Morphology and Ontogenesis of Psilotrichides hawaiiensis nov. gen., nov. spec. and Molecular Phylogeny of the Psilotrichidae (Ciliophora, Hypotrichia)

The Psilotrichidae are a family of middle-sized hypotrichs with unique morphological and ontogenetic features (e.g. the oral primordium develops in a deep pouch) that, however, did not provide a definite phylogenetic signal. Thus, we studied the 18S rRNA gene of Urospinula succisa (Müller 1786) Esteban et al., 2001 as well as the morphology and ontogenesis of Psilotrichides hawaiiensis, a new genus and species from an ephemeral swamp on Oahu Island, Hawaii. The molecular data classify the psilotrichids into the oxytrichids but without clear branching position. A brief revision, using the structure of the oral apparatus, the location of the contractile vacuole, and three ontogenetic features, showed four distinct genera: Psilotricha Stein, 1859; Urospinula Corliss, 1960; Hemiholosticha Gelei, 1954; and Psilotrichides nov. gen., which differs from the confamilials mainly by the obliquely oriented buccal cavity and the shape of the undulating membranes as well as by a distinct ridge along the right buccal margin. The pyriform species, P. hawaiiensis, is about 65 × 45 μm in size and is easily recognized by the table tennis racket-shaped appearance due to the elongated last cirrus of the left marginal row. Refined diagnoses are provided for the family Psilotrichidae Bütschli, 1889 and the genera contained.

into the Amphisiellidae, and Psilotricha Stein 1859a and Hemiholosticha Gelei 1954 into the Psilotrichidae. Esteban et al. (2001) revised the Psilotrichidae and redescribed the type species, P. acuminata, and accepted two genera, Psilotricha and Urospinula, and several species. They classified Psilotricha into the Oxytrichidae and Urospinula to the Orthoamphisiellidae, following Eigner (1997). However, Berger (2011) rejected the transfer of Urospinula to the Orthoamphisiellidae because the frontoventral cirral rows do not originate via primary primordia and the anlagen A1 and A2 of the opisthe originate from the oral primordium (Foissner 1983).
Obviously, the morphological and ontogenetic data did not unambiguously reveal the phylogeny of the Psilotrichidae. Thus, we applied molecular methods to Urospinula succisa whose morphology and ontogenesis were described by Foissner (1983). This showed an unexpected position of the Psilotrichidae within the oxytrichid clade. Further, we describe a new psilotrichid genus and species, showing that the psilotrichid diversity is not yet exhausted.

Materials
Psilotrichides hawaiiensis was discovered in a sample of dry surface soil and litter (0-3 cm) from an ephemeral swamp grown with fern (Marsilea sp.) on Koko Head, Oahu Island, Hawaiian archipelago, W157°41 0 44″ N21°15 0 52″. Unfortunately, we did not store specimens for sequencing because the species was discovered 20 yr ago when molecular characterization just began.
The sample was analyzed with the "nonflooded Petri dish method" as described by Foissner (1987Foissner ( , 1992. Briefly, this simple method involves placing 50-500 g airdried terrestrial material (soil, leaf litter, roots, etc.) in a Petri dish (13-18 cm wide and 2-3 cm high) and saturating, but not flooding it, with distilled water. Such cultures are analyzed for ciliates by inspecting about 2 ml of the runoff on days 2, 7, 14, 21, and 28. To obtain sufficient dividers, raw cultures were set up in Petri dishes containing Eau de Volvic (French Table water), a few ml of the eluate from the nonflooded Petri dish culture, and some crashed wheat kernels to stimulate growth of food, i.e. bacteria and protists.
Urospinula succisa was collected from a nonflooded Petri dish culture with soil from a rice field in the surroundings of the Lake Biwa Museum, Japan. It was cultivated as described for P. hawaiiensis above. Three voucher slides, reg. no. 2013/47-49, have been deposited in the Biology Centre of the Museum of Upper Austria (Biologiezentrum des Ober€ osterreichischen Landesmuseums), Linz (LI). Relevant specimens have been marked by black ink circles on the coverslip.
Hemiholosticha sp. was found in the Simmelried, i.e. in a moorland pond in Bavaria, N49°2 0 E10°45 0 . For details, see Kreutz and Foissner (2006). Environmental specimens were used for the investigations because the species was rather abundant; it will be described in a forthcoming paper.

