A weighted taxonomic matrix key for species of the rotifer genus Synchaeta (Rotifera, Monogononta, Synchaetidae)

Abstract A new, weighted matrix identification key for 34 largely undisputed species of Synchaeta was created with the aim of providing comparable, detailed and diagnostic character sets for each species that can be applied to live and/or preserved specimens. As part of this process, 14 species of Synchaeta were intensively re-investigated with respect to their habitus and trophi morphology using binocular, light, and scanning electron microscopy, which, together with behavioural observations, revealed several new discriminating characters. Whenever possible, missing information for any character was added for the remaining species from the literature, with the two recently described species Synchaeta arcifera and Synchaeta squamadigitata being considered for the first time in an identification key. Beyond its completeness, our key has two distinct advantages. First, the characters are supported by detailed illustrations of their respective character states whenever possible to both simplify identification and minimize any uncertainty in the descriptions themselves. Second, the new approach of weighting the characters according to their reliability, robustness and/or ease of determination was employed. This latter approach is especially advantageous for soft-bodied rotifers such as species of Synchaeta, where, for example, several external characters can be influenced by preservation and are therefore less diagnostic or reliable. Although the key is as comprehensive as possible, information for many species remains missing for many characters, thereby highlighting the need for additional comprehensive and detailed species (re-)investigations within Synchaeta.


Introduction
The rotifer genus Synchaeta (Monogononta, Synchaetidae) comprises approximately 37 (see Segers 2007) to 39 (see Jersabek et al. 2018) valid and truly planktonic species, of which approximately half occur in brackish and/or marine habitats (Hollowday 2002). Although their importance in aquatic food webs is unquestioned because of their often dominant role in the rotifer (Stemberger and Gilbert 1985) and metazooplankton communities (Arndt et al. 1990), specimens of Synchaeta in ecological studies are seldom identified to species level (Obertegger et al. 2006). To a large extent, this situation derives from the identification of and delimitation between species of this genus being regarded as being especially challenging (Pourriot 1965;Ruttner-Kolisko 1972;Koste 1978).
The several comprehensive revisions and keys of Synchaeta that exist (e.g., Voigt 1956Voigt -1957Ruttner-Kolisko 1972;Koste 1978;and most recently Hollowday 2002) tend to be restricted in that they limit themselves to describing the most concise set of characters that delimit each species. Although this represents a useful simplifying strategy, the inherently incomplete data set it entails presents two distinct disadvantages. First, because additional, alternative characters are not presented for many species, their identification is impossible when the respective, diagnostic ones are deformed or not clearly visible (e.g., foot morphology when it is retracted). Second, and more importantly, direct comparisons among species are usually not possible because the species are often described using different sets of characters.
To address both sets of issues, we have developed a new identification key for Synchaeta, with the dual aims of making it both easy to use and as comprehensive as possible by providing large, comparable data sets for each species. To accomplish this, we thoroughly re-examined live and prepared specimens of 14 species and intensively researched the literature for all members of Synchaeta, including the most recently described species Synchaeta arcifera Xu, 1998 and Synchaeta squamadigitata De Smet, 2006, which are presented for the first time in a comparative identification key. In addition to an in-depth analysis of the habitus, we focussed on the trophi in particular because they are considered to be both species-specific (De Smet 1998;Fontaneto and Melone 2006;Wulfken et al. 2010) and less susceptible to conservation (Kutikova 1970 as cited in Obertegger et al. 2006). Detailed information about the trophi are therefore of great advantage in ecological studies, for example, where the material is necessarily fixed for practical reasons (Obertegger et al. 2006;Labuce and Strake 2017), with the consequence that the species identity of distorted or contracted rotifer specimens might be determinable only via their trophi (De Smet 1998;Segers 2004). By providing comparable data sets for each species, our taxonomic key also functions as a matrix key (also known as free access or multi-access key; see Hagedorn et al. 2010), which is better able to incorporate missing information when identifying species than the more traditional dichotomous key.
In addition, we weighted all characters within our matrix key according to their perceived discriminatory power. This strategy of focussing on more robust and diagnostically conclusive characters potentially facilitates accurate species identification by giving characters that are more susceptible to variation or artefacts (e.g., body shape, which is strongly affected by the pressure of a cover slip (Koste 1978), fixation/preservation (Ruttner-Kolisko 1972;Segers 2004) or by developing eggs and stomach content in soft-bodied rotifers) less impact than more constant and robust ones (e.g., the number and position of the lateral antennae.) To further simplify the identification process, we supported the characters with detailed illustrations and photographs of the respective character states whenever possible and introduce a consistent and distinctive terminology for homologous structures. Although the latter point seems obvious, the use of synonyms for homologous structures is a widespread problem, even within the same key. For example, Hollowday (2002: 90) variously denotes the apical receptors as "sensory frontal antennae", "sensory antenna", "frontal prominence with tuft of setae" or "sensory setae" in his identification key for species of Synchaeta.
Our purpose here is to deliver a comprehensive and robust key for Synchaeta by which a reliable identification of live and preserved specimens is feasible through a comprehensive and comparable morphological data set. In so doing, the present study not only confirmed existing discriminatory characters, but also re-described some of them more explicitly (e.g., foot shape and morphology of the apical receptors) as well as established several novel ones for species demarcation (as e.g., behaviour, morphology of the pedal glands, detailed fulcrum and ramus morphology).

