and branch support was obtained through the implemented ultrafast bootstrap (1000 replicates) [19].BI and ML analyses were run with the model specified above (for IQ-TREE, see [20]).In addition, maximum parsimony (MP) analyses were performed using the heuristic search algorithm as implemented in PAUP* [21], with gaps treated as the fifth base.Support for nodes was estimated by bootstrap analysis (1000 bootstrap replicates with 10 random addition replicates each).
Mouthparts and branchiae: Mandibles (Figure 1I) without palp, incisor process ending in irregular teeth, molar process truncated.Lower endite of maxillula (Figure 1J) broadly rounded with long pappose setae and few rows of serrate setae near margin and scattered cuspidate setae proximal, upper endite elongate, with numerous distinct cuspidate setae flanked by a row of serrate setae on inner margin and few pappose setae subterminal, palp slender with few simple setae and one conical spiniform seta near tip.Upper endites of maxilla (Figure 1K) subdivided, furnished with numerous pappose, simple and papposerrate setae arranged in rows, palp slender, scaphognathite tapering posteriorly, fringed with long, curved serrulate setae on truncate posterior margin and plumose setae inferiorly.Palp of first maxilliped (Figure 1L,M) terminating in finger-like extension, caridean lobe and flagellum with plumose setae, base with numerous pappose, papposerrate and serrate setae arranged in rows and tufts.Podobranch on second maxilliped (Figure 1N) well developed; exopod slender, protruding propodal segment, with long plumose setae distally; basis and merus with rows of pappose setae, distolateral margin of penultimate segment with numerous pappose and serrate setae; dactylus with numerous serrate setae flanked by a row of pappose setae.Third maxilliped (Figure 2A) slender, with two arthrobranchs; exopod slender, exceeding distal margin of antepenultimate segment, with few long plumose setae distally; antepenultimate segment of endopod slender, with few simple setae basally; penultimate segment slender, with straight lateral margins, with few simple setae distally, inner margin with rows of short serrate setae; ultimate segment slightly shorter than penultimate segment, tapering distally, ending in a claw-like curved spiniform seta (nail), proximal to nail with few cuspidate setae; inner margin densely covered with rows of serrate setae; epipod well developed, with distal hook.Branchial apparatus as shown in Table 2; well-developed (with distal hooks) epipods present on third maxilliped and first pereiopod (Figure 2B).Vestigial epipods (visible only after cleaning and staining of appendages) on second and third pereiopods.
Pleopods: Endopod of male first pleopod (Figure 2I) elongated subtriangular, lacking appendix interna, 3.2 times as long as proximal width, 0.38 times as long as exopod, with long plumose setae on outer and distal margins, with few pappose setae on inner margin.Male second pleopod with appendix masculina rod-shaped (Figure 2J), 8.3 (n = 1) times as long as wide, 0.63 (n = 1) times as long as endopod, armed with strong spiniform setae on inner margin and distal margin; appendix interna originating from about 0.4 times the length of appendix masculina, reaching to about 0.75 times the length of appendix masculina.
Size: Postorbital carapace length of apparently adult specimens on the second and third pereiopods, the lack of an epipod on the fourth pereiopod vs. a well-developed (with distal hooks) epipod on the second pereiopod, a reduced epipod on the third and a vestigial epipod on the fourth pereiopod in C. temasek).Furthermore, the proposed new species differs from C. temasek by lacking an appendix interna on the first pleopod of males (vs.with well-developed appendix interna in C. temasek) and smaller embryo size (0.59-0.64 × 0.37-0.39mm vs. 0.70-0.80× 0.44-0.54mm in C. temasek according to the original description).There is still some confusion about the shape of the preanal carina in C. temasek.Choy and Ng [23] did not give a drawing of this part, but in the verbal description, they wrote "Pre-anal carina with about 10 setae", suggesting that they did not observe a distinct hook-like spine.In their paper on Atyidae from peninsular Malaysia and Singapore [24], Cai et al. wrote that the preanal carina in C. temasek is "triangular, no spine or with a very tiny spine".Four of five paratypes of C. temasek observed for the present study show a prominent hook-like spine on the preanal carina, as do all four topotypical specimens from Singapore listed above.The holotype of C. temasek lacks a spine on the preanal carina (Y.Cai, pers.comm.).It seems likely that specimens assigned to C. temasek with and without a hook-like spine on the preanal carina might belong to different species.All specimens of Caridina darwin n. sp.lack a spine on the preanal carina.
