New and cryptic species of intertidal mites (Acari, Oribatida) from the Western Caribbean – an integrative approach

ABSTRACT The present study highlights the distribution, systematics, morphology, genetics, and ecology of two newly discovered intertidal oribatid mites from the Western Caribbean. The fortuyniid Litoribates floridae sp. nov. represents a cryptic species as it looks nearly identical to L. bonairensis. The two species can be distinguished only by subtle morphological and morphometric characteristics, whereas cytochrome oxidase subunit I gene sequences clearly separate the two taxa. The absence of morphological divergence in these disjunct species may have resulted from stabilizing selection due to the extreme intertidal environment. Litoribates floridae sp. nov. is presently known from the Florida Keys, primarily in mangrove leaf litter. The selenoribatid Thalassozetes balboa sp. nov. can be distinguished from all known congeners by a unique cuticular notogastral pattern, the presence of only two pairs of adanal setae, and two ventral teeth on each leg claw. It is morphologically most similar to T. barbara from the Eastern Caribbean. Thalassozetes balboa sp. nov. was found in Panama and Florida. This species usually occurs on rocky substrate and feeds on the intertidal alga Bostrychia. Litoribates floridaehttp://www.zoobank.org/urn:lsid:zoobank.org:act:A4B830FC-A03F-405D-9DE4-DE4C39DB6211 Thalassozetes balboahttp://www.zoobank.org/urn:lsid:zoobank.org:act:EBF8C435-5C07-4B0E-8279-2101DC9E2CD4

These recent records demonstrate that the Caribbean intertidal oribatid mite fauna is probably highly diverse and still poorly known. In the course of ongoing studies in this region, further Thalassozetes and Litoribates specimens were found on the coasts of Panama and Florida, respectively. The Thalassozetes individuals showed obvious morphological deviations from T. barbara, which occurs in the Eastern Caribbean (Pfingstl et al. 2016), whereas the Litoribates specimens from Florida exhibited a remarkable similarity to L. bonairensis, known from far distant Bonaire. Therefore, the aims of the present paper are (1) to assess the taxonomic identity of these specimens with morphological and molecular genetic means, (2) to describe new taxa in detail and discuss their morphological peculiarities, and (3) to provide their distribution patterns.

Materials and methods
Samples of intertidal algae were scraped from rocks with a knife and decaying mangrove litter was collected during low tide. Algae and litter were put in Berlese-Tullgren funnels for about 24 h to extract mites, which were then stored in ethanol (100%) for morphological and molecular genetic investigation.

Sample locations (Figure 1)
Coordinates for locations of individuals used only for morphological studies are provided in this section; coordinates for locations of individuals used for both morphological and molecular genetic studies are given in Table 1

Morphometric analyses
For morphometric investigations, specimens were placed in lactic acid (temporary slides) and measured using a compound light microscope (Olympus BH-2) and ocular micrometer.
For the study of intraspecific variation, 20 continuous variables ( Figure 2) were measured in 90 T. balboa sp. nov. individuals from three different populations in Panama (PA_37, PA_39, and PA_43). For species discrimination, 15 continuous variables  were measured in 16 L. floridae sp. nov. specimens and compared to those of 19 L. bonairensis from Bonaire (data for the latter from ). The T. balboa specimens from Panama were compared with 19 T. barbara individuals from Barbados.
Univariate statistics were performed for the species of each genus: Mean, minimum, maximum, standard deviation, and coefficient of variation (cv) were calculated to assess variation within and between species. The Mann-Whitney U test was used for comparing the means of variables between the species.
In the subsequent multivariate analyses, males and females were coded with different symbols in order to detect possible sexual dimorphism in the studied populations. To reveal patterns of morphometric variation, we conducted Principal Component Analyses (PCAs) on ln(x+1) transformed raw and size-corrected data of Thalassozetes and Litoribates. Size correction was performed by dividing each variable through the geometric mean of the respective specimen, as described in .
A linear discriminant analysis (LDA) was performed on raw and size-corrected data to reveal possible differences and to determine the most important variables differentiating Thalassozetes from Panama and Barbados and Litoribates from Bonaire and Florida. The performance of the classification by LDA was tested by calculating the percentage of specimens correctly classified by all-samples LDA and by leave-one-out cross-validation LDA. For testing the equality of means of populations, Bonferroni-corrected Hotelling's T 2 tests were used.
All analyses were performed with PAST 3.11 (Hammer et al. 2001).