Study site and sampling
Using a 55-µm mesh plankton net, sampling for species of Synchaeta took place intermittently between June 2013 and August 2017 in northwest Germany from freshwater habitats in Oldenburg, Lower Saxony and Tecklenburger Land, North Rhine-Westphalia as well as from brackish and marine habitats in Wilhelmshaven, Lower Saxony (Table 1). Species of Synchaeta found in the samples (Table 2) were identified using the existing information in Rousselet (1902), Voigt (1956Voigt ( -1957, Ruttner-Kolisko (1972), Koste (1978) and Hollowday (2002).

Binocular and light microscopical (LM) investigations
Undisturbed, living specimens were initially observed in a petri dish using a binocular microscope to examine their (swimming) behaviour. For the LM analyses, single individuals were isolated and carefully sedated with carbonated water before being further immobilized through the slight pressure of a cover glass. For the latter step, extreme care was taken to avoid any deformation of the body, which could lead to morphological artefacts. All observations used differential interference contrast using a LEICA DMLB microscope and digital photographs were taken using a Canon EOS 5D Mark II camera.

Scanning electron microscopical (SEM) investigations
For SEM examinations of the habitus, single specimens were initially sedated with carbonated water before being euthanized with 1% OsO 4 buffered in 0.1 M NaCa cacodylate buffer and fixed with 240 mOsmol picric acid-formaldehyde (Melone and Ricci 1995). To examine the isolated trophi, the surrounding tissue of selected specimens was dissolved according to the protocol of Kleinow et al. (1990) by transferring them into a droplet of dissolving agent (0.1 g dithiothreitol added to a 5-ml stock solution of 5.2 g sodium dodecyl sulphate + 0.24 g NH 4 HCO 3 in 100 ml aqua dest; AppliChem, Darmstadt, Germany) for ca. 15 min before being rinsed with distilled water subsequently. Thereafter, samples of either the habitus or the trophi were dehydrated using an ascending, graded ethanol series. Following critical-point drying, each sample was attached onto an SEM stub and coated with gold-palladium before being examined on a Hitachi S-3200N SEM.

Illustrations
All new illustrations of the habitus were made using Adobe Illustrator CS4 based on representative digital photographs. References from drawings that we have obtained and modified from literature are listed below each illustration.

Included species and information from the literature
Overall, 34 species of Synchaeta were considered in our key, with six species that are recognized by either Segers (2007) and/or Jersabek et al. (2018 being excluded (see lists below). Information about species that we did not find in our samples derive from their respective initial descriptions and from the literature, with an emphasis on Rousselet (1902), Lauterborn (1905), Lie-Pettersen (1905), Peters (1931), Voigt (1956Voigt ( -1957, Ruttner-Kolisko (1972), Koste (1978) and Hollowday (2002). Information or interpretations that we have made from illustrations or photographs that are derived from other sources than the above-mentioned literature are indicated below each table (Tables 3-8).

