Morphology and taxonomy of the genus Ramazzottius (Eutardigrada; Ramazzottiidae) with the integrative description of Ramazzottius kretschmanni sp. nov

Abstract The species of the genus Ramazzottius (Ramazzottiidae, Eutardigrada) are among the most common and widespread tardigrade species in the world. Most of the 28 Ramazzottius species have been described only with morphological characters which were most of the time represented only with drawings. The discovery of a new species of this genus in the Black Forest (Germany) provided the opportunity to compare this species with the type specimens of ten Ramazzottius species, to propose the status of species dubia for Ramazzottius edmondabouti, and through new photographs to elucidate the anatomy of animals and eggs (in particular of the head sensory regions, eye spots, buccal tube, ornamentations of the dorsal posterior cuticle, and morphology of egg processes). These thorough observations led to a better understanding of the diversity and evolution, not only of this cosmopolitan genus, but also of other eutardigrade genera. The new species Ramazzottius kretschmanni is described with an integrative approach integrating morphological (light and electron microscopy observations and morphometric data) and molecular (cox1 and ITS2 genes) data. The PTP and ASAP analyses confirmed the validity of the new species from a molecular point of view. The new species is morphologically similar to Ramazzottius oberhaeuseri, but is distinguishable by the smooth cuticle, the presence of a “cheek-like” area on the head, and the size of egg processes as well as different sequences of the molecular markers.

Within this genus, only two species have been described with an integrative approach: R. oberhaeuseri (the neotype population; Stec et al. 2018) and Ramazzottius sabatiniae Guidetti, Massa, Bertolani, Rebecchi & Cesari, 2019b. All the other species were described only by morphological and morphometric characters using Light Microscopy [LM], plus Scanning Electron Microscopy [SEM] for only five of them (Kaczmarek et al. 2006;Dastych 2011;Stec et al. 2017Stec et al. , 2018Guidetti et al. 2019b). In the papers in which Ramazzottius species are described, the morphological characters are illustrated with photos (not always of good quality) in only half of them, and for 13 species, animals and eggs morphologies are represented only by drawings. The drawings can be useful because they can reproduce multifocal images of a structure/character and can emphasize the details useful for taxonomic identification. On the other hand, drawings are not objective representations, as they are subjected to the interpretation of authors, who may neglect details that could be useful for species comparisons.
The discovery of a new Ramazzottius species from the Black Forest in Germany provided the opportunity to compare this species with the type specimens of several Ramazzottius species, to take new photographs of these specimens elucidating the anatomy of animals and eggs, and to develop a better understanding of the diversity and evolution not only of this cosmopolitan genus but also of other eutardigrade genera.

Tardigrade sampling and morphological analyses
Tardigrades were extracted from a moss growing on tree bark (C4322-Probe103) collected in October 2016 by Ralph O. Schill in the Black Forest (Schwarzwald, Germany). These specimens were morphologically analyzed with LM, and those belonging to a new species were also analysed with SEM or characterized with a molecular approach.
To extract tardigrades, fragments of the moss sample were placed in distilled water for about half an hour. After soaking, the sample was sieved (sieve meshes: 500 μm and 38 μm) to separate tardigrades and eggs from the substrate. Animals and eggs were then isolated using a needle, removed with a glass pipette under a stereomicroscope, and mounted on slides in Hoyer's medium. Specimens for SEM observations were fixed in boiling absolute ethanol for few minutes, then were rinsed three times in absolute ethanol, desiccated by evaporation, mounted on stubs, and sputter coated with gold. Observations with SEM were carried out with EVO-LS 10 (Carl Zeiss company), available at the Institute of Evolution and Ecology at the University of Tübingen (Germany).
Observations with LM and measurements were carried out under both phase contrast [PhC] and differential interference contrast [DIC] up to the maximum magnification (100× oil objective) with a Leica DM RB microscope equipped with a Nikon DS-Fi 1 or an AmScope MU1803 digital cameras, at the Department of Life Sciences, University of Modena and Reggio Emilia (UNIMORE), Italy. Measurements of the lengths of the animals and their cuticular structures (i.e. claws, structures of feeding apparatus) were made according to Kaczmarek and Michalczyk (2017) and Stec et al. (2018); structures were measured only if they were in proper position. Morphometric data were handled using the "Parachela" ver. 1.6 template available from the Tardigrada Register (Michalczyk & Kaczmarek 2013), updated with the Thorpe's normalization of the data (as in Massa et al. 2021) according to Bartels et al. (2011a).

