Rough backs: taxonomic value of epicuticular sculpturing in the genus Milnesium Doyère, 1840 (Tardigrada: Apochela)

The phylum Tardigrada comprises ~ 1400 described species that inhabit a wide range of ecosystems throughout the globe. Tardigrades are generally considered taxonomically challenging due to a strongly limited number of taxonomically informative morphological traits and microscopic size. Of all tardigrade groups, the taxonomy of Milnesium Doyère, 1840 is particularly difficult because in comparison with most other eutardigrades, the genus lacks the taxonomically informative egg shell ornamentation and/or placoids in the muscle pharynx. Therefore, any new morphological traits that could be used in species delineation and identification are priceless. In this contribution, we review and evaluate taxonomic value of the dorsal cuticle morphology. Specifically, by means of experimental taxonomy, we demonstrate the first extreme case of ontogenetic variability in dorsal epicuticle sculpturing using a new species from Portugal, Milnesium decorum sp. nov. Furthermore, we verify the existence of dorsal gibbosities in Milnesium reticulatum Pilato, Binda, Lisi, 2002, the only species of the genus allegedly exhibiting these structures. Finally, we provide a diagnostic key to the Milnesium granulatum morphogroup.

Microscopy, imaging and morphometry. The specimens were mounted on permanent microscope slides in Hoyer's medium according to the method by 29 to examine general morphology in PCM and acquire morphometric data. The measurements follow 30 , the buccal tube widths were measured according to 11,12 and the body length was measured from the anterior to posterior margin of the body, excluding the hind legs. Pseudoplate row numbers are given according to 26 and poorly visible pseudoplates are marked with a dotted line. The pt index is a ratio of a given structure to the length of the buccal tube, expressed as a percentage 31 and in the text is given in italics. The number of measured specimens follow the recommendation of 32 when it was possible, otherwise all available and properly fixed and oriented specimens were measured. Structures were measured only if their orientation was suitable. We present the joined measurements of specimens of the same CC (i.e. juveniles and adults) in a single table. The morphometric data was handled using the Apochela spreadsheet ver. 1.3. available from Tardigrada Register 33 , www. tardi grada. net. All the measurements and photographs were taken with Olympus BX53 PCM associated with Olympus DP74 digital camera (PCM). Pseudoplate arrange-   39 40 18S_Tar_Rr1 Reverse  GCC GCA GGC TCC ACT CCT GG   28S rRNA  28S_Eutar_F  Forward  ACC CGC TGA ACT TAA GCA TAT   41  42   28SR0990  Reverse  CCT TGG TCC GTG TTT CAA GAC   42   ITS-2  ITS2_Eutar_Ff Forward GCA TCG ATG AAG AAC GCA GC 43 43 ITS2_Eutar_Rr Reverse  TCC TCC GCT TAT TGA TAT GC   COI Table 3). Moderate length Milnesium species, up to 783 µm ( Fig. 1), yellow. Eyes present in all living individuals and in the majority of Hoyer-fixed specimens (9/11; 82%). The dorsal cuticle covered with reticulum, which is clearly visible on pseudoplates (Figs. 1B, 2) and weakly developed in the remaining areas of the dorsum (Figs. 2-3). In larger specimens (4 th + instars), the reticulation may be poorly visible in PCM (Fig. 2C). This species is characterised by numerous pseudoplates (Fig. 1C-D) arranged in nine transverse rows, which are clearly visible both in PCM and UVM: (I) a single trapezoid pseudoplate (divided into four equal rectangular portions); (II) a large central rectangular pseudoplate (divided into four rectangular portions, with the two anterior rectangles being larger) + two lateral oval pseudoplates; (III) central rectangular pseudoplate (divided into four equal rectangular portions, concave laterally) + lateral square pseudoplates with protuberance matching the concave sides of the central plate; (IV) large roundish central pseudoplate (divided into six equal rectangular portions) + roundish lateral pseudoplates; (V) a central rectangular pseudoplate (divided longitudinally into two equal rectangles) + two lateral roundish pseudoplates (divided into four unequal rectangular portions); (VI) a large central rectangular pseudoplate (divided into six equal rectangular portions) + two lateral elongated pseudoplates with curvy edges (divided longitudinally into two unequal portions); (VII) a central rectangular pseudoplate (divided longitudinally into two equal rectangles) + two lateral rectangular pseudoplates (divided into four unequal rectangular portions); (VIII) the largest, most complex, trapezoid pseudoplate (divided into eight parts: a central triangle and seven quadrangles) + two roundish lateral pseudoplates with small projections; (IX) four pseudoplates arranged transversally (internal trapezoid and the lateral roundish). Mouth opening surrounded with six short peribuccal papillae (with the ventral one being the smallest) and six triangular peribuccal lamellae of unequal size (with the two lateral lamellae significantly smaller, i.e. the 4+2 configuration; Fig. 4A). The lamellae configuration is unambiguously visible only in SEM. Two short lateral cephalic papillae present. Buccal tube cylindrical and of moderate width (Fig. 4B).