Morphological methods
Living cells were studied using a high-power oil immersion objective and differential interference contrast microscopy. Protargol impregnation and scanning electron microscopy (SEM) were performed as described by Foissner and Xu (2007). For protargol impregnation, P. hawaiiensis was fixed with Stieve's solution, which produced rather mediocre preparations while alcohol and Da Fano fixation produced very good results in Hemiholosticha.
Counts and measurements of silvered specimens were performed at a magnification of 1,250X. In vivo measurements were conducted at magnifications of 40-1,000X. Drawings of live specimens were based on free-hand sketches and micrographs, those of impregnated cells were made with a drawing device. In the ontogenetic stages, parental structures are shown in outline while newly formed structures are shaded black. Each of the stages depicted has been seen in at least two specimens.
Terminology is according to Foissner (1983) and Lynn (2008) while details of the oral apparatus are according to Foissner and Al-Rasheid (2006).

Phylogenetic analyses
Prior to phylogenetic analyses, sequences were quality checked, and PHRED/PHRAP analyses were carried out using CodonCode Aligner v.3.0 (CodonCode Corporation, Dedham, MA). Vector and primer nucleotides were trimmed off. The sequence of U. succisa was first subjected to a BLAST analysis (Altschul et al. 1997) against GenBank's nr database. Then, the 18S rDNA sequence was aligned to available hypotrich families and choreotrichs as outgroup using MUSCLE (Edgar 2004) as implemented in SEAVIEW (Gouy et al. 2010), and subjected to Gblocks (Castresana 2000) for refinement. Manual inspection for further editing of the alignment was conducted in MacClade v4.05 (Maddison and Maddison 2005). The GTR-I-Γ evolutionary model was best fitting selected by the AIC in jModeltest v0.1.1 (Guindon and Gascuel 2003;Posada 2008). The resulting alignment used for phylogenetic analyses included 62 sequences and 1750 characters. Maximum likelihood (ML) analyses were carried out in RaxML-HPC v7.2.5 (Stamatakis et al. 2008). Support came from a majority rule consensus tree of 1,000 multiparametric bootstrap replicates. An evolutionary distance tree using Neighbor Joining (NJ) algorithm was calculated in SeaView (Galtier et al. 1996). Support came from 1,000 bootstrap replicates. Trees were visualized with FigTree v1.3.1 (Rambaut 2006). The GenBank accession number for U. succisa is KF411460.