Molecular genetic analyses
In total, 20 specimens of Litoribates spp. and Thalassozetes spp. collected in Barbados, Bonaire, Florida, and Panama were analysed (see Table 1). Total genomic DNA was extracted from single individuals preserved in absolute ethanol. Extraction was carried out using the Chelex method (Casquet et al. 2012) with the following adjustments: whole specimens were crushed against the tube wall in microcentrifuge tubes containing 55 µl of a 10% Chelex solution and were extracted for 20 min at 95°C. Two gene fragments were sequenced for this study: the mitochondrial cytochrome oxidase subunit I (COI) and the nuclear 18S rRNA (18S). A 567 bp fragment of the COI gene was amplified using the primer pairs Mite COI-2F and Mite COI-2R (Otto and Wilson 2001). The complete 18S rRNA (~1.8 kb) was amplified according to the PCR protocol of Dabert et al. (2010) using the recommended primers (Skoracka and Dabert 2010). PCR conditions for the COI gene fragment are given in . DNA purification (with the enzyme cleaner ExoSAP-IT, Affymetrix; and the Sephadex G-50 resin, GE Healthcare) and sequencing steps (using the BigDye Sequence Terminator v3.1 Cylce Sequencing Kit, Applied Biosystems) were conducted according to the methods published by Schäffer et al. (2008). Sequencing was performed in both directions on an automated capillary sequencer (ABI PRISM 3130xl, Applied Biosystems). Alignments were generated by means of the program MEGA6 (Tamura et al. 2013). For both gene fragments, Bayesian 50% majority rule consensus trees were generated with MrBAYES 3.1.2 (Ronquist and Huelsenbeck 2003) applying a MC 3 simulation with 20 million generations (5-7 chains, 2 independent runs, 10% burn-in, GTR+I+G model). Results were analysed in TRACER v.1.6 (Rambaut and Drummond 2007) to check for convergence and to ensure the stationarity of all parameters. Uncorrected p-distances were calculated in MEGA6. All sequences obtained from this study were deposited in GenBank (www.ncbi.nlm.nih.gov/genbank; accession numbers for COI: MK035001-MK035017, 18S: MK035018-MK035022; Table 1). For the 18S data set, all already-published fortuyniid and selenoribatid sequences were integrated in the alignment, and as there is no 18S sequence data from closely related families as for example Ameronothridae or Podacaridae, Hydrozetes served as closest related outgroup. For the COI, data set sequences of Alismobates pseudoreticulatus served as outgroup.

Drawings and photographs
Preserved animals were embedded in Berlese mountant for microscopic investigation in transmitted light. Drawings were made with an Olympus BH-2 Microscope equipped with a drawing attachment. These drawings were first scanned, then processed, and digitized with the free and open-source vector graphics editor Inkscape (https://inkscape.org).
For photographic documentation, specimens were air-dried and photographed in reflected light with a Keyence VHX-5000 digital microscope.
Morphological terminology used in this paper follows that of Grandjean (1953) and Norton and Behan-Pelletier (2009).

Etymology
The specific epithet is a noun in the genitive case, referring to Florida, the type locality of this species.

Remarks
Presently, there are only three species of this genus known.
Litoribates floridae can be easily distinguished from the Pacific L.
caelestis by the absence of the obvious reticulate cuticular surface pattern present in the latter and its larger size (mean length 360 µm vs. 322 µm). It is morphologically very similar with L. bonairensis but has several irregular longitudinal cuticular ridges in the humeral region that are either absent or only faintly developed in the latter species.
Legs. (Figure 6) Monodactylous and slender. Long, strong hook-like claws, with inconspicuous dorsal dentition. Cerotegument generally finely granular, larger granules only on distal third of all femora. Trochanter III and IV with obvious dorsal spur. All genua with ventral transversal ridge. Large elongate porose areas on ventral paraxial side of femora I and II and on paraxial dorsal aspect of femora III and IV. Kidney-shaped porose areas on paraxial dorsal aspect of trochanters III and IV. Dorsal seta d on all femora slightly thickened and barbed on outer curvature. Lateral setae of femora and genua I and II blunt, short, broadened, and slightly serrate. Ventral setae of all tibiae and tarsi long and slightly serrate ventrally. Tibia IV with two setae. Solenidia φ 1-2 on tibia I borne on vaguely delimited mound. Chaetome and solenidia given in Table 2.

Etymology
As this species was found in Panama and Florida, the name was chosen to somehow relate to both locations. The specific epithet "balboa" (given as noun in apposition) refers to the Spanish explorer Vasco Núñez de Balboa; he led the first European expedition that crossed the Isthmus of Panama in 1513 and the Panamanian currency is named after him. It also refers to Rocky Balboa, a well-known movie character, a descendant of immigrants who made his fortune in the USA. The new species similarly may be an immigrant that has now successfully settled in North America. Moreover, T. balboa sp. nov. was predominantly found on "rocky" substrate.