Caridina darwin n. sp. is also similar to Caridina excavatoides Johnson, 1961.Both species lack an appendix interna on the endopod of the first pleopod in males.The proposed new species differs from C. excavatoides by a longer sixth pleomere (0.60-0.67 times the length of the carapace vs. less than 0.6 times the length of the carapace in C. excavatoides).Furthermore, Caridina darwin n. sp.differs from this species in the shape of the palp of the first maxilliped (with a distinct finger-like extension vs. ending in a triangular extension in C. excavatoides).Caridina  (Figures 3 and 4).

Description
Cephalothorax and cephalic appendages: Small shrimp with carapace lengths of 2.1-4.3 mm (median 2.9 mm, n = 16) in apparently adult specimens.Carapace (Figure 3A,B) smooth, antennal spine almost fused with suborbital lobe, pterygostomial angle rounded.Rostrum (Figure 3A,B) extending to end of second segment of antennular peduncle or to distal end of antennular peduncle, 0.53-0.76(median 0.64, n = 8) times as long as carapace, nearly straight or slightly convex, ending in acute tip; dorsal margin armed with 17-29 teeth over entire length, including 5-8 postorbital; teeth above orbit larger than anterior and posterior teeth, ventral margin with 3-6 teeth located in distal half of rostrum.
Mouthparts and branchiae: Shown in Figure 3I-O and Figure 4A.Similar to Caridina darwin n. sp.except for the incisor process of mandible (Figure 3I,J), comparably well-developed and numerous irregular teeth and palp of first maxilliped (Figure 3M,N) terminates in a triangular extension.Third maxilliped (Figure 4A) was slender, with the last segment of maxilliped slightly shorter than penultimate segment.Branchial apparatus as shown in Table 3, with one well developed and one reduced arthrobranch on third maxilliped and one arthrobranch on first pereiopod.
Pleopods: Endopod of male first pleopod (Figure 4J) elongated triangular, lacking appendix interna, 2.0 (n = 1) times as long as proximal width, 0.40 times as long as exopod, with long plumose setae on distal margin, and few shorter plumose setae on inner and outer margins.Male second pleopod with appendix masculina rod-shaped (Figure 4K), 10.0 (n = 1) times as long as wide, 0.83 (n = 1) times as long as endopod, armed with strong spiniform setae on inner margin and distal margin; appendix interna originating from about 0.4 times the length of appendix masculina, reaching to about 0.65 times the length of appendix masculina.
Size: Postorbital carapace length of apparently adult specimens 2.1-4.3 mm.Colouration: Variable.Some specimens reddish with a pale median dorsal line extending from the rostrum to the sixth pleomere, some specimens mottled with reddish and creamy chromatophores, smaller specimens mostly transparent [4,25].
Distribution: The species is known from rivers draining into the Timor Sea.From the Fitzroy River in Western Australia eastwards to the Howard River in the Northern Territory.
Etymology: The species name magnovis is derived from the rather large size of the embryos of the proposed new species compared to C. serratirostris.
Common name: North Australian Chameleon Shrimp.Remarks: The general morphology of Caridina magnovis n. sp., with a straight rostrum, armed throughout close to the tip and a high number of postorbital teeth (5-8), a long stylocerite, slender chelipeds and distal margin of the telson armed with long plumose setae clearly exceeding a lateral pair of strong spiniform setae and the variable coloration of living specimens resembles that of Caridina serratirostris de Man, 1892.The proposed new species differs from C. serratirostris in the armature of the rostrum, i.e., ventral teeth located in the distal half of the rostrum vs. ventral teeth placed at midlength of the rostrum in C. serratirostris.The stylocerite of Caridina magnovis n. sp. is shorter, reaching to 0.89-1.05 of the first segment of antennular peduncle vs. clearly exceeding this segment (1.09-1.47 times the length of the first segment).Carpus of first pereiopod is 2.12-3.22times as long as wide in Caridina magnovis n. sp. vs. 2.63-4.67 times in C. serratirostris.Dactylus of third pereiopod with 6-8 spiniform setae vs. with 4-5 in C. serratirostris.Dactylus of fifth pereiopod with 16-18 spiniform setae on flexor margin, first seta not markedly enlarged vs. with 11-17 spiniform setae, first seta enlarged in C. serratirostris.Uropodal diaeresis with 12-15 movable spiniform setae vs. with 16-21 movable spiniform setae in C. serratirostris.Preanal carina rounded, without prominent tooth in most specimens (vs.preanal carina with a prominent hook-like tooth in all specimens of C. serratirostris).Embryos 0.66-0.76× 0.39-0.49mm vs. 0.37-0.44× 0.20-0.26mm in C. serratirostris.