Character clarification, character weighting, and species identification
To support the written descriptions, the morphology for each character state is also exemplified both through illustrations as well as the naming of at least one exemplar species that possesses the respective state. The character states are represented in detailed tables (Tables 3-8) and in a numerical list for each species where the text is formatted according to the perceived reliability and/or discriminatory power of the states: 1. "?": The character state is unknown or ambiguous for the respective species -further examinations are required. 2. brackets: The character state rarely occurs in the species.

italics:
The character is variably expressed within the species or its interpretation is either subjective or can be easily misunderstood because of potential artefacts that can arise during preparation. These characters should be applied with caution. 4. normal text: The character state is more or less robust, but shared by several, additional species of Synchaeta. Many characters of this quality are usually required for species demarcation in the form of a unique character set for each species. 5. blue color: The character state is robust and important insofar as it is unique for the species and/or shared by only a few, additional species of Synchaeta. Individual characters in this category typically exclude many other congeneric species to greatly simplify species demarcation. 6. bold: The character state is robust and species-specific (autapomorphy). To simplify the identification process, blank character checklists and tables for recording character states are appended (Suppl. material 1: Tables S1, S2).

Identification characters
Habitat and behaviour (  (Table 3) 7. Overall body length of mature specimens (measured from the apical field to the distal tips of the toes, excluding the cilia) a. less than 250 µm b. more than 250 µm Head and neck region (Table 4) 8. Apical field -Width in relation to the trunk width 1 a. as wide as the trunk ( Fig. 4A; exemplar S. tremuloida) b. wider than the trunk ( Fig. 4B; exemplar S. longipes, S. triophthalma) 9. Apical field -Elevation with respect to auricles a. level ( Fig. 4C; exemplar S. grimpei) to slightly elevated ( Fig. 4D; exemplar S. tremula) b. intermediate ( Fig. 4E; exemplar S. triophthalma) c. strongly elevated; distinctly convex ( Fig. 4F; exemplar S. grandis) 10. Dorsolateral styles -Elevation a. not raised to very slightly raised ( Fig. 4G; exemplar S. tremula) b. intermediate ( Fig. 4H; exemplar S. gyrina) c. strongly raised ( Fig. 4I; exemplar S. baltica) 11. Auricles -Size a. not clearly distinct from the rotatory organ ( Fig. 4J; exemplar S. grimpei) 1 This character strongly depends on the amount of ingested food or presence of developing eggs in some species (e.g., S. gyrina; character state "a/b").  Trunk (Table 5) 17. Trunk region -Shape 3 a. conical: trunk decreases gradually in width caudally ( The intensity of the colour is often variable within the species and a less intense colour can appear as colourless 3 The body shape is strongly influenced by the pressure of the cover glass or fixation/preservation (Koste 1978) as well as from developing eggs and stomach content.   The length of the pedal glands is related to the foot length. Caution should be exercised with individuals that carry their foot retracted to any degree because in such situations the glands can appear longer than they really are and/or that they extend into the trunk. 6 The present character overlaps with the previous one but is more specific. It should only be consulted when the state can be determined with certainty because the state "b" can be difficult to determine accurately in the case of a second rudimental toe that can be easily overseen. Otherwise, we recommend character 31 for species identification. . Foot, pedal glands and toes. A-I Presence of pseudosegments (arrows) and morphology of pedal glands (grey areas) A foot pseudosegmented, pedal gland single and of foot-length (S. neapolitana) B glands asymmetrical with the larger one being of foot-length, glands terminating in the toes (S. triophthalma) C glands asymmetrical with one gland terminating in the toe and one in the dorsal spur (S. hutchingsi) D symmetrical glands of foot-length, voluminous proximally and decreasing gradually in width moving caudally (S. tremula) E