Molecular characterization
Prior to the molecular analysis, individuals were observed and identified with LM using the method described in Cesari et al. (2011) to obtain photo voucher specimens. Genomic DNA was extracted from four separate animals. The extractions were performed with QuickExtract™ DNA Extraction Solution (Lucigen), following the manufacturer's protocol. Investigations of molecular genetic markers were carried out using fragments of mitochondrial (cytochrome c oxidase subunit 1: cox1) and nuclear (internal transcribed spacer 2: ITS2) genes. The cox1 gene was amplified using primers and PCR protocols described in Cesari et al. (2009) (cox1, Forward: LCO 5'-GGT CAA CAA ATC ATA AAG ATA TTG G-3', Reverse: HCOoutout 5'-CCT GGT AAA ATR AGA ATA TAR-3'; amplicon length: 549). The ITS2 was amplified utilizing primers and PCR protocols described in Wełnicz et al. (2011) (ITS2, Forward: ITS3 5'-GCA TCG ATG AAG AAC GCA G-3', Reverse: ITS4 5'-AGT TTY TTT TCC TCC GCT TA-3'; amplicon length; 501). The amplified products were gel purified using the Wizard Gel and PCR cleaning (Promega, Madison, WI, USA) kit. Sequencing reactions were performed using the ABI Prism Big Dye Terminator v. 1.1 Sequencing Kit (Applied Biosystems™) on purified amplicons. Each sequencing reaction contained 0.2 μM of a single PCR primer to initiate the sequencing reaction, 2 μL of BigDye, 70 ng of purified products, 4 μL of 5x BigDye Terminator v.1.1 Sequencing Buffer and H 2 O for a final volume of 20 μL. Cycling conditions for sequencing reactions consisted of 25 cycles of 96°C for 10s, 50°C for 5s and 60°C for 4 min. Both strands were sequenced with ABI Prism 3100 (Applied Biosystems™). Nucleotide sequences of the newly analyzed specimens were submitted to GenBank, the accession numbers for cox1 of the four sequenced specimens (C4322 T1-T4) are OM370801-04, for ITS2 are OM402517-20.
The cox1 and ITS2 nucleotide sequences were aligned with the MAFFT algorithm (Katoh et al. 2002) as implemented in the MAFFT online service (Katoh et al. 2017) and checked by visual inspection. For cox1 sequences, chromatograms were checked for presence of ambiguous bases, as sequences were translated to amino acids by using the invertebrate mitochondrial code implemented in MEGA X (Kumar et al. 2018) to check for the presence of stop codons and therefore of pseudogenes. Sequences of other tardigrade sequences from GenBank belonging to Ramazzottius species were also included in the analysis for comparisons (Tab. S1 Supporting information). Pairwise nucleotide sequence divergences between sequences were calculated using p-distance with MEGA X for each gene.
Furthermore, relationships between cox1 and ITS2 were estimated using a parsimony network, by applying the method described in Templeton et al. (1992), as implemented in TCS ver. 1.21 (Clement et al. 2000) and visualized using tcsBU (Múrias Dos Santos et al. 2016). A 95% connection limit was employed, as it has been suggested as a useful general tool in species assignments and discovery (Hart & Sunday 2007). Putative species were also inferred by using the Poisson Tree Process (PTP; Zhang et al. 2013) and the Assemble Species by Automatic Partitioning method (ASAP; Puillandre et al. 2021). The PTP method produces robust species diversity estimates, and the starting gene trees were maximum likelihood (ML) trees computed using RAxML ver. 7.2.4 (Stamatakis 2006), as implemented in CIPRES (Miller et al. 2010), under the GTR+G model, as inferred by using the Akaike Information Criterion on jModelTest2 (Guindon & Gascuel 2003;Darriba et al. 2012) for both genes. Sequences of Hypsibius convergens (Urbanowicz, 1925) (GenBank accession number: FJ435798) and Hypsibius exemplaris  (GenBank accession number: MG800336) were used as outgroups for the genes cox1 and ITS2, respectively. Bootstrap resampling with 1000 replicates was undertaken via the rapid bootstrap procedure of Stamatakis et al. (2008) to assign support to branches in the ML tree. Bayesian trees were also computed using different models as inferred by MrModeltest ver. 2 (Nylander 2004). For the cox1 gene, the following models were utilized to consider the different evolutionary models for the three codons: SYM+I+G for the first position of the codon, GTR for the second position of the codon and GTR+G for the third position of the codon; while for the ITS2 gene the model HKY+G was utilized. The Bayesian dendrograms was computed with the program MrBayes ver. 3.2.7a (Huelsenbeck & Ronquist 2001;Ronquist & Huelsenbeck 2003), as implemented in CIPRES. Two independent runs, each of four Metropolis

