Sample collection, rearing conditions, and microscopy

The ice-free area of Innhovde is situated at ca. 120 km southwest of Syowa Station (Figs. 1, 2). Along the north-eastern slope near the coast are exposed rocks, about 3 km long from the NW end (ca. 10 m above sea level) toward the SE end (160 m a.s.l.). We conducted a field study in the Innhovde area on 11 January 2015. Limited vegetation occurred in the northern quarter of the area; in contrast, the rest of the southern area was relatively dry with little vegetation, and lakes were still frozen at the time of our visit. Mosses and lichens from the northern area were collected in paper bags and kept dried in cold storage, at 4°C during the expedition, then later at –20°C in a cold room at National Institute of Polar Research (NIPR), Tokyo.

Several samples were examined during the expedition in our field laboratory at Kizahashi-hama, Skarvsnes (69°28′22″S, 39°36′06″E). The samples were divided into several portions, one of which was immediately soaked in Milli Q water and the small animals that were liberated were picked out under a stereo microscope (Wild M3C, Leica). Two adult females and a juvenile of the Milnesium species were found in the moss Ceratodon purpureus. The adult specimens were mounted on glass slides in gum-chloral solution, and the juvenile was fixed and stored in ethanol for DNA analysis. Images of the slide specimens were taken at the field laboratory with a digital camera (NEX5-N, Sony) on a compound microscope (Nikon L-Ke) with differential interference contrast (DIC), and again later using another microscope (BX50, Olympus) with DIC. The rest of the samples were returned to Japan, where the remainder of the moss sample that had yielded the Milnesium specimens at the field laboratory was examined. Additional specimens comprising an exuvia with three eggs, an adult female, a premature simplex individual, and a moulting male were extracted. The premature simplex specimen was fixed with ethanol for DNA analysis, while the moulting male and three hatchlings from the eggs were mounted on microslides. The adult female, which was presumed to be the mother of the three hatchlings, was reared in a plastic dish and the life history was observed under a stereo microscope (Wild/Leica M-10). The rearing conditions were as described in Suzuki (2003) with a slight modification, using rotifers Lecane inermis (Bryce, 1892) as a food source, and at 4 or 10°C under 16L:8D photoperiod. After this female individual died, the specimen was processed for scanning electron microscopy (SEM). It was fixed in 10% formalin, post-fixed in 1% OsO4 in 0.1 M sodium cacodylate, pH 7.2 for 1 hr at room temperature, dehydrated through a graded series of ethanol and 100% t-butyl alcohol, frozen at 4°C overnight, and lyophilized using a JFD-320 (JEOL). The specimen was mounted on an aluminium stub, sputter-coated with gold, and observed with a JSM 6510 (JEOL) at 20 kV.

Fig. 1.

Aerial view of Innhovde on 11 January 2015, from NW to SE direction.

Fig. 2.

Maps showing Innhovde. Based on maps by Geospatial Information Authority of Japan ( https://www.gsi.go.jp/antarctic/02.html) and modified.

Specimens for morphometry were observed under an Olympus BX-50 DIC microscope with a digital camera, and measurements were taken using ImageJ ( https://imagej.nih.gov/ij/). Structures were measured when their orientation was suitable. Body length was measured from the anterior extremity to the end of the body excluding the hind legs. Buccal tube length and stylet support insertion point were measured according to Tumanov (2006). Buccal tube width (diameter) was measured at the anterior, standard, and posterior positions (Michalczyk et al., 2012). The pt index (Pilato, 1981) was used for comparison, i.e., the percentage of the length of each structure relative to the buccal tube length, expressed in italics. Illustrations were made using a drawing tube attached to a BX-50 microscope and finished with Inkscape software ( https://inkscape.org/). The terminology for the Milnesium claw system generally followed Camarda et al. (2022), i.e., we now replace the terms “primary/secondary branches” in the previous works (reviewed in Suzuki, 2022) by “primary/secondary claws”, and instead we call the hooks, or points, on the secondary claws “branches”. Following Suzuki (2022), we avoid using the directional terms internal/external for leg I–III, and describe the CC by the revised notation system (Suzuki, 2022) with a modification, i.e., the number of “branches” on the “secondary claws” expressed in brackets as {anterior/posterior} in order through all legs. The position of legs was indicated by a superscript Roman numeral behind the brackets, if necessary. By this notation, the adult CC of M. tardigradum is expressed simply as {3-2}, or more precisely {3-2}^I–IV. Specimens of the new species in the present study have a different number of branches, even on the left/right legs in one individual, so that every CC was described for each leg.

