Echiniscidae from the Sierra Nevada de Santa Marta, Colombia, new records and a new species of Bryodelphax Thulin, 1928 (Tardigrada)

Abstract Three species of Echiniscus are recorded for the first time from Colombia: Echiniscus dariae, Echiniscus kofordi, and Echiniscus perarmatus. In addition, the description of the new species Bryodelphax kristenseni sp. n., is mainly based on the presence of ten paired plus two unpaired granularly sculptured ventral plates, the dorsal plate ornamentation with superficial irregular pores, no spine on the anterior legs, and the hind legs without papillae or dentate collar.


Materials and methods
This survey was based on tardigrade specimens deposited in the Centro de Colecciones Biológicas de la Universidad del Magdalena under the catalogue acronym CBUMAG:TAR. The material was collected between 2011 and 2012, from different localities (San Lorenzo, Bella Vista, and Medium basin of Garupal River) in the Sierra Nevada de Santa Marta, Colombia, from 543 and 2,200 m a.s.l. All specimens were preserved on slides in Hoyer's medium.
trema, growing on a tree trunk. Localities: Bella Vista, Sierra Nevada de Santa Marta, 11°05'47.8"N, 74°05'04.4"W, 1930 m a.s.l. Remarks. The morphological characters of the Colombian specimens agree with the original description of E. perarmatus (Murray, 1907). This species, according to the literature, has a tropical and subtropical distribution; with the type locality, Cape Colony (South Africa), it has also been recorded in Indonesia, Hawaii, United States, and Venezuela (McInnes 1994). The apparent wide distribution we suggest indicates E. perarmatus might be a species complex. Therefore further work with original material, or specimens from the type locality, would be required to solve this problem. This is the first record (sensu lato) for Colombia.
Description of the holotype. Body colourless, eyespots absent or not visible after mounting. Total body length, 126.5 µm. Scapular and terminal plate not distinctly divided but unsculptured folds indicate the different portions of the plates (Fig. 1B, 2A). In particular on the scapular plate, a pair of lateral longitudinal grooves differentiates clearly the small lateral portions from the main median, which shows a median longitudinal fold crossed by three less defined transversal bends ( Fig. 2A). A median longitudinal fold, not always obvious, is also present in the unpaired plates. The terminal plate has a pair of longitudinal folds ( Fig. 2A), which separate the plate into a median and two lateral portions; the former appearing crossed by irregular transversal folds that are not always clearly visible (Fig. 1A, 2A). All three median plates divided, but the posterior portion of the third plate is narrow and rectangular. The first median plate transversally subdivided in two parts by a suture devoid of sculpture, a trapezoid anterior portion, and a triangular posterior section with a rounded caudal edge. The anterior portion of the second median plate is triangular and the posterior section trapezoid; in the posterior area an unsculptured triangular region is visible but, due to its appearance, was not interpreted as a plate. The third median plate is divided, the main anterior plate triangular, with an anterior rounded edge, and a posterior sculptured portion, but sometimes hidden by the terminal plate in contracted specimens. Paired plates also divided into an anterior moderately narrower portion, and a posterior wider portion (Figs 1A, 2A). The shape and arrangement of all plates and their sub-portions is outlined in Fig. 2A. All plates show apparent double sculpture: big pores more or less irregular in size and distribution, often fused to one another ( Fig. 1A-B) in some cases almost forming patterns, and a lower level of regular "granules" (i.e. cuticular pillars). The pores tend to form three transversal bands on the scapular plate (Fig. 1B), while on the terminal plate they tend to be grouped in areas that are separated by stripes without pores, thus almost outlining "facets". On the remaining plates, the pores tend to be arranged on each sub-portion of the plate in a more or less defined transversal band (or a single line on the narrowest sub-plates) (Fig. 1), i.e. a band on the anterior and a band on the posterior portions of the paired plates, a band on the anterior and a single line on the narrow posterior portions of the unpaired plates.