glands shorter than the foot, each spherical proximally and abruptly decreasing caudally before widening again to form a reservoir (S. oblonga) F glands longer than the foot (S. prominula), G glands of foot-length and tubular (S. longipes) H foot with wrinkles, glands tubular, suspended from the trunk (S. pachypoda) I foot pseudosegmented, each gland with two voluminous sections, suspended from the trunk, glands terminate proximally to the toes (S. pachypoida) J-Q Symmetry, size and separation of the toes J single toe (S. triophthalma) K asymmetrical, toes of different shape (S. cecilia) L symmetrical toes of medium size, bases of the toes are in contact (arrow), tips are close to one another or very slightly divergent ( (Table 7) 38. Cerebral eye -Morphology a. single ( Fig. 10A; exemplar S. pectinata) b. paired but partially fused or connected by pigment granules (Fig. 10B; exemplar S. triophthalma) c. paired and distinctly separated from one another ( Fig. 10C; exemplar S. oblonga, S. lakowitziana) 39. Cerebral eye -Size a. small to medium, evenly shaped (Fig. 11A A general re-examination is recommended here to verify if any apparent pseudosegmentation is actually an artefact caused by wrinkles (e.g., S. hyperborea; described as "wrinkled" in Friedrich and De Smet (2000), but "pseudosegmented" in Hollowday (2002)), by the insertion point of the spur (e.g., S. neapolitana) or by the preanal-fold overlapping the foot (e.g., S. oblonga; Wilke et al. 2018b). 8 The granules are often regarded as being frontal eyespots, but they are more likely frontal aggregations of pigmented granules of the retrocerebral organ (Wilke et al. in prep.) 9 The streams of pigment granules appear to be present in the ducts of the retrocerebral organ (Wilke et al. in prep.).  43. Apical receptors -Elevation a. on a flat or slight central elevation of the apical field ( Fig. 10D; exemplar S. oblonga) b. on a strong central elevation of the apical field ( Fig. 10E; exemplar S. grandis) c. on two bulges or pimples ( Fig. 10F; exemplar S. triophthalma) d. on strong, paired elevations (tentacles) (Fig. 10G; exemplar S. pectinata) e. on a single, tubular elevation ( Fig. 10H; exemplar S. fennica, S. johanseni, S. vorax) 44. Lateral and dorsolateral styles -Length 10 a. minute ( Fig. 10I; exemplar S. squamadigitata) b. short ( Fig. 10J; exemplar S. grimpei) c. medium ( Fig. 10K; exemplar S. pectinata) d. long ( Fig. 10L; exemplar S. vorax) 45. Dorsal antenna -Elevation a. none to a slight elevation ( Fig. 10M; exemplar S. oblonga) b. distinct prominence to a snout-like projection ( Fig. 10N Fig. 13A; exemplar S. hutchingsi, S. triophthalma) b. one; right lateral antenna of normal size, left one is absent ( Fig. 13B; exemplar S. tamara) c. paired symmetrical lateral antennae of normal size ( Fig. 13C; exemplar S. tremula) 10 These states are admittedly subjective, especially in the absence of any reference point. However, for those species for which literature data only was available, it was often not possible to be more precise.
As such, we have listed all states except the obvious "minute" as subjective in the associated table and highly recommend comparison with our illustrations of the exemplar species for each condition.  48. Lateral antenna(e) -Location relative to the median transversal axis of the body a. directly lateral on the median transverse axis ( Fig. 13D; exemplar S. tremula) b. ventrolateral to the median transverse axis ( Fig. 13E; exemplar S. pectinata) c. mid-dorsal and slightly displaced to the right of the body axis ( Fig. 13F; exemplar S. tamara) 49. Lateral antenna(e) -Location relative to the longitudinal plane a. in the posterior third of the trunk region ( Fig. 13G; exemplar S. oblonga) b. in the caudal-most trunk region at or near the base of the foot ( Fig. 13H; exemplar S. tremula) c. on lateral lobes caudally to the cloaca ( Fig. 13I; exemplar S. grimpei) 50. Lateral antenna(e) -Base a. surrounded by a tubular (Fig. 13J; exemplar S. johanseni) or papillary ( Fig.  13K; exemplar S. oblonga) epidermal fold b. surrounded by a low epidermal fold ( Fig. 13L; exemplar S. pectinata) Trophi (

Spine of frontal hook 14
a. absent (e.g., Fig. 14A     ? a a/b † Adheres to objects only when disturbed and then only for a short time and without any twisting movement. ‡ Adherence to objects observed by Lauterborn (1905) and Remane (1929), but not by Hood (1893).