RESULTS
Description of P. hawaiiensis nov. spec.
Psilotrichides hawaiiensis has a size of 55-75 9 40-50 lm in vivo, usually it is about 65 9 45 lm, as calculated from in vivo measurements and the morphometric data in Table 1, adding 15% preparation shrinkage. The bluntly pyriform or table tennis racket appearance of the body is caused by the prominent terminal cirrus belonging to the left marginal row. Rarely, the cells are broadly ellipsoidal with acute rear end or have a sigmoidal left body margin. Usually, P. hawaiiensis is dorsoventrally flattened up to 2:1. In lateral view, most cells are roughly hemiellipsoidal, i.e. have a convex ventral side and a flat or sigmoidally curved dorsal side (Table 1; Fig. 1A-F, I-K and 2A-C, E-G).
The nuclear apparatus is slightly anterior of the central quarters of the cell and left of body midline, usually being composed of two macronuclear nodules and one micronucleus in between. The macronuclear nodules are rotund to ellipsoidal, on average 14 9 10 lm in protargol preparations; the average distance between the nodules is 3 lm, occasionally they are connected by a fine strand; the nucleoli are usually rotund, 0.5-2 lm across, rarely up to 4 lm. The micronucleus is globular to broadly ellipsoidal, on average 3.6 9 2.9 lm in protargol preparations (Table 1; Fig. 1A, J and 2C).
The contractile vacuole is in midbody at the left cell margin, and the cytopyge is near the acute posterior end (Fig. 1A, B). The cortex is inflexible, colorless, and lacks specific granules. The cytoplasm is colorless and studded with food vacuoles up to 12 lm in diam., some ordinary crystals about 4 lm in size, and lipid droplets 1-5 lm across (Fig. 1A, I). Psilotrichides hawaiiensis feeds on colorless flagellates of the genera Polytoma and Hyalogonium, both having a red eye-spot subapically; most specimens are packed with this kind of food, both in the nonflooded Petri dish culture and in the raw cultures (Fig. 1A, I and 2A). The ciliate glides slowly to rapidly on the microscope slide and can perform short, fast jumps.
There are 18-26 cirri in four ventral, one postoral, and in the right and the left marginal row (Table 1; Fig. 1A,K and 2B,E). The length of the rows is highly variable (CV usually > 20%, Table 1) while the total number of cirri is rather stable (CV 9.5%). Most cirri are thin and long, i.e. are composed of 3 9 3 to 3 9 4 (rarely 5) cilia 20-25 lm long in vivo, those of the left marginal row are short or even absent (but the basal bodies are present, Fig. 2D, E), except of an approximately 30 lm long terminal cirrus that inserts on the acute body end and is composed of cilia of various lengths (Fig. 3B); frontal, buccal and transverse cirri are not distinguishable. The cirri are widely spaced and their number in the individual rows is highly variable (CV usually > 20%), making the pattern difficult to discern. The anterior cirrus of ventral row R1 is near the distal end of the adoral zone and thus possibly homologous to frontal cirrus III/3 of other hypotrichs; the distance between the two anteriormost cirri of ventral row R3 is often increased; and the cirri of ventral row R4 usually form a group each at the ends of the row (Table 1; Fig. 1A, K, 2B, D-G and 3B).
The dorsal bristles are 3-4 lm long in vivo and in protargol preparations. They are arranged in three meridional rows slightly (rows 2 and 3) to distinctly (row 1) shortened anteriorly; some parental bristles are preserved in six of 23 specimens. Caudal cirri are absent (Table 1; Fig. 1J and 2C, F).
The adoral zone of membranelles occupies an average of 43% of body length; the largest bases are 6-7 lm wide in vivo. When the cell is viewed ventrally, the zone looks like a question mark cut proximally while it is semicircular when the cell is viewed apically (Table 1; Fig. 1A, G, K, 2A, B, E and 3A). On average, the adoral zone is composed of 21 ordinary membranelles with the intermembranellar distances strongly increasing from proximal to distal. Likewise, the length of their cilia gradually decreases from about 20 lm distally to 5 lm proximally; further, the cilia of the short fourth row are only 1.5 lm long. The 10-12 frontal membranelles, which have ciliary row 4 distant from the centripetal end, are often separated from the ventral membranelles by a membranelle composed of only three kineties (Fig. 1G, H). In vivo, the frontal membranelles are partially covered by an approximately 3 lm high scutum that is strongly asymmetrical, i.e. the right half is much shorter than the left one which merges into the margin of the membranellar stripe ( Fig. 1A and 2A). Under the scutum is an accumulation of lipid droplets up to 5 lm across. Lateral membranellar cilia, intermembranellar ridges, and membranellar bolsters are absent (Table 1; Fig. 1A, G, H, K, 2A, B, E and 3A).
The buccal cavity is about half as long as the ventral portion of the adoral zone, i.e. about 11 lm long, 8 lm wide, and 4 lm deep in the protargol preparations. The right margin is occupied by a conspicuous, 2-3 lm thick ridge anteriorly merging into the frontal cortex and posteriorly into the buccal vertex. A buccal seal is not recognizable (Table 1; Fig. 1A, E, 2E and 3A, C, D).
The paroral membrane inserts in a shallow slit at the right margin of the buccal ridge and forms an acute to very acute angle (22-44°) with the longitudinal axis of the cell. The cilia produce an undulating, up to 7 lm high membrane gradually decreasing to 3 lm at both ends. In many specimens more or less large parts of the paroral are doubled or even triplicated. The endoral membrane extends side by side with the paroral, and its cilia form always an approximately 4 lm long plate, indicating that they are motionless (Fig. 3C, D). The pharyngeal fibers extend obliquely backwards and are 8-10 lm long in protargol preparations (Table 1; Fig. 1A, G, H, K, 2E and 3C, D).