Remarks
T. balboa sp. nov. can be distinguished from the other three Thalassozetes species by its specific reticulate-foveate cuticular pattern, with the centre of notogaster covered with regularly distributed conglomerates of irregular circular or elliptic depressions, fading laterally and caudally in smaller foveate pattern; by two pairs of adanal setae instead of three; and by having two ventral teeth on each claw instead of one. Integument. Colour dark brown.
Gastronotic region. (Figures 7(a) and 8(a)) Oval in dorsal view, convex in lateral view; dorsosejugal suture incomplete. Nearly triangular clear spot (vestigial lenticulus) with irregular borders on anterior notogastral median area. Cuticle with specific reticulate-foveate pattern, centre of notogaster covered with regularly distributed conglomerates of irregular circular or elliptic depressions, fading laterally and caudally in smaller foveate pattern. Pair of weak and highly variable (in shape) concave notogastral ridges framing light spot. Laterad of ridges, elliptic depressions lined with fine and densely packed granules. Fourteen pairs of setiform notogastral setae (approx. 10 µm), c 1-2 , da, dm, dp, la, lm, lp, h 1-3 , p 1-3 ; seta c 3 absent. Five pairs of notogastral lyrifissures present; ia in humeral area adjacent to anterior border of elliptic cavity; im laterad of seta la; ih laterad and anterior to h 3 ; lyrifissures ip and ips laterally of seta p 3 and p 2 respectively. Orifice of opisthonotal gland (gla) laterad between seta la and lm.
Lateral aspect. (Figures 7(c) and 8(c)) Cerotegument basically finely granular, larger granules on pedotectum I and in areas surrounding acetabula. Lateral sejugal furrow broad and deep, with dense granulation. Next to lateral border of bothridium triangular protrusion orientated caudally, slightly projecting above lateral sejugal furrow. Pedotectum I small rounded, slightly projecting. Pedotectum II absent. Lateral enantiophysis present, anterior projection triangular, well developed, posterior protrusion triangular and small. Discidium present, developed as strongly projecting triangular bulge between acetabulum III and IV.
Podosoma and venter. (Figures 7(b) and 8(b)) Cerotegument with larger granules next to acetabula and fine granules surrounding anal opening. Epimeral setation 1-0-1-1, all setae setiform and smooth, seta 1b longest (approx. 45 µm), others significantly shorter (ca. 7 µm). Internal borders of all epimera well visible. A densely granulated median sternal cavity on epimeron I with slightly concave lateral borders resulting in inverted pear-like shape. Rounded genital plates with three pairs of fine setae (approx. 8 µm) arranged in longitudinal rows. Aggenital setae absent. Lyrifissure iad oblique adjacent to anterior corner of anal orifice. Outer part of preanal organ rectangular with rounded edges, inner part shaped like a transverse bar. Two pairs of short anal setae, an 1-2 and two pairs of short adanal setae, ad 1-3 (all approx. 8 µm).
Legs. (Figure 10) Monodactylous. Long, strong hook-like claws, dorsally with slight dentation, hardly visible. Two ventral teeth on each claw, proximal one conspicuous and distal one weakly developed. Cerotegument generally finely granular, larger granules on all trochanters and femora. Femora with ventral carina. No porose areas discernible. Femoral setae serrate. Lateral setae of all genua thickened, blunt and slightly serrate. Ventral setae of tarsus serrate. Famulus ε developed as short broad knob. Solenidion ω 2 in lateral position. Chaetome and solenidia see Table 2.