348
R. Guidetti et al. coupled Markov chains Monte Carlo method, were launched for 3 × 10 7 generations, and trees were sampled every 1000 generations. Convergence of runs was assessed by tracking average standard deviation of split frequencies between runs and by plotting the log likelihood of sampled trees in TRACER ver. 1.7 (Rambaut et al. 2018) and the first 3 × 10 6 sampled generations were discarded as burn-in. In the distance-based ASAP method, the sequences are sorted into hypothetical species based on the barcode gap (i.e. whenever the divergence among organisms belonging to the same species is smaller than divergence among organisms from different species). The method first detects the barcode gap as the first significant gap beyond a model-based one-sided confidence limit for intraspecific divergence, and then uses it to produce several partitions of the data. The ASAP then computes an ad hoc ASAP-score for each defining partition, with the lower score indicating the better partition. The analysis was performed on the ASAP website (https://bioinfo.mnhn.fr/abi/public/ asap/).

Results
Comparisons of the new species with type specimens of several Ramazzottius species provided the opportunity to describe characteristics not reported in the original descriptions of those species and to obtain new photographs of several characters, not always present in the original papers. The new species description and the comparisons are provided below.
Ramazzottius kretschmanni sp. nov. (Figures 1-4 Cuénot, 1929. Etymology. The species is dedicated to Winfried Kretschmann, the political mastermind and founder of the Black Forest National Park. Description. (morphometric data in Table I, Supplementary Tab. S2): Body colour is reddish. Eye spots are absent. Elliptical sensory structures are present on the head and visible with LM ( Figures 1B, 3A,B). One small gibbosity is present on the external side of each leg of the fourth pair ( Figures 1J, 3G), not clearly detectable in all specimens. Entire surface of the body is smooth without visible ornamentation with both LM and SEM ( Figures 1L, 3A,B). With SEM, a "cheek-like" area is visible on each side of the head, at the level of the mouth opening ( Figure 3C). It is an oval area slightly raised above the body surface that shows a different cuticular pattern (i.e. a net of very small meshes, < 0.1 µm; Figure 3E, F) compared with the rest of the body cuticle ( Figure 3H). Within this "cheek-like" area there are: a dorsal region with several very small pores (more concentrated dorsally), a proximal cribrose area (for muscle attachment), and a ventral, almost rectangular, region with a few scattered, very small pores ( Figure 3E).
Feeding (bucco-pharyngeal) apparatus has a narrow buccal tube that is bent ventrally with slightly thicker walls located posteriorly to the stylet support insertion points ( Figure 1C, D). The buccal armature, visible only with SEM ( Figure 3D), is formed by a tiny anterior band of small teeth at the frontal extremity of the buccal tube (whose opening is rectangular in transverse section; Figure 3D) and a line of posterior teeth positioned in the anterior part of the buccal tube at the same level of the anterior part of the stylet sheaths ( Figure 3D). With SEM, it was possible to observe only the four dorsal teeth in a line, but symmetrical ventral teeth could be present, as in R. oberhaeuseri . Apophyses for the insertion of the stylet muscles on the buccal tube are asymmetrical and with the typical shapes for the genus: dorsal apophysis is shorter and stumpy, with the caudal apex clearly prominent ("blunt hook") ( Figure 1D); ventral apophysis has a less developed caudal apex. Stylet supports have an enlargement increasing from the proximal to the distal part ( Figure 1D). Each stylet furca has two wide spherical condyles laterally flattened and internally sclerified, supported by short branches with large apophyses. Pharynx has large triangular apophyses and two macroplacoids (the first is clearly longer than the second in larger specimens; Figure 1C -F). The shape and size of the placoids can change slightly between specimens. When the placoids are observed in lateral view ( Figure 1F), the first macroplacoid is grain-shaped (sub-spherical in smaller specimens; Figure 1D), while the second is sub-spherical; in dorsal view ( Figure 1E), the first macroplacoid is drop-shaped (a small median incision is visible in some specimens), while the second is rectangular with rounded corners and without incision.
Claws are of the Ramazzottius type (oberhaeuseri variant; according to Guidetti et al. 2019b) and moderately sized ( Figure 1G -I, 3F-G). Claws of the same leg are extremely different from one another in size and shape (Table I). The main (primary) branch of the external claws is straight and curved only distally, with small accessory points (difficult to see with LM) that run parallel and