Other samples of vegetation collected from Antarctica during JARE-56 have been kept at NIPR for future investigation.

DNA extraction and sequencing

DNA extraction was performed as described by Kagoshima et al. (2013). Each specimen was transferred individually to 20 µL of 0.25 N NaOH in a 0.2 mL tube and kept at room temperature for 12 hr. The tube was heated for 3 min at 95°C, and 4 µL of 1 M HCl and 10 µL of 0.5 M Tris-HCl (pH 8.0) were added to neutralize the base followed by 1 µL of 2% Triton X-100. The lysate was heated for a further 3 min at 95°C and stored at – 20°C. Polymerase chain reaction (PCR) amplification was performed with primers as follows: for the 18S rRNA gene we used the primers SSU04F and SSU81R (Blaxter et al., 1998), for the 28S rRNA gene the primers were 28S_Eutar_F (Gąsiorek et al., 2017) and 28SR0990 (Mironov et al., 2012), for ITS-2 they were ITS3 and ITS4 (White et al., 1990), and for COI the primers were LCO1490 and HCO2198 (Folmer et al., 1994). PCR conditions for the 18S rRNA gene and COI were: 94°C for 2 min, followed by 40 cycles of 94°C for 10 sec, 52°C for 30 sec, and 72°C for 10 min. Conditions for the 28S rRNA genes and ITS-2 were: 94°C for 2 min, 40 cycles of denaturing at 94°C for 30 sec, annealing at 50°C for 30 sec, extension at 68°C for 75 sec, and final extension for 7 min. Sequencing reactions were performed with Big-Dye terminator cycle sequencing kits and run on the ABI 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA) for 18S rRNA gene and COI, and sequencing for the other two regions was performed by Eurofins Genomics (Japan). The sequences were checked and trimmed with SnapGene Viewer, a software from Insightful Science (snapgene.com).

Phylogenetic analyses

For phylogenetic analyses, we used the publicly available Milnesium spp. sequences of 18S rRNA, 28S rRNA, ITS-2, and COI in the DDBJ/ENA/GenBank databases as shown in  Supplementary Table S1 (zs220085_TableS1.xlsx). Antarctic taxa and some species with published sequences were added to the list. The sequences were aligned using MAFFT v7.222 (Katoh et al., 2002; Katoh and Toh, 2008) with the default setting, then obtained alignments were checked by using MEGA7 (Kumar et al., 2016) and trimmed manually to 193 bp in 18S rRNA, 564 bp in 28S rRNA, 104 bp in ITS-2, and 478 bp in COI. The p-distances were calculated with MEGA7.

Phylogenetic analyses were run with COI and 18S + 28S + ITS-2 + COI concatenated sequences. The concatenated sequences included only the taxa with the COI sequence and at least one nuclear DNA sequence. Four Antarctic COI sequences: KJ857001, KJ857002, KP013598, and KP013613, were added for the concatenated sequences even though the taxa with only COI sequence was available. The best-fit evolution model and Maximum Likelihood (ML) topologies were calculated with IQ-TREE v1.6.12 (Minh et al., 2020) by 10,000 rapid bootstrap replicates. TVM+F+G4, GTR+F+I+G4, K3Pu+F+G4 and HKY+F+R7 were estimated to ITS-2, 1st, 2nd and 3rd codon of COI, respectively. In addition, K2P+G4 and T2P+R5 were estimated to the respective partitions for 18S and 28S of the concatenated sequence. Bayesian Inference (BI) probabilities were calculated with MrBayes v3.2.6 (Ronquist and Huelsenbeck, 2003; Ronquist et al., 2012) under GTR+I+G model with “nst = 6 rates = invgamma” option, and the analyses were run for over 5,000,000 generations, sampling the Markov chain every 1,000 generations. Moreover, an average standard deviation of split frequencies of < 0.01 was confirmed. The trees were visualised in FigTree v1.4.3 ( http://tree.bio.ed.ac.uk/software/figtree).