Ontogenesis of P. hawaiiensis
Very early dividers ( Fig. 3E and 4A-C) The oral primordium, i.e. an anarchic field of basal bodies, is formed de novo on the cell surface between the postoral and the left marginal row. The macronuclear nodules show a reorganization band. The dorsal infraciliature is unchanged.
Early dividers (Fig. 3F, G, 4D-F, 5A, B and 6A) The oral primordium increases in size and begins to invaginate. When the first protomembranelles, which are composed of only two kineties, are formed, the pouch becomes deep and prominent (Fig. 3F, G, 4D-F and 5A). Some basal bodies remain on the right margin of the pouch and will later form opisthe's anlage A1 (Fig. 3F, G, 4F and 5A). The parental undulating membranes begin to reorganize (Fig. 4F, 5B and 6A). Four streaks of basal bodies appear within ventral cirral rows R3 and R4 to form the anlagen A3 and A4 in proter and opisthe (Fig. 3G, 4D-F and 5B). In the protargol preparations ( Fig. 4D-F) but not in the scanning electron micrographs ( Fig. 3G and 5B), the two opisthe streaks are posteriorly connected by scattered basal bodies, as in U. succisa which produces opisthe anlage A4 in this way (Fig. 32 in Foissner 1983 and Fig. 4G in the present publication). The proter anlage A1 is formed by the dedifferentiated posterior cirrus of row R1, and anlage A2 by the anterior cirrus of row R2. The opisthe's anlage A1 is produced by the oral primordium while anlage A2 is formed by the dedifferentiated posterior cirrus of row R2 (Fig. 3G, 4E, F, 5B and 6A). Some cirri of the right and left marginal row disintegrate and form the marginal anlagen in proter and opisthe (Fig. 3G, 4E, F, 5A and 6A).
Early mid-dividers ( Fig. 6B-D) The distal half of the oral primordium, which is still growing and differentiating, evaginates. The last adoral membranelles are formed inside the pouch, which becomes partially covered by the cortex. An anlage for the undulating membranes of the opisthe now separates from the right posterior portion of the oral primordium (Fig. 6D) and is thus not visible in the SEM (Fig. 5B). The parental undulating membranes continue reorganization. All cirral anlagen have been formed and produce cirri. Supernumerary minute anlagen and basal bodies occur frequently between the anlagen A2 and A3 (Fig. 6B, D). The dorsal kineties form within-row primordia in proter and opisthe; both basal bodies of the dikinetids are ciliated, making the anlagen prominent (Fig. 6C). The micronucleus slightly inflates and shows a fibro-granular structure, which remains up to nuclear division (Fig. 6C).

Mid-dividers (Fig. 5C, D and 6E-H)
The oral primordium is still evaginating and curves to the right so that the membranelles form a convex zone; the proximal third of the primordium is still in the pouch ( Fig. 5C and 6E). The buccal cavity is developing and the undulating membranes, which are already ciliated and orient increasingly parallel to the proximal part of the adoral zone of membranelles, migrate onto its dorsal wall; the buccal ridge is not yet recognizable (Fig. 5C, D). In late mid-dividers, a remarkable process occurs ( Fig. 5D and 6G): the undulating membranes which are separating, and the proximal third of the adoral zone become inclined and orient almost transversely to the main body axis. The parental buccal cavity and undulating membranes are reorganizing, i.e. the cavity disappears and the cilia shorten, according to the SEM investigations (Fig. 5C, D). The parental membranellar zone does not show any sign of reorganization.
In the anlagen, the new cirri are migrating to their specific sites, in both proter and opisthe (Fig. 5C, D and  6E, G). Anlage A1 migrates to the distal end of the adoral zone. Anlage A2 migrates to the buccal vertex. Anlage A3 splits: the anterior portion remains at the level of the gap in anlage A4 while the posterior portion migrates leftwards to become the postoral row, as does anlage A4 in U. succisa (Fig. 49 in Foissner 1983 and Fig. 7D in the present publication). Anlage A4 elongates and splits in two portions separated by a wide gap. Cirri not involved in anlagen formation become resorbed.
The macronuclear nodules have fused in the cell center and the micronucleus commences division (Fig. 6F, H). The new dorsal bristle rows elongate, replace the parental bristles, and the dikinetids lose the posterior bristle (Fig. 6F, H).
Late dividers (Fig. 5E and 7A-C) When the division furrow becomes recognizable, the proter forms a new posterior body end at the left margin (Fig.  7A, B). The buccal cavity of the proter commences redeepening and the undulating membranes finish reorganization and obtain their final position (Fig. 5E). The adoral zone and undulating membranes of the opisthe are inclined to the longitudinal body axis by about 90°or, in other words, the posterior half of the adoral zone and cirral row R1 are oriented transversely to the main body axis ( Fig. 5E and 7B). Parental cirri are continuously resorbed. The new dorsal bristle rows reach their final length and most dikinetids lose the posterior cilium (Fig. 7C). The macronucleus elongates and divides. The micronucleus has already divided but both are still connected by a fibrous structure (Fig. 7C).