Morphometrics
Univariate Statistics: Litoribates. The two Litoribates species L. bonairensis and L. floridae differ significantly in only 5 of the 15 measured variables (Table 3). The variability as indicated by cv is moderate (cv ≤0.10) in all characters.
Leg I d, (l) ( l), σ (l), v´, φ 1 , φ 2 (pl), (pv), s, (a), (u), (p), (tc), (ft), (it), ε, ω 1 , ω 2 0-3-2-3-18 (explaining 25.0% of the total variation) reveals a sexual dimorphism. In L. floridae, males and females are clearly separated, while the two sexes overlap in L. bonairensis. The variable contributing the most to separation along PC 2 is the length of the genital opening gl (Table 4). Size correction resulted in a decrease of 87.1% of the total variation. After size correction, PCA no longer separates the two Litoribates species (Figure 11). The sexual dimorphism of L. floridae is not visible in the size-corrected data, but a slight separation of the sexes becomes apparent along PC 2 in L. bonairensis. The variable with the highest loading on PC 2 is dcg ( Table 4).
The LDA on both raw and size-corrected data revealed that the two species are mainly separated by the variable ll, and the Hotelling's T-test showed highly significant differences (p < 0.001) between them. In the raw data, 100% of the individuals can be correctly classified by all-samples LDA and 80% after leave-oneout cross-validation. After size correction, all-samples LDA still correctly classifies 97.14% and leave-one-out cross-validation LDA correctly classifies 74.29%.
Univariate Statistics: Thalassozetes. Thalassozetes barbara from Barbados and T. balboa from Panama differ significantly in 11 of the 20 measured variables ( Table 5). The variability of almost all characters is moderate with a cv ≤0.10 in all populations, only efw 2 shows slightly higher values (cv between 0.11 and 013).
Multivariate analyses: Thalassozetes. In PCA on both raw and size-corrected data, PC 1 (explaining 33.6% of the total variation in raw data and 29.1% in size-corrected data) separated T. barbara from T. balboa with a very small overlapping area ( Figure 12). The variation within T. balboa seems to be larger than that of T. barbara, and the larger variation is not caused by different sample sites as the three populations from Panama always overlap. In PCAs on both raw and size-corrected data, the variables with the highest loadings on PC 1 were ll and efw2 (Table 6). A pronounced sexual dimorphism is revealed by PC 2 (explaining 25.6% and 17.0% of the total variation, respectively) in both raw and size-corrected data (Figure 12). The differences between the two sexes are more pronounced in T. barbara than in T. balboa PC 2 is strongly correlated with size (geometric mean) in the raw data (r = 0.96), indicating that females of both species are considerably larger than males. Variables con-tributing most to PC 2 were gl and gw, length and width of the genital opening. In the size-corrected data, there is only a weak correlation between PC 2 and size (r = 0.68). Here, the differences between the two sexes are again mainly caused by the variables gl and gw, which have the highest loadings for PC 2 ( Table 6). The total variation decreased strongly, by 88.7%, after size correction.
LDA showed again that separation of T. barbara from T. balboa was mainly caused by the variables ll and efw2 in both raw and sizecorrected data. All-samples LDA on raw data correctly classified 100% of all specimens and 96.3% after leave-one-out cross-validation. After size correction, the percentages of correctly classified individuals slightly dropped to 99.1% in all-samples LDA and increased to 97.3% after leave-one-out cross-validation, Hotelling's T-test revealed highly significant differences between the two species.

Molecular genetic analyses
To ensure the genetic distinctness of the investigated Litoribates species, classical barcoding by means of COI data was applied ( Figure 13). The COI topology clearly separates L. floridae and L. bonairensis. The distinctness of these two species is underlined by a clear "barcoding gap" with high genetic differences. The highest intraspecific distance (uncorrected p-distance) amounted to 1.4% and the lowest interspecific distance to 9.6%.
The 18S topology clearly demonstrates the monophyly of all investigated genera (Figure 14). Although the 18S rRNA gene is typically not suitable for species identification, T. balboa and T. barbara could be clearly separated with high statistical support. Thalassozetes species form the sister group of the Schusteria and Thasecazetes clade and together with the recently described Indopacifica species and Rhizophobates (="Thalassozetes shimojanai," for explanation of diverging taxon name, please refer to discussion) they build a monophyletic clade of Selenoribatidae. Litoribates spp. are placed as sister group of the Pacific Alismobates species; however, members of Fortuyniidae are shown to be paraphyletic with the genus Fortuynia being placed closer to Selenoribatidae.

Systematics
Litoribates floridae sp. nov. and L. bonairensis have nearly identical morphology and represent cryptic species, which Knowlton (1993) defined as species that are difficult to distinguish using traditional morphology-based taxonomy. Apart from the irregular notogastral ridges in the humeral area being more pronounced in L. floridae, both species share all diagnostic characters and are difficult to distinguish under the microscope (see Figure 15). The morphometric analyses also demonstrate that the species are similar in size and shape, and only a few characters differ significantly between them. Litoribates floridae is basically slightly larger with a broader prodorsum (variables dPtI and db) and a longer lenticulus (ll); additionally, the sexual dimorphism is more pronounced in this species, with females having obviously larger genital openings than males. In contrast to morphology, genetic data clearly separate the species as shown by the large barcoding gap between the taxa; hence; the distinctness of the species is confirmed despite the striking morphological similarity.
We speculate that either of two evolutionary processes could be responsible for their near-identity. First, speciation may have been too recent for detectable morphological differences to have accumulated (Holland et al. 2004). Second, extreme or homogeneous environmental conditions may cause stabilizing selection, resulting in highly conserved morphologies (e.g. Bickford et al. 2007). The intertidal habitat represents an extreme environment, and both these intertidal species dwell in the same littoral habitat; therefore, stabilizing selection may well be responsible for their morphological similarity.
Within the genus Thalassozetes, T. balboa sp. nov. shares most characteristics with T. barbara. They differ only slightly, particularly in diverging notogastral surface structure, and numbers of adanal setae and of ventral teeth on the leg claws. Since the two species occupy the same geographic region, they may be derived from a common Caribbean ancestor. Molecular genetic data are not comprehensive enough to address this issue as there are no comparable data for the Mediterranean T. riparius and the Indo-Pacific T. tenuisetosus. In GenBank, there is a sequence presented under the name Thalassozetes shimojanai but this should be corrected to Rhizophobates shimojanai Karasawa & Aoki, 2005(Pfingstl et al. 2018. The separation of R. shimojanai from true Thalassozetes species is further supported by their isolation in the phylogenetic tree based on 18S rRNA sequences (Figure 14).