Mouthparts and branchiae: Shown in Figure 5I-N and Figure 6A.Similar to Caridina darwin n. sp.except that palp of first maxilliped (Figure 5L,M) terminates in a short blunt extension.Third maxilliped (Figure 6A) with last segment shorter than penultimate segment.Branchial apparatus as shown in Table 4.
Size.Postorbital carapace length of apparently adult specimens 1.9-4.3mm.Colouration: Body transparent.Rostrum red, red chromatophores also arranged in a narrow line running along the sternites of the pleon, on the distal part of the exopod of the uropod and a red transverse line at the dorsal margin of the third pleomere.Embryos are greenish (Wilson 2008; Short, Page and Humphrey 2019).
Distribution: The species is known from rivers flowing into the Gulf of Carpentaria (Queensland) and to the Timor Sea near Darwin (Northern Territory).
Etymology: The species is named after the biologist and aquarist Dave Wilson, collector of some of the type material and an enthusiast on Australian fish and shrimp fauna.
Common name: Darwin Red Nose Shrimp.Remarks: Caridina wilsoni n. sp.Resembles Caridina gracilirostris De Man, 1892, both in coloration and in general morphology (body very slender; rostrum strongly curved upwards with few widely spaced teeth on proximal part of dorsal margin, no postorbital teeth; preanal carina with one hook-like tooth; distal margin of telson with few spiniform setae).The proposed new species differs from this species in the larger size of the embryos (0.59-0.66 × 0.34-0.43mm vs. 0.34-0.39× 0.21-0.25 mm in C. gracilirostris), a slightly stouter carpus of the second cheliped (3.73-4.94times as long as wide vs. 4.69-5.26times as long as wide in C. gracilirostris), merus of third pereiopod 8.56-9.67 times as long as wide vs. 9.55-10.22 in C. gracilirostris, propodus of fifth pereiopod 3.07-3.54times as long as dactylus vs. 4.0-4.25 in C. gracilirostris.Some of these morphological characters overlap slightly (length-to-width ratio of carpus of second cheliped, length-to-width ratio of merus of third pereiopod), but by combining the morphological characters, these two species can be distinguished.The proposed new species also resembles C. gracillima Lanchester, 1901, known from the Malayan peninsula, especially in the size of the embryos (0.59 -0.66 × 0.34-0.43mm vs. 0.55-0.66× 0.35-0.40mm in C. gracillima) [26], but differs from this species in its preanal carina, bearing a hook-like spine vs. preanal carina rounded without a spine in C. gracillima.