Typical Milnesium claws. Primary branches with tiny accessory points visible both in PCM and SEM (   Mature males (from the third instar onwards; morphometrics in Table 4). In the sample only single male was found (preserved on SEM stub) but the culture yielded additional 10 specimens. Smaller than females (Fig. 1E), with narrower buccal tube and with modified first pairs of claws into rigid hooks (Fig. 4H), as in all other Milnesium species. Eyes present in living animals, but absent in 7/10 (70%) of Hoyer-fixed specimens. Cuticular bars under claws always absent (these are the first Milnesium males reported to be lacking cuticular bars).  Table 3). Morphologically similar to adult females but with a better developed dorsal reticulation and more weakly outlined dorsal pseudoplates (Fig. 2B). Eyes present in living animals but absent in both Hoyer-fixed specimens. Cuticular bars absent. Other qualitative traits as in adult females. Table 5). Morphologically similar to juveniles but with a better developed dorsal reticulation and more weakly outlined dorsal pseudoplates ( Fig. 2A). All secondary branches   (Fig. 4C,D). Eyes present in living animals but absent in all 11 Hoyer-fixed specimens. Cuticular bars absent. Other qualitative traits as in adult females.

Hatchlings (first instar, morphometrics in
Ontogenetic variability. Milnesium decorum sp. nov. undergoes developmental changes in two key taxonomic traits, cuticular sculpturing and CC. The dorsal cuticle sculpturing becomes less clear with every consecutive instar. Specifically, under PCM, it is most pronounced in hatchlings, slightly less developed in juveniles, and it is very weakly outlined in adults or even not visible at all in large adult females. Under SEM, the reticulum also fades with subsequent moults, but it is detectable in all life stages (Fig. 3). The CC changes from -  in hatchlings to [2][3]-  in juveniles, i.e. the species is characterised by early negative CC change.
Eggs. Smooth, oval, slightly yellowish; deposited in exuviae; up to 4 in a single clutch were found in the culture. Remarks. The species was accompanied with a Milnesium sp. from the almatyense complex (Milnesium sp. #5 PT.010 A in 8 ), which was much more abundant in the sample. All of the 26 eggs of M. decorum sp. nov. incubated in the laboratory culture hatched, and nine hatchlings and seven juveniles were fixed on microscope slides. All remaining ten juveniles kept in the culture eclosed into males. Thus, with no adult females obtained in the laboratory, the culture was terminated. As a result, most of the type series consists of hatchlings and males. Because of that, the type series contains small number of mature females, which are in addition poorly to moderately preserved, thus as a result we designated a juvenile as the holotype.
Adult and juvenile phenotypic differential diagnosis. Milnesium decorum sp. nov. is one of the 13 described spe-   Ramazzotti, 1962 15 , known only from Chile (the only confirmed record), by a relatively narrower standard buccal tube width (25.5-28.7 in the new species vs 46.3 in M. granulatum paratype; morphometrics from 11 ).