350
R. Guidetti et al. coplanar to the branch ( Figures 1I, 3G). The primary branch is connected by a couple of thin cuticular filaments to the basal portion of the claw (that is continuous with the secondary branch) forming a non-sclerotized portion of the branch (the light refracting unit, LRU; Figure 1G -I). Length of branches increase slightly from the first to the fourth legs. The secondary branch of external claws is short and stumpy; it is inserted on a short basal portion and has evident accessory points ( Figure 1G -I, 3F-G). Pseudolunules are visible in the hind claws, although thin and barely visible; in the external claws they are extended towards the internal leg. Eggs are laid freely in the environment and have an ornamented shell (Figures 2, 4A,B). Eggs are circular or slightly oval, with a diameter without processes of 49.1-66.9 µm (mean 56.5 µm, SD 5.0 µm; N = 15). Egg shell has hemispherical processes (the size and appearance of which can vary between eggs; Figure 2 A-F), interspersed with few processes of irregular shape (e.g. resembling cones and truncated cones; Figure 2H -J, 4A). The heights and diameters of the hemispherical processes can vary between eggs (height: mean 3.8 µm, SD 0.5 µm, min 2.7 µm, max 4.8 µm; diameter: mean 7.9 µm, SD 1.3, min 4.6 µm, max 11.5 µm; N = 54 from 10 eggs). The process heights are generally lower or similar to half of the process diameters, with a mean percentage (ratio diameter/height) of 211.7% (SD 35.7;min 135.4,max 303.6; N = 54 from 10 eggs); two eggs, considered aberrant, show very flat processes (not measurable) with irregular margins ( Figure 2B, F). The thickness of the wall of the egg processes increases distally ( Figure 2G, K); in several eggs,

352
R. Guidetti et al. within the process wall, empty "bubble-like" spaces can be present ( Figure 2G), leading to an irregular appearance of the process surface with LM ( Figure 2C, E-F). There are about 15-22 processes (mean 18.2, SD 2.1; N = 15) on the egg circumference and on an egg surface of 1000 µm 2 is possible to count from 9 to 18 processes (mean 11.9, SD 2.4; N = 15). Surface of the processes is smooth. The base of the processes is round with an irregular indented margin visible only with LM ( Figure 2A -E); in some eggs, the indentations of the margin can be very long and evident ( Figure 2D). This indented margin develops below the shell surface (internally) and is visible with SEM only in broken processes ( Figure 4A). Egg shell surface between processes is generally smooth (Figure 2A), but some eggs show dots with LM ( Figure 2C, E) and small irregular crests with SEM ( Figure 4B). Most eggs were found in pairs within the sample; each pair was kept together   Guidetti et al. by connections between conical shaped processes and hemispherical processes ( Figure 2K -M). Three eggs with a fully developed embryo were found ( Figure 2N).