The cuticular pillars (Fig. 1A, B) appear regularly distributed, and vary in size between plates and the different part of each plate/sub-plate. The largest (about 1.2 µm) are on the scapular and terminal plates, and the proportion among the granules of the different plates is as follows: scapular = terminal > posterior portions of the paired plates > anterior portions of the paired plates > unpaired plates. In the latter, the "granules" appear not only smaller but also fainter. On each plate or sub-plate, the terminal plate excluded, the "granules" are larger on the median transverse band, and gradually decrease in the more cephalic and caudal portions as well as the very lateral portions, almost at the borders. On the terminal plate the largest granules lie on a band at about ¾ of the length of the plate, gradually decreasing, going to the more cephalic portion and at the very lateral and caudal portions, almost at the borders. Six pairs of lateral supplementary platelets, difficult to see, are present between the paired plates ( Fig. 2A).
Ventral plates present, but faint and difficult to observe, consist of ten paired plus two unpaired (IX/X:2-(1)-4-4-2-4-2-1-2-1 according to Kaczmarek et al. 2012), as depicted in Fig. 2B; the brackets in the formula and the question mark in Fig. 2B indicate the difficulty in ascertaining the presence of an unpaired ventral plate at the level of legs I: if that plate is present, then the unpaired plates are three and not only two, and the plate rows are ten instead of nine. All ventral plates show a faint granulation under PCM (Fig. 1C, arrows). The rest of the ventral cuticle is smooth. All legs with a band of small dots.
Spines on the first pair of legs absent or not visible under PCM. The dentate collar on the fourth pair of legs absent (but the sculptured platelet of the dentate collar present). No papilla visible on the hind legs under PCM. External claws smooth. Internal claws with spur oriented towards the base (Fig.1D, arrow a indicating a spur). Measurements of some structures of the holotype and ranges among the paratypes (larva excluded) are given in Table 1 (Supplementary Data provides all measurements for these specimens). No eggs were found.
Remarks. The paratypes exhibit the same morphology, but with a certain degree of variation with regard to appearance of the bands of the scapular and terminal plates, more visible in less relaxed specimens. The narrow posterior portions of the median plates (especially with regard to the third), are more visible in well-relaxed specimens but can be totally hidden in contracted specimens. In such specimens, the posterior elements of each couple of supplementary platelets, especially the third, could also be hidden. In addition, the orientation of the specimen on the slide meant the supplementary lateral platelets were not always clearly visible. With regard to the ventral plates, these were evident in some specimens, e.g. that chosen as holotype, but were not always easy to see. In general, the ventral plates varied from faint to almost invisible (without clear indication of a link with life stage); it took a very accurate, long observation under both PCM and DIC to identify all the plates and to be certain of the number and arrangement. Such plates show in PCM a faint granulation, which is actually what, in some cases, made them visible; their borders often being unclear. In some of the specimens not even one ventral plate was apparent at first sight, requiring very careful observation to detect at least some of them; thus this character can pass unnoticed. We therefore recommend great care in observing Bryodelphax before considering whether a specimen is without ventral plates, and also without supplementary lateral platelets.
Another character, for which considerable individual variability is noted, is the distribution of the cuticular pores, which may be arranged from a relatively regular distribution, as described in the holotype, to a quite random distribution. Additionally, the transversal bands of pores of the sub-portions of the paired and unpaired plates can be reduced to a single, more or less regular row.
Finally, the papilla of the hind legs in most specimens was not visible, but in a couple of individuals, there appeared to be an extremely small, faint papilla. However, the presence of particles in the slides preparation prevented us from being sure that what we observed was a papilla and not some out of focus particle. This character, therefore, needs to be confirmed.
Etymology. The species is named in honour of Professor Reinhardt Møbjerg Kristensen, in particular for his valuable contribution to the taxonomy of Echiniscidae (Kristensen 1987). Differential diagnosis. According to Kristensen et al. (2010), the new species falls into the weglarskae group due to the presence of ventral plates. Within this group the dentate collar is absent from only two species: Bryodelphax sinensis (Pilato, 1974) and B. aaseae. However, due to the difficulty in observing the ventral plates in some specimens of our new species, we also compare species descriptions where ventral plates were not reported (stressing the fact that especially in early publications the ventral plates may have passed unnoticed or not considered a valuable character). Species with the same type of cuticular ornamentation, and without the dentate collar include: B. parvulus, B. asiaticus Michalczyk, 2004 andB. ortholineatus (Bartoš, 1963). Bryodelphax aaseae (Kristensen, Michalczyk & Kaczmarek, 2010) is the most similar species, sharing with the new species the same ventral plate configuration (if the median unpaired plate at the level of legs I is also present in the new species). These plates were described as smooth by the authors, but in Kristensen et al. (2010 -figs 9-11 and 19), there is the appearance under PCM of granulation, and this might be similar to the new species. Another character we noted was the apparent presence of lateral supplementary platelets between the paired plates in B. aaseae, (see: Kristensen et al. 2010 - fig. 7 The diagnosis of B. parvulus was revised by Pilato et al. (2010), and this species should lack ventral plates. Moreover, another difference with the new species is the length of the clava: 3.4-4.3% of the body vs. about 2.5% of the body in B. parvulus (see Pilato et al. 2010, in which a specimen from Poland attributed to B. parvulus by Węglarska 1959, andconfirmed by Pilato, was measured, slide No. 1476 of Pilato andBinda collection).