Genetic species delimitation

Genetic species delimitation was performed at the bPTP server ( https://species.h-its.org/ptp) with the following settings: 100,000 MCMC generations, 100 thinning, 0.1 burn- in via Poisson Tree Processes (PTP) model, excluding the outgroups (Zhang et al., 2013).

Taxonomic account

Fig. 3.

Image of live specimen of Milnesium rastrum sp. nov. under the stereomicroscope. Adult female (Female-3). (A) ‘Tun’ on the nylon mesh. (B) Body after 7 minutes of rehydration. (C) Expanded body, 15 minutes after rehydration. (D) Revived animal walking on agar, 38 minutes after rehydration. Scales, 100 µm.

Fig. 4.

Adult female of Milnesium rastrum sp. nov. (A) Holotype habitus. (B) Claws on each leg. The Roman numeral indicates the position of the leg, with the superscript showing right or left.

Fig. 5.

Adult female of Milnesium rastrum sp. nov. (A) Female-2 habitus. (B) Claws on each leg. The illustration of the leg II^L is missing due to its unfavorable position.

Fig. 6.

Female of Milnesium rastrum sp. nov. (A) Head region showing the buccal apparatus, peribuccal structures and eyes. Female-2. Scale, 20 µm. (B) Claws of right leg II of holotype, {5-6}^IIR CC. Scale, 10 µm. (C) Anterior secondary claws of right leg IV of female-2, the secondary claw with seven branches. Scale, 10 µm. (D) Claws of right leg II, female-3, SEM. Scale, 5 µm.

Fig. 7.

Moulting male of Milnesium rastrum sp. nov. (A) Habitus. Scale, 100 µm. (B) Anterior half. Scale, 20 µm. (C) Right leg I. The juvenile claws at the top, with {5-5} CC. Arrow, the new male claw I under the old cuticle. Scale, 10 µm. (D) Left leg I, the juvenile claw (top) and the robust male claw (arrow). Scale, 10 µm.

Fig. 8.

Hatchlings of Milnesium rastrum sp. nov. Habitus of hatchling-1 (A), hatchling-2 (B), and hatchling-3 (C). (D) Right leg II of hatch-ling-1, with {5-5} CC. Scale bar, 10 µm. (E) Leg IV of hatchling-3 with {5-5} CC. Scale bar, 10 µm.

Fig. 9.

Exuviae of Female-3. (A) Posterior half of the exuviae, including the trophi of rotifer in the hind gut. (B) Enlargement showing the trophi.

Observation of a solitary adult female in culture

When the exuviae including three eggs was found, an adult female (female-3) was also retrieved from the same sample. This female, which might have been the mother of the above three eggs in spite of no firm evidence, was temporarily dried and kept in a refrigerator. After a week it was rehydrated (Fig. 3A–D). The dried body was 244 µm long and expanded to 707 µm after rehydration. Once rehydrated it was put into a 35-mm dish filled with water and provided with rotifers as the food source. Table 2 shows the summary of the observations of female-3 for more than 3 months. The movement of the female was very slow, and its eating behavior was not directly observed. However, its gut sometimes looked full of some grey substance.

At day 19 after rehydration, the female lost her buccal apparatus going into simplex (i.e., the moulting stage), but no distinct eggs were discerned in her ovary. It took about 10 days at 10°C to form new buccal apparatus (day 29). Two days later (day 31), the locomotive activity ceased, and apolysis, i.e., detachment of the old cuticle from the newly formed epidermis, was observed. It took another week to finish the ecdysis (day 38). The remains of a rotifer trophi in the exuviae between the 3rd-leg pair (in the old cuticle of the hind gut) provided evidence that this female had been eating the rotifers (Fig. 9).

The new-instar female continued her slow locomotion for a month until day 68, and then the next moulting began, indicated by the loss of the buccal tube again. The ovary, as before, had no large oocytes. It took 14 days to recognize the beginning of the buccal tube regeneration. At day 86, the observation was interrupted for 10 days at 4°C (caused by a conference trip of the first author). By day 96, when the animal was returned to 10°C, the new buccal tube had already formed. However, she did not complete ecdysis and apparently died at day 103. The body was fixed and processed for SEM observation (Fig. 6D). Overall during this observation, this female moulted twice without egg laying.