Morphological differential diagnosis
Ramazzottius kretschmanni sp. nov. is characterized by smooth cuticle, "cheek-like" area (described above), and egg shell with two types of processes (i.e. most hemispherical and some conical/truncoconical). Based on the claw morphology, Guidetti et al. (2019b) identified two groups of species within the genus Ramazzottius: the "oberhaeuseri group" characterized by claws of Ramazzottius type with oberhaeuseri variant (main branch connected to the secondary branch by two thin cuticular filaments, forming an LRU), and the "nivalis group" characterized by claws of the cataphractus variant (main branch detached from secondary branch). Based on egg morphology, Stec et al. (2018) identified the "oberhaeuseri complex" characterized by species with hemispherical egg processes. Accordingly, the new species, R. kretschmanni sp. nov., belongs to the "oberhaeuseri complex" based on egg morphology and to the "oberhaeuseri group" based on claw morphology. Within this group of species, the claw morphology is generally constant with differences among species only related to morphometric traits.
Ramazzottius kretschmanni sp. nov. differs from all the other species of the genus that have smooth cuticle or very weak dorsal posterior cuticular ornamentation. In particular, it differs from: Ramazzottius andreevi Biserov 1997/98 by having a completely smooth cuticle in all specimens (in R. andreveei the cuticle sculpture is poorly developed, completely smooth only in some specimens; Figure 5N) and shape of the egg processes (small and thin cones in R. andreveei; Figure 5P); Ramazzottius anomalus (Ramazzotti, 1962a) by the shape of the egg processes (long cone/aculeus in R. anomalus; Figure 6G), and egg surface smooth or with small dots (large granules in R. anomalus; Figure 6F); Ramazzottius caucasicus Biserov, 1997/98 by the shape of the egg processes (long cones with enlarged base in R. caucasicus; Figure 6K) and egg surface smooth or with small dots (smooth with scarcely distributed pores in R. caucasicus; Figure 6K); Ramazzottius montivagus (Dastych, 1983) by more slender main branch of external claws, with smaller accessory points, in the first three pair of legs (e.g. compare Figs. 5-7 in Dastych 1983 to Figure 1 H,J). Morphometric comparisons are difficult due to the absence of clear morphometric data for R. montivagus. The egg of R. montivagus are unknown; Ramazzottius oberhaeuseri (Doyère, 1840) (following the redescription by Stec et al. 2018) by evident elliptical organs on the head (poorly visible in R. oberhaeuseri), smooth cuticle (both with LM and SEM; R. oberhaeuseri shows a weak posterior polygonal sculpture), lower height of the egg processes (4.6-8.6 µm, mean 6.6 µm in R. oberhaeuseri), and higher percentage ratio of process diameter/height (97-197%, mean 141% in R. oberhaeuseri); Ramazzottius semisculptus Pilato & Rebecchi, 1992 (Figure 7) by completely smooth cuticle (a weak posterior polygonal sculpture can be present in R. semisculptus), shorter placoid row (pt 26.5-30.8 in R. semisculptus) and shape of the egg processes (conical in R. semisculptus); Ramazzottius subanomalus Biserov, 1985 by shape of the egg processes (conical in R. subanomalus; Figure 8D); Ramazzottius valaamis Biserov & Tumanov, 1993 by the shape of the egg processes (filamentous in R. valaamis; Figure 9F). A net-like sculpture, previously undescribed, was found on the cuticle of a paratype of R. valaamis ( Figure 9I), but is absent in the new species.
Ramazzottius kretschmanni sp. nov. differs from the other species of the genus that have hemispherical egg processes [i.e. R. affinis, Ramazzottius libycus Pilato, D'Urso & Lisi, 2013, R. oberhaeuseri, R. thulini] by the presence of smooth cuticle as all the other species have ornamented cuticle with hemispherical tubercles with a polygonal base. The cuticle of R. oberhaeuseri appears weakly ornamented with LM only in freshly mounted specimens or with SEM observations ). The eggs of R. kretschmanni sp. nov. differ from those of R. affinis (Figure 5 G,H) by the more regular shape of the processes (i.e. only few processes are not emispherical).

Molecular characterization
It was possible to amplify cox1 sequences from four specimens of R. kretschmanni sp. nov. (C4322 T1-T4), obtaining sequences of 549 bp, representing three different haplotypes with a p-distance of 0.2-0.4% (Figure 10; Tab. S1). The most similar haplotype (p-distance: 17.2%) to another species belongs to a population of Ramazzottius from Denmark (Tab. S1). The other available haplotypes from GenBank have p-distances ranging from 17.7% to 22.4% (Tab. S1) compared to that of R. kretschmanni sp. nov.
It was also possible to amplify ITS2 sequences of the same four specimens (C4322 T1-T4), obtaining sequences of 501 bp, representing four different haplotypes with a p-distance of 0.2-1.9% (Figure 11; Tab. S1). The most similar haplotype (p-distance: 2.3%) belongs to a population of Ramazzottius from Austria (Tab. S1). The other available haplotypes from GenBank have p-distances ranging from 3.1% to 21.6% (Tab. S1). In the comparison with R. oberhaeuseri (the type species of the genus), R. kretschmanni sp. nov. has p-distances of 18.9-19.1% for the cox1 sequences and 13.3-14.0% for the ITS2 sequences.  The validity of R. kretschmanni sp. nov. is further confirmed by the ASAP analysis of ITS2 gene (Figure 11, centre), whereas the haplotype network analysis (Figure 11, right) shows a further partition inside the German population, flagging specimen C4322 T2 as belonging to a different partition with respect of the other analysed specimens.