Genetic differential diagnosis. The ranges of uncorrected p-distances between the new species and available sequences for other congeners are as follows: The original description of M. reticulatum highlights the importance of providing the readers with raw data, such as photomicrographs, as this is the only way the scientific community may widely and at any time evaluate the interpretation and conclusions laid out by the authors of the original contribution (the re-examination of specimens is not always possible and much more difficult than accessing raw data provided in the article, supplementary materials or in open data repositories). In 22 , all images, including the dorsum and the alleged gibbosities, are in the form of drawings, thus the reader is presented only with an interpretation. It has been demonstrated that morphological interpretations may vary considerably between researchers 56 and the original description of M. reticulatum is a striking example of this phenomenon. Another misinterpretation of Milnesium morphology was recently exposed by 57 who showed that the alleged three spines on the dorsum of the invalid now "Milnesium tardigradum trispinosa" 58 5F and 6). Since only a fraction of Milnesium species have been imaged in SEM, the taxonomic value of fine sculpturing identifiable only in SEM is yet to be evaluated when more data are available. Nevertheless, the term "smooth cuticle" has to be used carefully, always with the reference to the type of microscope that was used to make the distinction. However, as more new species in the genus are uncovered, new types of cuticular sculpturing may be revealed. In fact, we have found such a new morphotype represented by an undescribed species collected in Colombia (Milnesium sp. nov. CO.004; Table 1; Figs. 8 and 10). This species is characterised by a genuine granulation present on the entire body, including the ventral side, which has never been reported in any Milnesium species before. The granulation is slightly larger on the dorsum than on the ventral side and in the caudal part compared to the cephalic part of the body, but all granules are evident in LCM (Fig. 10). The granules are in the shape of irregular polygons, most often concave and with 7-10 edges (Fig. 10D). Besides granulation, we observed pseudopores, but only in the cloacal cuticle. (Fig. 10E, insert). Even though this is clearly a new species, we refrain from describing it as a new taxon because of the lack of associated DNA sequences and the low number of available specimens (N = 2). The small sample size prevents the assessment of intraspecific variability and the exclusion of morphological aberration as the explanation for this extraordinary phenotype. Moreover, if there are more species exhibiting this type of sculpturing, describing this Colombian species without genetic data could make it difficult to delineate these hypothetical similar species, creating a potential taxonomic impediment that we have already seen too many times in the history of tardigrade research (e.g. see 59 ). In other words, we are of the opinion that the species should be described only when more individuals are found and their DNA is sequenced (see also 13 ).
In addition to epicuticular sculpturing and endocuticular pseudopores, some Milnesium species also exhibit endo-or sub-cuticular areas of thicker cuticle described and termed as pseudoplates independently by 13 and 60 ; however, they have been noted before although without naming them (e.g. 17,61 ). Moreover, 13 suggested that the number, shape and arrangement of these structures could possibly be used for species delineation and identification, but this view was questioned by 26 , who hypothesised that pseudoplates do not exhibit variation within the genus and therefore should not be used as a taxonomic trait. However, our extensive analysis of numerous species, some represented by multiple populations, under both PCM and UVM showed that there are species, such as M. tardigradum (Fig. 11), that never exhibit pseudoplates. Thus, although it needs to be thoroughly tested whether in species with pseudoplates the shape and arrangement of these structures may be subject to interspecific variation, the presence vs absence of pseudoplates appears to be a valid discriminative taxonomic trait. In the great majority of Milnesium species, for which ontogeny has been described, cuticle appears similar or the same both in sexually immature and mature instars, except for endocuticular pseudopores and pseudoplates that are usually absent or less developed and therefore more difficult to identify in hatchlings and juveniles than in adults (see M. variefidum 13 and M. tardigradum in 19 ). However, there are two species in which ontogenetic variability in the epicuticular sculpturing has been observed: M. pacificum 27 and M. decorum sp. nov. (the present study). In both these species, the reticulation is most developed in hatchlings and it becomes weaker with each consecutive instar, but the differences between the life stages are more pronounced in the latter taxon. Given that ontogeny has been inves-

Conclusions and future directions
We have integratively described the 45th species of the genus Milnesium. The new species, M. decorum sp. nov., represents the granulatum morphogroup and is the most striking example of ontogenetic variability in epicuticular sculpturing to date. We also amended the description of M. reticulatum, demonstrating that gibbosities are not present in any of the known Milnesium species. Moreover, our study showed that more research is needed to clarify the types of fine epicuticular sculpturing that are identifiable only under SEM, but appear as smooth cuticle under LCM. Further studies should also address the taxonomic value of pseudoplate number, shape and arrangement. Finally, the lack of evidence for phenotypic differences between M. krzysztofi and M. pacificum, noted when constructing the diagnostic key, calls for an integrative redescription of the senior species and is a reminder that utmost care must be takes when differentiating new and described Milnesium species.