Fig. 10.

Maximum likelihood phylogenetic trees based on COI. (A–C) indicate the expanded tree corresponding to the legend. The number on each node indicates bootstrap value (only above 90) and Bayesian posterior probability (only above 0.9). Symbols indicate the distribution of each taxon explained in the legend. N.B. M. pacificum was not attributed clearly to a specific locality, requiring two symbols. Scale bar indicates substitutions per position.

Continued.

Continued.

Genetic species delimitation

The DNA sequences of Milnesium rastrum sp. nov. were deposited in the DDBJ with the following accession numbers: LC721285 (18S rRNA), LC721286 (28S rRNA), LC721288 (ITS-2), and LC721287 (COI). A summary of genetic p-distances is shown as follows:

18S rRNA: 0.0%–1.9%, with the most similar species being Milnesium sp. CJS-2007a MilnC 010 (EF632492), Milnesium sp. CJS-2007a MilnC 025 (EF632493), Milnesium sp. Miln06 108 (EU266922), and Milnesium sp. Miln05 141 (EU266923) from Antarctica, and the least similar being M. dornensis Ciobanu et al., 2015 RO.008 (MK484071) from Romania.

28S rRNA: 5.1–9.5%, with the most similar species being Milnesium sp. UG.006 from Uganda (MK484006), and the least being Milnesium sp. AU.080 from Australia (MK483992).

ITS-2: 15.3–44.4%, with the most similar species being Milnesium spp. including PH.014 from the Philippines (MK484029), AR.470 from Argentina (MW538052), GF.196 from French Guiana (MW538055), ID.432, ID.711, ID.940 and ID.950 from Indonesia (MW538057, MW538058, MW538060 and MW538062, respectively); and the least similar species being Milnesium sp. ZA.087 from South Africa (MW538083).

COI: 18.0–32.2%, with the most similar species being Milnesium sp. Ta47-01 from Antarctica (KJ857002), and with the least similar species being M. berladnicorum Ciobanu et al., 2014 ZA.040 from South Africa (MW560657).

Phylogeny

The phylogenetic trees based on the COI and concatenated sequences are shown in Fig. 10 and  Supplementary Figure S1 (zs220085_FigS1.pdf), respectively; the supported values of BI trees were usually low, thus the phylogenetic trees based on the ML trees are shown. The COI ML tree indicates that M. rastrum sp. nov. is a separate species with 100% probability by both maximum likelihood and Bayesian supported solutions, and the new species is included in a clade with Antarctic specimens (Milnesium spp. Miln06 224, MilnC 025, Ta47-01, and Ta46-01).

Differential diagnosis

Milnesium rastrum sp. nov. is clearly distinguished from all other congeners by having four–seven branches on the secondary claws, which CC is probably expressed as {(4–7)-(4–7)}. Milnesium quadrifidum has {4-4} CC, M. wrightae has {3-3}{4-4}^IV with a minute 4th branch near the base of the secondary claw, and all other reported species have either CC of {3-3}, {3-2}, or {2-2}, or a mixture of these combinations (Suzuki, 2022).

Two species have already been described from Antarctica: Milnesium antarcticum Tumanov, 2006 from King George Island and M. validum Pilato et al., 2017 from Victoria Land, both have {3-3} CC. In addition to the CC difference, M. rastrum sp. nov. also differs from:

Milnesium antarcticum by having a shorter buccal tube for a similar body length (range: 62.5–64.1 µm in the new species and 67.0–74.7 µm in M. antarcticum); the adult body length for M. antarcticum was reported to exceed 800 µm (Tumanov, 2006) and thus may be similar to that of the new species; and by having a more cephalic position of stylet support insertion point (range: [62.0–64.2] in the new species and [70.0–73.7] in M. antarcticum).

Milnesium validum, by having a shorter buccal tube with respect to the body length (percent ratio 6.9–7.9 in the new species and 10.4–11.5 in M. validum).

