Two new species were isolated from a subtropical coastal environment near Ningbo, China ( Figures 1A, B ). Aspidisca spina sp. n. was collected from a sandy beach (29°46′27″N, 121°56′8″E) in Meishan Bay on 3 September 2020; the water temperature was about 32°C, and salinity was 19 PSU ( Figure 1C ). Aspidisca shini sp. n. was collected on 24 December 2020, from a brackish lake (29°45′3″N, 121°54′30″E); the water temperature was about 10°C, and salinity was 20 PSU ( Figure 1D ). Samples were taken back to the laboratory and maintained as raw culture; rice grains were added to promote the growth of bacteria as food for ciliates. A. spina sp. n. survived only a few days under laboratory conditions. A. shini sp. n. could be maintained for about 3–4 weeks.

The behaviors of two species were observed in Petri dishes using a dissecting microscope. Living cells were observed under bright field and different interference contrast (DIC) microscopy at ×400–×1,000 magnification to obtain morphological features in vivo. The ciliary pattern and nuclear apparatus were revealed using silver staining methods (Foissner, 2014; Wu et al., 2021). The protargol powder was made according to Pan et al. (2013). Counts, measurements, and drawings of stained specimens were performed at protargol powder 1,000 magnification. The terminology is mainly according to Wu and Curds (1979).

The determination of the SSU rRNA gene sequences was performed on single cells for each species. The single cells were first washed with sterilized in situ water and distilled water, then transferred to a 1.5-ml microfuge tube with a minimum volume of water.

The DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany) was used for single-cell DNA extractions, following the manufacturer’s instructions modified by Wang et al. (2021). Each species has a parallel sample, and the sequencing information of the two experiments was identical. The SSU rRNA gene sequence was amplified with the primers 18s-F (5′-AAC CTG GTT GAT CCT GCC AGT-3′) and 18s-R (5′-TGA TCC TTC TGC AGG TTC ACC TAC-3′) (Medlin et al., 1988). The cycling conditions were as follows: an initial denaturation at 98°C for 30 s, 35 cycles of 10 s at 98°C, 20 s at 56°C touchdown, and 100 s at 72°C, and a final extension of 5 min at 72°C (Ye et al., 2021). Q5 Hot Start High-Fidelity DNA Polymerase (New England Biolabs Co., Ltd., M0493, Beijing) was used to minimize the possibility of amplification errors. Sequencing was performed bidirectionally at Tsingke Biological Technology Company (Hangzhou, China).

A phylogenetic tree was constructed by using the newly characterized SSU rRNA gene sequences and 83 related sequences in the NCBI GenBank database. Six Discocephalida species, namely, Leptoamphisiella vermis (FJ865203), Paradiscocephalus elongatus (EU684746), Pseudoamphisiella alveolata (DQ503583), Pseudoamphisiella quadrinucleata (EU518416), Prodiscocephalus borrori (DQ646880), and Discocephalus ehrenbergi (FJ196397), were selected as the outgroup taxa. The 85 SSU rRNA gene sequences were aligned using MUSCLE with default parameter settings on the GUIDANCE web server (Penn et al., 2010). BioEdit v.7.1.3.0 was used to manually trim both ends (Hall, 1999). Hypervariable sites were masked and removed using Gblocks v.0.91b with a default setting (Castresana, 2000; Talavera & Castresana, 2007). Maximum likelihood (ML) and Bayesian inference (BI) analyses were carried out online on the CIPRES Science Gateway (Edgar, 2004; Miller et al., 2010). Maximum likelihood (ML) analysis with 1,000 bootstrap iterations was constructed with RAxML-HPC2 v.8.2.11 on XSEDE (Stamatakis, 2014). Bayesian inference (BI) analysis were performed with MrBayes on XSEDE v.3.2.6 (Ronquist et al., 2012) under the GTR + I + G model selected by the program MrModeltest v.2.4 (Nylander, 2004). The chain length was 1,000,000 generations and sampled every 100 generations, whereby the first 2,500 trees were discarded as burn-in. Trees were visualized in MEGA v.7.0 (Kumar et al., 2016) and TreeView v.1.6.6 (Page, 1996). Systematic classification and terminology mainly followed Lynn (2008); Gao et al. (2016), and Lian et al. (2021).