Ecology
The genus Litoribates has been associated mainly with mangrove leaf litter , and the present study supports this generalization. All L. floridae sp. nov. specimens were found in high numbers in a mangrove leaf litter sample, except for one specimen found in a High loadings explaining differences between the species are given in bold. Figure 11. Scatter plots of PCA of L. bonairensis from Bonaire and L. floridae sp. nov. from Florida. Open symbols refer to females; filled symbols refer to male specimens. For better display of sexual dimorphism, each gender is given in a separate hull.
sample of algae growing on rock under a mangrove tree, which may have been accidental. The actual food source of Litoribates species, whether the decaying leaf litter itself or some other material, such as coccal algae, bacteria, or fungi, is not yet identified . Members of the genus Thalassozetes also seem to be associated with specific coastal habitats. With one exception, there is a clear preference for intertidal rocky substrates, e.g. cliffs, rocks, and stones. The type species T. riparius was considered a characteristic inhabitant of the Mediterranean rocky littoral shore (Schuster 1963), T. barbara was suggested to prefer rocky habitats in the Caribbean area (Pfingstl et al. 2016), and T. balboa was predominantly found in algae growing on rocky substrate, with only four specimens collected from algae on mangrove roots. The exception is T. tenuisetosus, which was reported from intertidal algal sediment on a beach in India (Bayartogtokh and Chatterjee 2010). The specific food source used is not known for T. riparius and T. tenuisetosus, but T. barbara and T. balboa were observed to feed on the intertidal alga Bostrychia.

Distribution
When the genus Litoribates was recorded for the first time in the Caribbean region, a wider distribution was anticipated, even though it was known from only Bonaire and the close Venezuelan coast , and its trans-Caribbean distribution is now confirmed by our discovery of L. floridae. Their mangrove leaf litter habitat may facilitate wide dispersal as fallen leaves are easily washed into the sea and may drift on the surface over long distances.
The selenoribatid T. balboa sp. nov., on the other hand, was found in Panama and Florida, and despite the large distance between these areas, we suggest this species has a wide distribution along the Caribbean Central American coastline, with ongoing gene flow between the populations. Hydrochorous dispersal, i.e. drifting along ocean currents, has been considered the main mode of long distance transport for oribatid mites (e.g. Schatz 1991;Pfingstl 2013c), and in this case, the strong northward flowing Gulf Stream could facilitate dispersal along this vast shoreline. Since the Gulf of Mexico represents a large potential Minimum (min), maximum (max), mean (x), standard deviation (sd), and coefficient of variation (cv), marked grey if equal or higher than 0.10. Results of Mann-Whitney-U test are given; *p < 0.05, **p < 0.01, ***p < 0.001. barrier, either T. balboa sp. nov. has colonized the whole Gulf coast or it has crossed the ocean from the Peninsula of Yucatan to Florida by drifting on the Gulf Stream. Presently, T. balboa sp. nov. seems to be restricted to the Western Caribbean, while T. barbara was only found in the Eastern Caribbean. As both species use the same habitat and food source, a syntopic occurrence or overlapping distribution areas seem rather unlikely.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by the Austrian Science Fund (FWF) under Grant [P 28597]. High loadings explaining differences between the species are given in bold. Figure 13. Bayesian inference tree of three fortuyniid species based on COI sequences constructed by means of MrBayes applying the GTR+I+G model. Posterior probabilities for main nodes are shown above branches. Sequences obtained from GenBank are marked by *. For details, see Table 1. ** Mean of intra-(blue, red) and interspecific (violet) uncorrected p-distances given in per cent.   Table 1.