Other records of Antarctic Milnesium specimens

The distribution of Milnesium spp. recorded so far in and around Antarctica is shown in Fig. 11. The morphological trait of multiple branches observed in the new species from Innhovde has already been reported from other areas within the East Antarctic region. The specimens from Langhovde (Sudzuki, 1964) showed two–five branches on the secondary claws. Although the quality of the illustrations was far from perfect, this description suggested that the samples possibly included specimens with up to five-branched secondary claws. Moreover, although this report did not mention any males, one of the illustrations clearly showed the modified secondary claw of a male (see fig. 12 in Sudzuki, 1964). This indicates that a dioecious Milnesium population also occurred at Langhovde, and it is possible these belong to the new species. Co-localization of other Milnesium sp. with {3-3} CC has been suggested by the literature that described M. tardigradum sensu lato from Molodeznaya, in the vicinity of Lützow-Holm Bay (Utsugi and Ohyama, 1991).

Dastych (1984) compared specimens from King George Island and Evening Mountain (67°39′S, 46°06′E), near Molodeznaya Station, both recorded as Milnesium tardigradum sensu lato, describing them as follows: “Specimens from King George Island are typical [...], those from Antarctic Continent belong to the form [...] having variable middle branch of claw. That branch is single in typical specimens […], in mentioned form it is divided into 2–5 teeth, mostly 3–4; […]. A number of teeth is different within this species, often even on one leg.” Dastych's (1984) description of the ‘typical’ three-branched specimens from King George Island would correspond with the more recent description of M. antarcticum, also from King George Island (Tumanov, 2006). The Continental Antarctic ‘form’ that Dastych (1984) described with four–seven branches on the secondary claws, mostly five–six branches, would match our description for M. rastrum sp. nov.

Table 1.

Measurements of Milnesium rastrum sp. nov.

A ‘form’ of ‘Milnesium tardigradum’ was also described from two nunataks, Baileyranten (73°33.9′S, 14°33.8′W, 1141 m a.s.l.) and Engenhovet (74°34′S, 11°01′W, 1757 m a.s.l.) in Heimefrontfjella, Dronning Maud Land (Sohlenius et al., 1996). The paper stated, “Milnesium tardigradum is the form already reported from Antarctica (H. Dastych, personal communication).” Although this statement implied a ‘form’ (Dastych, 1984), it is not clear as to which ‘form’ the authors were referring.

The COI sequence of Milnesium sp. Ta46-01 was obtained from a specimen collected from the Stornes Penisula, Larsemann Hills (Velasco-Castrillón et al., 2015). Our unpublished data indicates there is a population of Milnesium in the Larsemann Hills with multiple secondary claw branches.

Table 2.

Life history of an adult female of Milnesium rastrum sp. nov. (female-3).

Another study noted the distribution of Milnesium sp. from the inland nunataks at the base of the Antarctic Peninsula, Palmer, and Ellsworth Land, and suggested that morphology indicated, “there may be distinct ‘forms’ or speciation occurring at the different sites within the continental, maritime, and sub-Antarctic” (Convery and McInnes, 2005). The details of morphological differences were not stated. However, the COI sequence of Milnesium sp. Miln06-224 (in Velasco-Castrillón et al., 2015) was obtained from Ellsworth Land, the region under discussion in Convery and McInnes (2005). There is currently no information about CC of Ta47-01 from Queen Maud Mountain.

Phylogeography

In the ML phylogenetic tree constructed from Milnesium COI, M. rastrum sp. nov. appeared in a clade with Antarctic strains: Milnesium sp. Ta47-01 from the Queen Maud Mountains, Milnesium sp., MilnC 025 from Charcot Island, Milnesium sp. Miln06 224 from the Ellsworth Mountains, and Milnesium sp. Ta46-01 from Stornes Peninsula, Larsemann Hills. The most basal specimens were apparently Milnesium sp. Miln06 123 and Miln06 124, both from Marion Island, followed by Milnesium spp. from South Africa (ZA.436, ZA.015 and ZA.367.01), M. matheusi Kaczmarek et al., 2019 from Madagascar, and Milnesium sp. from Australia (AU.080 and AU.105). Afrotropical strains are also found among many other clades in the phylogenetic trees. In comparison, the Nearctic/Palaearctic strains appeared in a more derived position, after the seventh or eighth node (Fig. 10B, C), and are supported by more studies from these regions. This view of the tree accords with the Gondwana origin hypothesis of several lines of tardigrades (McInnes and Pugh, 1998; Guidetti et al., 2017).