The distinctive features of Aspidisca spina sp. n. are the beak-like peristomial spur and two transparent protrusions on the posterior end of the cell. In terms of the conspicuous peristomial spur and the “polystyle-arrangement” of frontoventral cirri, five species should be compared with A. spina sp. n., that is, A. magna Kahl, 1932, A. sedigita Quennerstedt, 1867, A. hexeris Quennerstedt, 1869, A. fusca Kahl, 1928, and A. dentata Kahl, 1928 ( Table 2 ).

Aspidisca magna was firstly collected from Kiel by Kahl (1932). Based on the Qingdao population, Li et al. (2010) redescribed this species in detail using modern research methods. It is very similar to A. spina sp. n. in the body shape, arrangement of transverse cirri, number of membranelles in AZM1, and dikinetids in the dorsal kineties. However, these two species can be clearly separated by the protrusion structure (absent in A. magna vs. two posterior marginal protrusions in A. spina sp. n.), the number of dorsal ridges (4 in A. magna vs. 6 in A. spina sp. n.), and membranelles in AZM2 (15–18 in A. magna vs. 19–21 in A. spina sp. n.) ( Figures 7A–C ; Table 2 ). In addition, the SSU rRNA gene sequences of the two species differ from each other by 159 nucleotides (Li et al., 2010).

Aspidisca sedigita was initially found by Quennerstedt (1867) in Visby, Gotland Island; Wu and Curds (1979) made a short review of this species. A. sedigita and A. spina sp. n. share the same nuclear apparatus (single horseshoe-shaped macronucleus) and the presence of posterior protrusion (population of A. sedigita in Quennerstedt’s original report also with posterior protrusions, but not obvious in other populations). A. sedigita differs from the new species in the number of dorsal ridges (2 or 3 vs. 6) and membranelles in AZM1 (6 or 7 vs. 8) and in AZM2 (12–14 vs. 19–21). ( Figures 7G, H ; Table 2 ) (Dragesco, 1963; Agamaliev, 1967; Wu & Curds, 1979).

Aspidisca hexeris was also collected from Visby, Gotland Island, by Quennerstedt (1869) and was regarded as synonym of A. sedigita in the revision of Wu and Curds (1979). Jiang et al. (2013) clarified the taxonomy position of this species based on the comprehensive morphological and molecular information. It resembles A. spina sp. n. in the body shape, eight membranelles in AZM1, and the horseshoe-shaped macronucleus. However, A. hexeris can be separated from the A. spina sp. n. by the smaller body length (20–25 µm vs. 75–80 µm), smooth cell ends (vs. two posterior marginal protrusions in A. spina sp. n.), fewer dorsal ridges (3 or 4 vs. 6), fewer membranelles in AZM2 (13–15 vs. 19–21), and fewer dikinetids in the fourth dorsal kinety (7–11 vs. 16–21) ( Figures 7D–F ; Table 2 ). Furthermore, A. hexeris can be distinguished from A. spina sp. n. by the 136-nucleotide substitutions in the SSU rRNA gene sequences (Jiang et al., 2013).

Aspidisca fusca was originally collected from Bad Oldesloe, Germany, by Kahl (1928), and the infraciliature was firstly provided by Jiang et al. based on a Chinese population (2013). Although A. fusca also has a prominent peristomial spur, it differs from A. spina sp. n. for the absence of posterior protrusions, the fewer membranelles in AZM1 (4 vs. 8) and AZM2 (15–19 vs. 19–21), and the morphology of the macronucleus (two sausage nodules vs. single horseshoe-shaped). ( Figures 7I–K , Table 2 ). Additionally, A. fusca is genetically distinct from A. spina sp. n. by 197-nucleotide substitutions in the SSU rRNA gene sequences (Jiang et al., 2013).