Molecular Systematics and General Discussion
In the molecular phylogeny based on a fragment of the mitochondrial 16S rRNA gene, the sequences of each of the three new species form a strongly supported clade (Figure 7).Their relationships with their closest relatives are also well supported.Caridina darwin n. sp. is closely related to Caridina excavatoides, which is its sister species, and the allied species Caridina malayensis and C. temasek.Caridina magnovis n. sp. are found in a sister group relationship with C. serratirostris.Caridina wilsoni n. sp. is the sister species of C. gracillima and is also closely related to Caridina gracilirostris and C. neglecta.These relationships closely reflect the morphological similarities of the new Australian species as described above.Each species also differs greatly genetically from its closest relative (minimum p-distances: 6.5% Caridina wilsoni n. sp.-C.gracillima; 9.1% C. darwin n. sp.-C.excavatoides; 12.6% C. magnovis n. sp.-C.serratirostris), providing additional clear support for their status as distinct species.Page et al. [7] found close evolutionary relationships of Australian species with non-Australian taxa from the surrounding regions and suggested that Caridina wilsoni ("Gulf1") and C. magnovis ("W WA4") are endemic to Australia.This is not surprising for amphidromous species, nor in cases such as Caridina magnovis n. sp., whose closest relative, C. serratirostris, is a species that is widespread throughout the Indo-Australian Archipelago (IAA), and from which C. magnovis n. sp.appears to have evolved into an Australian endemic species.The sister species of Caridina wilsoni n. sp., C. gracillima, has a similar embryo size to C. wilsoni n. sp., but both are closely related to the very widespread amphidromous species Caridina gracilirostris (the extent of the distribution of the equally closely related C. neglecta is poorly known), so it is not too surprising that Caridina gracillima and C. wilsoni n. sp.occur at the western and eastern ends of the IAA, respectively.Perhaps most interesting of all in this context is C. darwin n. sp., which has relatively small embryos in comparison to its closest relatives with medium-to large-sized embryos and a distribution restricted to the Sunda Shelf.Although Australian endemism is likely for all three new species, their medium embryo size differs from other Australian endemic species with large embryos, indicating highly suppressed larval development, such as Caridina biyiga from Leichhardt Springs, Kakadu National Park, Northern Territory (embryo size 0.75-1.0mm) [4].
In conclusion, the rather limited molecular dataset presented here to support morphospecies in an integrative taxonomic approach is not suitable to discuss evolutionary and biogeographic aspects in more detail.But the few observations made above underline once again the diverse biogeographic origin of the Australian atyid fauna and its potential for more extensive studies of the adaptations associated with the colonization of this continent.
darwin n. sp.differs from C. excavatoides in the development of the epipods on the third maxilliped and pereiopods in the same way as C. temasek.The dactylus of the fifth pereiopod has a smaller number of spiniform setae on the flexor margin (21-49 vs. 50-60 in C. excavatoides).Caridina sp.WA 4 Page, von Rintelen and Hughes 2007; Caridina sp.'WA 4' Wilson, 2008; Caridina sp.WA 4 Cook, Page and Hughes 2011; Caridina 'sp.WA4' Short, Page and Humphrey, 2019;

Figure 7 .
Figure 7. Molecular phylogenetic tree (BI, 16S) showing the relationship of the three new species to morphologically similar congeners.Numbers on branches are, from the top, Bayesian posterior probabilities and ML/MP bootstrap values.The scale bar indicates the substitution rate.

Figure 7 .
Figure 7. Molecular phylogenetic tree (BI, 16S) showing the relationship of the three new species to morphologically similar congeners.Numbers on branches are, from the top, Bayesian posterior probabilities and ML/MP bootstrap values.The scale bar indicates the substitution rate.

Table 1 .
List of sequenced specimens with geographic origin, voucher information and data source."ZMB DNA" vouchers are DNA samples from specimens with an origin in the pet trade without any specimen remains.
1.8-3.7 mm.Colouration: Transparent to brownish with cream-colored transversal stripes on the pleon (Wilson 2008; Short, Page and Humphrey 2019).Distribution: The species occurs in rivers around Darwin, Northern Territory, Australia.Etymology: The species name darwin is derived from the Darwin river catchment, the natural habitat of the proposed new species, and is used as a noun in apposition.Common name.Blackmore River Caridina.Remarks: Based on morphology, Caridina darwin n. sp. is related to a group of species that includes Caridina propinqua de Man, 1908 and Caridina laevis Heller, 1862 as the best known members.Within this group of species related to C. laevis, two species with large embryos-Caridina temasek Choy and Ng, 1991 and Caridina excavatoides Johnson, 1961, both occurring in freshwaters of the Malayan peninsula-are most closely related to the proposed new species.Caridina darwin n. sp.differs from C. temasek by a higher number of teeth on the dorsal margin of the rostrum (15-21 vs. 14-16 in C. temasek), the development of epipods (vestigial epipods (only visible after clearing and staining of the appendages)