Weighted matrix key for Synchaeta -detailed tables
a/c † As already noted by Hollowday (2002), this species requires further revision (preferably on living, non-preserved specimens) because of inconsistencies in the published morphological data for it, especially for the neck region and the apical field. ?  Hollowday (2002), we found that S. oblonga exhibits a foot with only one instead of two pseudosegments. The impression of two pseudosegments being present might derive from the distinct preanal-fold that distinctly overlaps the foot, which itself is predominantly withdrawn (Wilke et al. 2018b).  ?    Friedrich and De Smet (2000) | Ramus according to a LM image of the habitus of an individual by Jersabek et al. (2003b), where the trophi were visible ¶ Trophi according to SEM and LM images from Obertegger et al. (2006) # Trophi according to LM image from Rougier et al. (2000) † † Trophi according to Jersabek et al. (2003a) Pale red aggregations of pigment granules located around the darkly pigmented cerebral eyes can make the latter appear large.

Discussion
Our weighted taxonomic matrix key constitutes the most comprehensive and comparable compilation of morphological and behavioural characters to date for the 34 species of Synchaeta that we consider to be valid. Through it, we hope to facilitate the reliable identification of both live as well as of preserved specimens, in part by highlighting those features that are more susceptible to the effects of preservation and, more generally, by indicating the reliability of different characters or individual character states for species identification.
In attempting to make our key as comprehensive as possible, we undertook detailed re-examinations of 14 species of Synchaeta (Wilke et al. 2017;Wilke et al. 2018a, Wilke et al. 2018b, supplemented by information from the literature where necessary. Nevertheless, we were restricted exclusively to literature information for 20 species of this genus, resulting in numerous cases of both missing information and uncertainty (indicated with a "?" in the tables 3-8). A pervasive problem in the literature is that many species have not been re-discovered since their initial description (e.g., S. atlantica and S. rousseleti; see Hollowday 2002) so that little information exists for them at all and that many species descriptions are extremely brief and exclusively restricted to the most obvious, diagnostic characters that discriminate the species from other known species and usually highly similar ones (e.g., S. tremuloida; see Pourriot 1965). Thus, it is not uncommon that important, but basic information is missing for many species such as for example the location and number of sensory antennae for S. atlantica, S. rousseleti and S. cylindrica (see Hollowday 2002), information that could also distinguish the species from new ones discovered in future. In addition, information is often missing or conflicting for those characters for which data are hard to obtain. For example, although the trophi are important for species identification in rotifers (De Smet 1998;Segers 2004), special skills and equipment are needed for their investigation (Telesh and Heerkloss 2002) such that they are often disregarded and so incompletely known for several species of Synchaeta (e.g., S. johanseni; see Hollowday 2002).
A further problem is that many illoricate species like those in Synchaeta have been described on the basis of preserved material only and, despite repeated calls not to do so (e.g., Donner 1959), fixation is commonly used in rotifer research (Labuce and Strake 2017), with its practical applications making it a necessary evil. However, in soft bodied rotifers such as Synchaeta, preservation is far more evil than it is necessary insofar as it causes deformations and/or distortions (Ruttner-Kolisko 1972;Koste 1978;Shiel and Koste 1993), with the consequence that species potentially include preservationinfluenced characters in their respective descriptions. This, in turn, might explain why several species have only ever been found once (e.g., S. atlantica and S. rousseleti, both of which were described using preserved material). However, even more commonly reported members of Synchaeta were described on the basis of preserved material as well (see Hollowday 2002), including S. lakowitziana, which is "notoriously disputed" (Hollowday 2002;p. 103) for some aspects of its characteristic morphology (e.g., the sharp neck constriction) that are suspected to be a preparation artefact.
Altogether, these problems highlight the need for standardized and comprehensive species descriptions in Synchaeta as well as in rotifers more generally comprising morphological (habitus and trophi), behavioural and molecular data from both live and preserved specimens (e.g., in Wilke et al. 2017;Wilke et al. 2018a, Wilke et al. 2018b. Such an integrative approach ensures the most comprehensive data set for the respective species and facilitates an assessment of which characters are potentially affected by preservation-based deformations and to which degree. Depending on the context (e.g., ecological assessments), it will often be difficult to avoid preservation. However, knowledge of its specific effects and providing sets of characters that are robust to them will facilitate better species identification. As such, we hope that our weighted taxonomic matrix key for Synchaeta, both through its comprehensiveness as well as through its use of weighting to indicate character reliability and utility, will not only make species identification in Synchaeta easier, but will also serve as a model for future keys within rotifers.