Molecular phylogeny of U. succisa (M€ uller 1786)
The Japanese population is morphologically highly similar to that from Austria studied by Foissner (1983). Thus, a redescription is not necessary. The 18S rRNA sequence is 1,671 base pairs (bp) long and its closest relative sequence deposited in public databases is the 18S rRNA sequence of Bistichella variabilis, an unclassified genus (accession number HQ699895.1, He and Xu 2011). The two taxa share a sequence similarity of 97.79%. In spite of this comparatively high sequence similarity, Urospinula and Bistichella are at rather different sites in the phylogenetic tree (Fig. 8). However, all clades involved have poor statistical support, indicating that clade composition and phylogenetic relationships are not yet settled and thus may change significantly when more sequences become available. Morphologically, Urospinula and Bistichella have little in common: cortex rigid vs. highly flexible; cirri hardly differentiated vs. highly differentiated into, e.g. frontal, buccal, and transverse cirri; oral primordium in deep pouch vs. on cell surface.

Psilotrichides hawaiiensis as a new genus and species
Foissner (1989) characterized the oxytrichid genera Oxytricha, Stylonychia, Cyrtohymena, and Steinia by the shape of the buccal cavity and undulating membranes. This has been widely acknowledged (for a review, see Berger 1999). A similar (convergent?) diversity occurs in the Psilotrichidae where Psilotrichides is unique in having a strongly oblique buccal cavity and undulating membranes (Fig. 9A-D). A further character is the buccal ridge, which is not only unique to the family but very likely to the entire subclass. Psilotrichides differs from Urospinula, the sole genus whose ontogenesis has been investigated (Foissner 1983), mainly by the oral apparatus ( Fig. 9A-D) and the number of cirral anlagen originating from the oral primordium (see below).
Psilotrichides hawaiiensis is unique in having a pyriform body with a narrow posterior end, where an elongated left marginal cirrus causes a table tennis racket-shaped appearance of the cell. The body shape resembles P. acuminata (Fig. 10A-H) and, especially, B. viridis Penard 1922 (Fig. 11N). However, Penard's species has the contractile vacuole posterior of the buccal vertex (vs. at left body margin) and possibly possesses transverse cirri and symbiotic green algae both absent from P. hawaiiensis.

Ontogenetic comparison
There are three distinct differences and similarities each in P. hawaiiensis and U. succisa, as described by Foissner (1983): the postoral cirral row is ontogenetically inactive in Psilotrichides while it produces cirral row R3 in Urospinula; the parental undulating membranes are reorganized in Psilotrichides while they appear unchanged in Urospinula; and the oral primordium produces one cirral row in Psilotrichides while two in Urospinula. The similarities are: the oral primordium develops in a deep pouch, as in euplotids and oligotrichs (Foissner 1996;Lynn 2008) while it develops on the surface or in a shallow concavity in the stichotrichs, e.g. in Steinia sphagnicola (Voß and Foissner 1996); the undulating membranes do not produce cirri, as in Schmidingerothrix extraordinaria (Foissner 2012); and both have a migrating kinetofragment that produces the postoral cirral row.
The 18S rRNA sequence classifies Urospinula into the large Oxytricha-clade with a tentative relation to Kahliella, which would support the above mentioned assignment to the family Kahliellidae. However, statistically, this position is unsupported and the branching of the Psilotrichidae in phylogenetic analyses remains elusive. However, there are two strong ontogenetic arguments that Psilotricha, Psilotrichides, Urospinula, and Hemiholosticha belong to the same family: the oral primordium develops in a deep pouch and a migrating part of a ventral cirral row becomes the postoral cirral row. Unfortunately, the deep pouch development of the oral primordium might be a plesiomorphic and thus a phylogenetically weak character because it is found also in the euplotids (Foissner 1996). An oxytrichid relationship, already proposed by Stein (1859a,b), is indicated by the enigmatic genus Pachycirrus whose organization is not very different from that of a typical Oxytricha ( Fig. 11I-K). Indeed, an oxytrichid relationship of the psilotrichids is indicated by the migrating cirri of row 4, which are reminiscent of the cirri in anlage VI of the 18-cirri hypotrichs (the frontoterminal cirri are distinctly separated from the corresponding pretransverse and transverse cirri), and rows R1-4 which might be homologuous to the rows formed from anlagen III-VI. Increased taxon sampling as well as the analyses of genes with different rates of evolution may shed light on the phylogenetic position of the Psilotrichidae as well as of other oxytrichid families.