Bryodelphax kristenseni sp. n. differs from B. asiaticus, which lacks the ventral plates, in having non-granulated ventral cuticle, anterior portions of median plates 1 and 2 markedly larger than the posterior portions (which are almost a stripe), while in B. asiaticus the posterior portions of those plates are only slightly smaller.
The new species differs from B. ortholineatus in having spurs on internal claws (absent in B. ortholineatus), in having supplementary platelets (not mentioned in the original description (Bartoš, 1963) and reported as absent by Fontoura et al. 2008), and the shape of the median plates looks different in Bartoš' original drawing, but it must be stressed that the drawing was very stylised.

Discussion
As mentioned above, in B. kristenseni sp. n. the third median plate is divided into an anterior and a posterior portion; this character, until now, has been considered typical of the genus Bryochoerus, while in Bryodelphax the third median plate has been considered undivided.
At first, there were doubts on the identification due to the division of the third median plate, although hidden in some specimens, which led us to Bryochoerus; on the other hand, there was evident resemblance between these specimens and Bryodelphax aaseae, and with other species, e.g., Bryodelphax weglarskae (Pilato, 1972). This encouraged careful examination of that species as well as many others congeners, and it was noted that the divided third median plate was also present in B. aaseae (figs 7 and 8 in Kristensen et al. 2010), B. asiaticus ( fig. 6 in Kaczmarek and Michalczyk 2004), B. parvulus (Ramazzotti and Maucci 1983, page 221, fig. 70 from Thulin's 1928 original description), and the following material from the Pilato and Binda collection, which was examined: B. brevidentatus (paratype, slide No. 5386, Fig. 3 A), B. meronensis (holotype, slide No. 5350, Fig. 3 B) (Fig. 3 D).
The definition of the genus Bryodelphax is therefore suggested as follows: Small Echiniscidae with non-flexible buccal tube with CaCO 3 encrusted stylet supports. No lateral or dorsal appendages present except cirrus A. Median plates all divided, but the caudal portion of the third median plate may be hidden by the terminal plate. Without pseudosegmental plates. Ventral plates may be present.
With respect to Kristensen's (1987) definition, we have preferred not to include the presence of "red granulate eyes" as a character of the genus because the presence or absence of eyes is variable in many genera, and can be lost in the slide mounting process (e.g. B. kristenseni sp. n.).

Conclusions
Authors from an earlier period (ca. 1900-1950s) traditionally considered many species to be cosmopolitan, which in tardigrade taxonomy created species-groups. These speciesgroups, along with past errors and misinterpretations, now require careful analysis in order to amend taxa descriptions and differentiate the sibling species. As taxonomic knowledge has progressed, key characters have been added that were not considered essential in older references. We are left with a legacy of early species descriptions that are often impossible to identify without type material, which in many cases is sadly unavailable. Taking into account the absent or poor state of older type material, and the difficulty in resampling a vaguely described locus typicus, the possibility of abolishing or classifying a suspect species as "species dubia" should be considered. This would help prevent further confusion created by non-taxonomists or beginners using an apparently simple diagnostic key (e.g., Ramazzotti and Maucci (1983) monograph) to record questionable species.
With our present contribution, based on material in a museum collection, three new records enrich the list of Echiniscidae recorded for Colombia. In addition, a species new to science was discovered, which provided the occasion to make evalu-ations at a higher taxonomic level. The fact that a relatively limited study provides new and interesting results, in spite of the great efforts from past decades, is evidence of how little is known about tardigrade fauna and biogeography for most regions of the world. It further highlights how much our taxonomic knowledge has grown and can still grow.