Observations and taxonomic considerations on
Ramazzottius species Ramazzottius affinis (Figure 5A-H) The original description (Bertolani et al. 1994) provided only drawings of the species. Being the species description in Italian, we report the description of

358
R. Guidetti et al. the main characters of the species as reported by Bertolani et al. (1994) and confirmed by our observations (for morphometric data see Bertolani et al. 1994). Eye spots are absent. Sculptured dorsal cuticle with 5-6 bands of small hemispherical tubercles (diameter 2.7-3.8 µm; Figure 5F), on the posterior two-thirds of the animal, alternated with thin smooth bands; the sculpture results absent in the anterior one-third or sometime in the first half of the animal. Two evident elliptical organs present on the head. A small papilla (not cited in the original description) presents on the external side of each leg on the fourth pair (in the holotype; Figure 5E). Transversal bands of epidermal cells with brown-reddish pigments (posteriorly corresponding to the bands of tubercles of the cuticle) alternated with not pigmented bands are present. Buccal ring without lamellae but, dorsally and ventrally, with a line of six very tiny teeth. Apophyses for the insertion of the stylets muscles (AISM) are asymmetrical respect to the frontal plane in shape of blunt hooks (dorsal crest thicker than the ventral; Figure 5A,B). Buccal tube with thicker walls after the stylet support insertion (not cited in the original description; Figure 5A). Pharyngeal bulb with evident apophyses, more developed transversally, and two granular macroplacoids (the first with a small indentation in the middle and the largest). Long claws of the Ramazzottius type, oberhaeuseri variant (according to Guidetti et al. 2019b), with not very evident accessory points, especially in the external claws, and with thin pseudolunula (Figure 5C,D). Eggs are free laid and ornamented with two type of processes: conical processes with relatively large base, the most abundant, and truncated-cone processes ( Figure 5G,H).

Ramazzottius andreevi (Figure 5I-P)
The original description (Biserov 1997/98) provided only drawings of the species, we provide LM pictures of all the taxonomic characters considered in the description of the species. The characters of the type specimens examined correspond to the original description. Based on our obeservations, the following

360
R. Guidetti et al. characters can be added to the original description as: the presence of clearly visible cuticular sculpture, with polygonal flat tubercles ( Figure 5N) and visible elliptical organs on the head ( Figure 5O) (according to Biserov (1997/98) both characters are inconspicuous and/or not visible); an increase of the thickness of the buccal tube wall after the stylet support insertion point ( Figure 5L,M); a light refracting unit (LRU) in the main branch of external claw on all legs ( Figure 5J).

Ramazzottius anomalus (Figure 6A-G)
The original description (Ramazzotti 1962a) provided only drawings of the species, we provide LM Results of the Poisson tree process analysis are provided using differently coloured branches: putative species are indicated using transitions from blue-coloured branches to red-coloured branches. Newly scored haplotypes are in bold. The scale bar shows the number of substitutions per nucleotide position. Centre: rectangles denote specimens grouped by ASAP analysis (asap-score: 3.00). Right: haplotype network analysis. Circles represent haplotypes, while circle surface denotes haplotype frequency. Networks falling below the value of the 95% connection limit are disconnected.
photographs of the bucco-pharyngeal apparatus, claws and eggs of the species. Some photographs of the type series were published in Stec et al. (2017). We report photographs of the dorsal and ventral crests of the AIMS in lateral view ( Figure 6A) never represented before. The characters of the type specimen examined correspond to the original description. To avoid future misunderstanding, we specify that the original description of the species is not in Ramazzotti (1962b), as reported by some authors (e.g. Kaczmarek et al. 2015;Stec et al. 2017), but in Ramazzotti (1962a).

Ramazzottius caucasicus (Figure 6H-N)
The original description (Biserov 1997/98) provided drawings of the species and four SEM pictures of a claw, the cuticle and eggs, we provide LM photographs of all the taxonomic characters considered in the description of the species, including the pseudolunules in the external claws of second and third pair of legs and in the posterior claws of the hind legs ( Figure 6M,N), considered by Biserov (1997/98) poorly visible. The characters of the type specimen examined correspond to the original description. We want to emphasise the presence of an increase of the thickness of the buccal tube wall after the stylet support insertion point ( Figure 6I), and a LRU in the main branch of external claw on all legs ( Figure 6L,M), not evidenced in the original description. Séméria, 1993 Due to the poor original description (Séméria 1993) of R. edmondabouti, it was not possible to verify the status of most of the characters useful to discriminate this species from the other species in the genus. For this reason and because of the unknown egg morphology of this species, we propose to assign to R. edmondabouti the status of species dubia, pending analyses of the type specimens (available at the Natural History Museum of Nice, France; Séméria et al. 2018).