It is interesting that Marion Island harbored the most basal Milnesium strain. This island is the tip of an active oceanic intraplate volcano, considered to be less than one million years old and with an earliest subaerial eruption estimated by K–Ar dating at ca. 450,000 years ago (McDougall et al., 2001). On the other hand, the split between Parachela and Apochela is estimated to be at 432 million years ago (Mya) (range, 323–540 Mya) during the Palaeozoic, and the most recent common ancestor of Milnesium was estimated to appear 162 Mya (range, 116–207 Mya) (Morek et al., 2021), which would coincide with the onset of the Gondwana breakup about 164 Mya (Mueller and Jokat, 2019). Thus, the island is quite young in comparison with Milnesium diversification. However, Marion Island is on the Marion Hotspot that has been stationary since 88 Mya and played an active role in the breakup of Madagascar and India (Storey et al., 1995; Torsvik et al., 1998). According to this geographic view, Madagascar would have migrated North, with its Milnesium ancestor, away from the Marion Hotspot about 88 Mya. This Milnesium population would have evolved in response to various environments, with the extant Madagascan species M. matheusi and M. wrightae, as the modern representatives. However, this does not explain how Milnesium spp. arrived on Marion Island. We cannot say where the ancestral Milnesium species existed until we get more spatiotemporal data and hopefully some fossil records. Nevertheless, with the molecular results of Figs. 10 and 11, it is possible to speculate that the vicinity of the Marion Hotspot might have been the center of the explosion of Milnesium diversity. If so, the ancestors of extant species would have dispersed from about 164 Mya, on the breakup of the Continents, West-Gondwana (Africa and South America) and East-Gondwana (Indo- Madagascar, Antarctica and Australia). Further sampling is required to prove or disprove this hypothesis.

Fig. 11.

Localities of Milnesium spp. found in and around Antarctica. 1, King George Island; M. antarcticum (Tumanov, 2006). 2, Charcot Island; MilnC-010 and MilnC-025 (Sands et al., 2008a). 3, Ellsworth Mountains; Miln06-224 (Convey and McInnes, 2005; Sands et al., 2008b). 4, Heimefrontfjella; Milnesium sp. (Sohlenius et al., 1996). 5, Innhovde; M. rastrum sp. nov. 6, Langhovde; Milnesium sp. (Sudzuki, 1964). 7, Molodeznaya; possible M. rastrum and Milnesium sp. with {3-3} CC (Dastych, 1984; Utsugi and Ohyama, 1991). 8, Larsemann Hills (Stornes Peninsula); Milnesium sp. Ta46-01 (Velasco-Castrillón et al., 2015). 9, Victoria Land (Carezza Lake); M. validum (Pilato et al., 2017). 10, Queen Maud Mountains; Milnesium sp. Ta47-1 (Velasco-Castrillón et al., 2015). 11, Signy Island, South Orkney Islands; Milnesium sp. with {3-3} CC (McInnes, 1995). 12, South Georgia; Milnesium sp. Miln06-108 (Sands et al., 2008b). 13, Marion Island, Prince Edward islands; Milnesium sp. Miln06-123 and Miln06-124 (Sands et al., 2008b; Velasco-Castrillón et al., 2015). 14, Kerguelen Islands; Milnesium sp. with “vierten Beinpaar 3/4 Krallen”, i.e., {3-4}^IV or {4-3}^IV (Richters, 1908). 15, Macquarie Island; Milnesium sp. with {3-3} CC (Miller et al., 2001). The distribution map was created using an Antarctic map (Map ID 13469, Australian Antarctic Data Centre) with a modification.