Materials for a revision of the psilotrichids
Our brief revision is based on Foissner (1983), Esteban et al. (2001), the present and some unpublished data, and two assumptions: P. acuminata Stein 1859a; type of the family and genus, has not been restudied and Pachycirrus How can we be sure that Urospinula, Hemiholosticha, and Psilotrichides are confamilial with P. acuminata? First, all have a rigid cortex already described by Stein (1859a,b) as "euplotid". Second, they all have a similar size and an undifferentiated ciliature without, e.g. distinct frontal, marginal, and buccal cirri. Third, P. hawaiiensis has a great overall similarity with P. acuminata.
Stein (1859a,b) discovered P. acuminata in a poorly studied habitat, viz. in a puddle strongly contaminated by liquid manure. All other species and populations described later are from more ordinary limnetic habitats, such as clean and eutrophic ephemeral puddles. The oral apparatus is distinctly different from that of other psilotrichid genera (Fig. 9A-D). It is rather large and has a deep buccal cavity, the right margin of which is occupied by a long, vertical paroral membrane. The adoral zone of membranelles is semicircular, and the length of the cilia abruptly decreases in the proximal third, a distinct feature present also in Hemiholosticha spp. (W. Foissner, unpubl. data). In contrast, the somatic cirral pattern is quite similar to that of Psilotrichides and Urospinula ( Fig. 1A and 10I). A further feature shared by Psilotricha and Psilotrichides is the lateral location of the contractile vacuole while all other described populations have it posterior of the buccal vertex, i.e. near the body center (Fig. 10I,Q and 11C,N).
Pachycirrus costatus, which was discovered in a nonflooded Petri dish culture with grassland soil from Scotland, poses several problems. It has been "fully" described as a new genus and species by Olmo and Esteban (1999) in a congress abstract. Although not recommended, the name is very likely available according to Article 9 of the International Commission on Zoological Nomenclature (1999) because the genus and species were diagnosed in the abstract. Thus, this must be considered as the original description. Later, Esteban et al. (2001) identified this population as P. acuminata Stein 1859a. We disagree because P. costatus has a quite different oral apparatus (buccal cavity narrow, paroral minute) and three distinct caudal cirri (Fig. 11I-K). We classify P. costatus as incertae sedis as we do with P. dragescoi Groli ere 1975, which is possibly a kahliellid (Fig. 11L); and with B. viridis Penard 1922, which has possibly transverse cirri (Fig. 11M, N).  Gelei (1944). T, U. Urospinula sinistrocaudata after Gelei (1944). V. Urospinula simplex after Dragesco and Dragesco-Kern eis (1986).

TAXONOMIC SUMMARY
Class Spirotrichea B€ utschli 1889 Subclass Hypotrichia Stein 1859b Family Psilotrichidae B€ utschli 1889 Improved diagnosis. Medium-sized, ellipsoidal hypotrichs with posterior body end rounded, acute, or with one or two spines. Two macronuclear nodules, usually one micronucleus in between. Contractile vacuole at left body margin or near body center slightly posterior of buccal vertex. Cortex rigid, in some species with distinct ridges. Cirri long and sparse, arranged in several ventral rows, one right and one left marginal row, and a postoral row originating from the posterior, migrating fragment of a ventral row; frontal, buccal, and transverse cirri not distinguishable. Three to five dorsal kineties; caudal cirri absent. Oral apparatus occupies about one-third to one-half of body length, in four distinct patterns ( Fig. 9A-D Remarks. In the absence of new data (see above), the diagnosis remains incomplete but is sufficient to separate Psilotricha clearly from the confamilials. Of particular significance is the abrupt shortening of the cilia in the proximal half of the adoral zone because this rare feature is present also in Hemiholosticha (W. Foissner, unpubl. data). One must credit Stein (1859b), who shows this feature clearly in his figures (Fig. 10A, B).