Ramazzottius rupeus (Figure 7A-F)
The original description (Biserov 1999) provided drawings of the species and four LM photographs of the cuticle, claws and an egg. We provide new LM photographs of most taxonomic characters Figure 11. Left: tree resulting from both the Bayesian inference and the maximum likelihood analysis of ITS2 in Ramazzottius kretschmanni sp. nov. specimens and sequences from GenBank. Values above branches denote posterior probability values, while values under branches represent bootstrap values. Results of the Poisson tree process analysis are provided using differently coloured branches: putative species are indicated using transitions from blue-coloured branches to red-coloured branches. Newly scored haplotypes are in bold. The scale bar shows the number of substitutions per nucleotide position. Centre: rectangles denote specimens grouped by ASAP analysis (asap-score: 1.50). Right: haplotype network analysis. Circles represent haplotypes, while circle surface denotes haplotype frequency. Networks falling below the value of the 95% connection limit are disconnected.

362
R. Guidetti et al. considered in the species description derived from specimens different to those photographed by Biserov (1999). The characters of the type specimen examined correspond to the original description. We want to emphasise the presence a LRU in the main branch of external claw on all legs ( Figure 7B), not evidenced in the original description.

Ramazzottius semisculptus (Figure 7G-K)
The original description (Pilato & Rebecchi 1992) provided only drawings of the species, we provide LM photographs of some taxonomic characters considered in the species description. The characters of the type specimen examined correspond to the original description. We want to emphasise the presence a LRU in the main branch of external claw on all legs ( Figure 7I-K), not evidenced in the original description.

Ramazzottius subanomalus (Figure 8A-E)
The original description (Biserov 1985) provided drawings of the species and three LM photographs of the animals and an egg. We provide LM photographs of some taxonomic characters considered in the species description from type specimens already analysed by Stec et al. (2017) in the redescription of the species. We want to emphasise the presence a LRU in the main branch of the external claw on all legs ( Figure 8B), not evidenced in previous descriptions (Biserov 1985;Stec et al. 2017).

Ramazzottius thulini (Figure 8F-J)
The original description (Pilato 1970) provided only drawings of the species, while some pictures of the type specimens were published in Pilato et al. (2013). We provide new LM photographs of the holotype, in toto, and of its bucco-pharyngeal apparatus and claws. The characters of the type specimen examined correspond to the original description. We want to emphasise the presence of an increase of the thickness of the buccal tube wall after the stylet support insertion point ( Figure 8G), and a LRU in the main branch of external claw of all legs ( Figure 8I,J), not evidenced in the original description.

Ramazzottius tribulosus (Figure 9A-D)
The original description ) provided drawings of the species, one LM photograph of the egg, and one SEM picture of mouth opening. We provide new LM photographs of the bucco-pharyngeal apparatus, claws and an egg. The characters of the type specimen examined correspond to the original description. We want to emphasise the presence of a LRU in the main branch of external claw on all legs ( Figure 9C,D), not evidenced in the original description.

Ramazzottius valaamis (Figure 9E-I)
The original description (Biserov & Tumanov 1993) provided drawings of the species and four LM photographs of the animals, feeding apparatus and eggs. We provide new LM photographs of the cuticle, bucco-pharyngeal apparatus, claws, and an egg ( Figure 9E-I). We want to emphasise that contrary to the original description, a net-like sculpture is visible in the dorsal posterior cuticle ( Figure 9I) in a paratype, and a LRU is present in the main branch of the external claw on all legs ( Figure 9G).

Discussion
The discovery of a new species of Ramazzottius, observations of type materials hosted in public collections, and a review of the literature related to this genus led to an analysis of morphological characters within the genus and the comparison with other eutardigrades in an effort to understand their characteristics, distribution, and evolution.