On the signification of male Milnesium

The male form of Milnesium rastrum sp. nov. has a robust secondary claw on leg I. This characteristic claw shape was noted by Murray (1907) and Richters (1908), and then described in detail as a male-specific character (Thulin, 1928). Generally, Milnesium males were considered rare, with Marcus (1936) describing a single male for 25 females (Marcus, 1936), and Ramazzotti (1972) thought that there were fewer. A detailed study of male/total numbers in several populations of ‘Milnesium tardigradum’ revealed: 18/75 from Sardinia, 9/26 from Buffalo Mountain Park, TN, U.S.A., 29/76 from Land Between the Lakes, TN, U.S.A., and 2/38 from New Zealand (Rebecchi and Nelson, 1998). Previously, males have been recorded for 12 species of Milnesium. These are: M. beasleyi Kaczmarek et al., 2012 from Turkey, M. burgessi Schilabach et al., 2018 from Kansas, U.S.A., M. decorum Morek et al., 2022 from Portugal, M. dornensis from Romania (Ciobanu et al., 2015), M. eurystomum Maucci, 1991 GB.005 from Scotland (Morek et al., 2020b), M. fridae Moreno-Talamantes et al., 2020 from Mexico, M. inceptum Morek et al., 2019 from Japan (‘M. cf. tardigradum’ in Suzuki, 2008), M. lagniappe Meyer et al., 2013 from Louisiana, U.S.A., M. matheusi from Madagascar (Kaczmarek et al., 2019), M. swansoni Young et al., 2016 from Kansas, U.S.A., M. tetralamellatum Pilato and Binda, 1991 from Tanzania, and Milnesium sp. (MilnC_010) from Charcot Is, Antarctica (Sands et al., 2008a, additional file 1); the last of which may be M. antarcticum. Milnesium rastrum sp. nov. is the latest addition to this list.

It is unknown whether any species in the above list have facultative parthenogenesis. Although M. inceptum always reproduces without males, male individuals emerged at a very low frequency even in such a thelytokous population (Suzuki, 2008); the reason for this has not yet been elucidated. The first author has observed that individuals of M. inceptum, since 2000, have laid at least one egg during the moulting period, even if the nutritional condition was not good. On the other hand, the solitary female of M. rastrum sp. nov. did two cycles of moult without oviposition, suggesting an obligate mode of bisexual reproduction. In dioecious tardigrades, females isolated from males never lay eggs (Lemloh et al., 2011; Bingemer et al., 2016). The same condition was apparently observed in M. eurystomum GB.005, in which males were found, and described as follows: “Since mature females extracted from the GB.005 sample did not lay additional eggs, the culture was terminated after a short period of time” (Morek et al., 2020b).

While parthenogenesis is considered more common in harsh environments (Nelson et al., 2015), it is interesting to note that an isolated ice-free area like Innhovde contains the presumed non-parthenogenetic species M. rastrum sp. nov. This may account for the rarity of Milnesium over the East Antarctica area, and suggests that M. rastrum sp. nov. had retained the plesiomorphic trait of bisexual reproduction since the isolation of Antarctica.

Tardigrade dispersal mechanisms are relatively poorly studied and the exact means are unknown. In the genus Ramazzottius Binda and Pilato, 1987, which has both reproductive modes, the diploid bisexual populations were found mostly in large moss or lichen encrustations on rocky outcrops, while the triploid or tetraploid parthenogenetic populations were from newly formed patches of moss or lichen on tree trunks (Bertolani et al., 1990). This observation suggests a theory that parthenogenesis with multiploidy originated somewhere, and then settled and propagated in harsh and isolated habitats. However, this theory cannot be applied to the genus Milnesium because parthenogenetic populations of M. tardigradum, M. pacificum Sugiura et al., 2020, and M. inceptum all showed diploid chromosomes (Sugiura et al., 2020). So, parthenogenetic Milnesium flourish in harsh environments regardless of the ploidy.

Antarctica was a warm and well-vegetated land mass before becoming isolated and frozen about 30 Mya. Tardigrades, able to survive the harsher environment, would have survived, staying in the remaining vegetation but becoming more isolated by glacier formations. This, and the more plesiomorphic bisexual reproduction as the ancestral mode may account for the isolated pockets of Antarctic dioecious tardigrades, e.g., M. rastrum sp. nov., Mopsechiniscus franciscae Guidetti et al., 2014, (Victoria Land), and Echiniscus corrugicaudatus McInnes, 2010 (inland nunataks in Ellsworth Land). While these Antarctic environments are extremely isolated, they have also provided very stable and rich habitats as refugia for tiny bisexual creatures for several tens of millions of years. However, this also makes these ecosystems very fragile against settlement of more fertile parthenogenetic animals as well as contemporary invaders (see Hughes et al., 2020) or possible future destruction of the natural barrier around the Antarctic Continent (Convey and Peck, 2019).