Genus Urospinula Corliss 1960
Improved diagnosis. Psilotrichidae with one or two posterior spines. One or two micronuclei between or near macronuclear nodules. Contractile vacuole near midbody. Adoral zone of membranelles C-shaped, length of cilia gradually decreasing from distal to proximal. Right margin of buccal cavity and undulating membranes usually straight in main body axis. Two cirral rows produced by the oral primordium; parental undulating membranes not reorganized; postoral cirral row ontogenetically active. Type species. Trichoda succisa M€ uller 1786. Foissner (1983) neotypified this species with an Austrian population and combined it with Psilotricha. This is outdated, according to the new data. Later, Esteban et al. (2001)  Corliss 1960 (syn. of U. succisa), P. acuminata sensu Dingfelder 1962 (syn. of U. succisa), U. calcibia (Gelei 1944) Corliss 1960, U. sinistrocaudata (Gelei 1944) Corliss 1960, and Urospinula simplex Dragesco and Dragesco-Kern eis 1986. Remarks. The diagnosis is based on the study of Foissner (1983) and the present data. Possibly, U. calcibia (Gelei 1944) lacks a postoral cirral row. Species are distinguished by the number of ventral cirri and cirral rows, respectively (about 54 in U. succisa, only 15 in U. simplex), the number of posterior spines (only one in U. sinistrocaudata), and the number of dorsal kineties (U. succisa with three, U. calcibia with four, and U. sinistrocaudata with five). The GenBank code for the Japanese population of U. succisa is KF411460.
Genus Hemiholosticha Gelei 1954 Improved diagnosis. Ellipsoidal Psilotrichidae with micronucleus usually between macronuclear nodules. Contractile vacuole near body center. Adoral zone of membranelles C-shaped, length of cilia abruptly decreasing in proximal half. Right margin of buccal cavity and paroral membrane distinctly curved (cyrtohymenid). Two cirral rows produced by the oral primordium; parental undulating membranes reorganized; postoral cirral row ontogenetically active. Type species. Hemiholosticha viridis Gelei 1954. Species assignable. Hemiholosticha viridis Gelei 1954 and P. viridis sensu Kahl (1932). Remarks. The majority of the diagnosis is based on three unpublished species from Germany and Brazil (for an example, see Fig. 9E, F) because the description of Gelei (1954) appears rather incomplete and bewildering. He describes the buccal apparatus as follows (Fig. 11A-E): "The peristomial field is not a deepening but appears as a projecting wedge which runs into the adoral zone. Thus, the lip membrane, which is very short, is associated with a very short row of membranelles, forming a ciliary row". This text is difficult to interpret, especially because Fig. 11E shows a well developed buccal cavity. Generally, the cyrtohymenid oral apparatus in this kind of psilotrichs is difficult to analyze because part of the buccal cavity is covered by the buccal lip. Perhaps, Gelei (1954) recognized only the endoral membrane. We shall discuss this matter in detail when we describe these species.

Genus Psilotrichides nov. gen
Diagnosis. Pyriform Psilotrichidae with micronucleus usually between macronuclear nodules. Contractile vacuole at left margin of body. Adoral zone of membranelles semicircular, length of cilia gradually decreasing from distal to proximal. Buccal cavity and undulating membranes curved and distinctly slanted (~20-45°) to longitudinal axis of cell; with buccal ridge, i.e. conspicuously thickened right margin of buccal cavity. Oral primordium produces one cirral row; parental undulating membranes reorganized; postoral cirral row ontogenetically inactive.
Diagnosis. Size in vivo about 65 9 45 lm; bluntly pyriform. Two broadly ellipsoidal macronuclear nodules and one rotund micronucleus in between. On average a total of 23 cirri in four ventral, one postoral, and one right and one left marginal row; left marginal cirri usually short and partially unciliated, last cirrus in center of posterior pole and distinctly elongated, providing the species with a table tennis racket shape. On average 33 dorsal bristles in three kineties. Adoral zone occupies about 43% of body length, on average composed of 21 membranelles widely spaced in anterior half. Type locality. Surface soil and litter (0-3 cm) from an ephemeral swamp on Koko Head, Oahu Island, Hawaiian archipelago, W157°41 0 44″ N21°15 0 52″. Etymology. The epithet refers to the type locality. Type material. One holotype slide and eight paratype slides with morphostatic and dividing, protargol-impregnated specimens have been deposited in the Biology Centre of the Museum of Upper Austria (Biologiezentrum des Ober€ osterreichischen Landesmuseums), Linz (LI) reg. no. 2013/38-46. Relevant specimens have been marked by black ink circles on the coverslip.