Characteristics of the dorsal posterior cuticle
Ramazzottius kretschmanni sp. nov. has a smooth cuticle, but in most Ramazzottius species, the dorsal cuticle is characterized by small or large "bulges" (i.e. gibbosities, protuberances, tubercles), or even spines as in Ramazzottius belubellus Bartels, Nelson, Kaczmarek & Michalczyk, 2011b, or large hemispherical tubercles as in R. saltensis and R. szeptycki (which are absent in all the other species of the genus). When present, these "bulges" are always larger and more evident posteriorly. Similar "bulges" in the dorsal posterior portion of the cuticle are present in many other species of different genera of Parachela (Eutardigrada) belonging to different evolutionary lineages and living in different environments and habitats. In these species, the posterior-dorsal cuticle (i.e. generally after the third pair of legs) is characterized by one of the following types of "bulges": gibbosities, granulations, tubercles, dots, crests, spines, outgrowths, and wrinkles. When these "bulges" are present in a more anterior-dorsal position, they are always reduced and/or less evident.
A similar unknown selective pressure probably acts not only in Parachela, but also in limnoterrestrial Heterotardigrada, resulting in a similar phenomenon. In the heterotardigrades without cuticular plates, as e.g. Orella mollis Murray, 1910, there are posterior small gibbosities, while in the Echiniscidae with cuticular dorsal plates, the dorsal spines and/or filaments are in many cases present only on the posterior edge of the dorsal-posterior plates, and when other dorsal spines are present (except for the anterior sensory cirri), they are often smaller in size (e.g. see Guil 2008).
The possible selective pressure that led to this convergent evolution is unknown. Guidetti et al. (2019b) hypothesized that the dorsal-posterior granules in Crenubiotus species increase the animal's adhesion to the substrate, similar to the dot-like structures present on the legs of many Macrobiotoidea that very probably increase the grip of the leg on the substrate, but further data are needed to test this hypothesis. In tardigrades the ventral side of the body is always smooth and species of the same genus, living in similar habitats, can have different cuticular characteristics (e.g. the new species here described and other species of Ramazzottius have a smooth cuticle, although most species of the genus have an ornamented posterior cuticle). Other hypotheses to explain the phenomenon described could encompass cuticle permeability, sensory structures, hydrodynamics of the body, defence mechanisms, or it may not even be an adaptive trait and be caused by a non-adaptive developmental model.
Understanding the origin and function of these "bulges" will require more accurate phylogenetic analyses of genera and families. For example, our phylogenetic analysis for the cox1 gene shows Ramazzottius kretschmanni sp. nov. in a basal position with respect of all other available Ramazzottius species (Figure 10), suggesting that a smooth cuticle could be the ancestral state. Given that this situation is not confirmed in the analysis of the ITS2 gene ( Figure 11), a more accurate (molecular) phylogenetic analysis is required.

Head sensory regions
Ramazzottius kretschmanni sp. nov. has at least three sensory regions on the surface of the head, two of have the cup-cell at all). This could explain the presence or absence of an eye spot even within a single specimen (Bąkowski et al. 2016). This hypothesis should be investigated in detail to evaluate the evolution of tardigrade vision and to determine the value of eye spots as a taxonomic trait.

Morphology of egg processes
The main morphological characters used to discriminate among Ramazzottius species are features related to egg morphology and dorsal cuticle ornamentation. Therefore, intraspecific variability in the egg processes can be an issue for species discrimination and identification as evidenced for R. subanomalus which can present a high variability of egg process morphology (Stec et al. , 2017. The egg of R. kretschmanni sp. nov. is characterized by two types of processes: "hemispherical" (the most abundant) and "conical" (few and with variable shapes). In species of the "oberhaeuseri complex" (see above), the "hemispherical" egg processes maintain their shape, although they can differ in size and appearance on egg surface (Figure 2A Pilato et al. 2013); while the shape of the "conical" process can vary greatly within each egg but are substantially similar between the species (see Figure 2H-J and Pilato et al. 2013). In R. kretschmanni sp. nov., the "conical" processes are responsible for the connection between two different eggs ( Figure 2K-M), and their shape seems to be determined by the nature of this connection, e.g. the process can be pulled/ stretched as in Figure 2K,L, or pushed/deformed as in Figure 2M. One hypothesis is that the "conical" processes are randomly present on the egg surface to increase the chance of touching and then connecting to another egg, and/or that they derived from "hemispherical" processes that have been deformed by the adhesion to other processes of a different egg. More information is needed to evaluate the true nature of the shape and numbers of the "conical" processes and their taxonomic value.