Aspidisca dentata was firstly discovered by Kahl (1928) but has not been studied in more detail since then. It is similar with A. spina sp. n. in the elliptical body shape and single horseshoe-shaped macronucleus. However, A. dentata has fewer dorsal ridges (4 vs. 6 in A. spina sp. n.) and fewer membranelles in AZM1 (4 vs. 8 in A. spina sp. n.) and AZM2 (6–10 vs. 19–21 in A. spina sp. n.). Besides, A. dentata has a prominent dorsal thorn (vs. absence in A. spina sp. n.), but A. spina sp. n. has two posterior protrusions (absence in A. dentata) ( Figures 7L, M ; Table 2 ) (Kahl, 1928; Kahl, 1932).

Aspidisca shini sp. n. is a brackish water species with a small size. It is mainly characterized by the oval-shaped outline, the “polystyla-arrangement” frontoventral cirri, and the horseshoe-shaped macronucleus. With reference to its morphological characteristics and ciliature, it can be compared with A. steini Buddenbrock, 1920, A. aculeata (Ehrenberg, 1838) Kahl, 1932, A. polypoda Kahl, 1932, and A. quadrilineata Kahl, 1932.

Aspidisca steini is a well-known species first found in a marine aquarium in Germany (Buddenbrock, 1920). Although the original description was very brief and incomplete, the diagrams provided the accurate arrangement of ventral cirri ( Figures 7T, U ). About 10 years later, Kahl (1932) identified an Aspidisca collected from Germany (Kiel) as this species and presented some morphological information; however, the arrangement of frontoventral cirri was slightly different from the original description, that is, two lines arranged in two parallel rows ( Figure 7V ) (Kahl, 1932). Borror (1963) provided the ciliary pattern of this species for the first time based on a population collected from USA ( Figure 7W ). The arrangement of frontoventral cirri was the same as Kahl’s population (Borror, 1963). Subsequently, several populations from Africa were investigated by Dragesco and Dragesco-Kernéis (1986) and more taxonomic data were provided. Unfortunately, those morphological data were unstable, including the arrangement of the frontoventral cirri, the number of membranes in AZM1, and the number of dorsal kineties, and even the number of transverse cirri was not stable. Thus, we suggest that Dragesco and Dragesco-Kernéis’ isolation is likely a complex Aspidisca species (Dragesco & Dragesco-Kernéis, 1986). The detailed morphological description of this species was provided by Song and Wilbert (1997) based on a population collected from China, and the molecular information was also supplied by the same laboratory several years later (Chen & Song, 2002). The arrangement of frontoventral cirri of a China population was very similar with the original population ( Figures 7Q–S ). Till now, the China population should be the most well-described population of A. steini (Song & Wilbert, 1997; Chen & Song, 2002). Therefore, we mainly compare the new species with the China population of A. steini in this work. It is very similar with A. shini sp. n. in the body size, number of dorsal ridges, number of membranelles in AZM1 and AZM2, and number of dikinetids in the dorsal kinety. However, the former can be clearly distinguished from the latter by the arrangement of the frontoventral cirri (the second line arranged in curved arc, name cirrus III/2 not in the same line as the others vs. approximately arranged in two parallel rows) ( Figures 7Q–S ; Table 3 ). Furthermore, A. steini is genetically distinct from A. shini sp. n. by 121-nucleotide substitutions in the SSU rRNA gene sequences. Based on the arrangement of frontoventral cirri, we conclude that only the China population of A. steini is identical to the original description; the populations from USA and Germany (Kiel) of A. steini should be synonymized with A. shini n. sp.

Aspidisca aculeata was discovered by Ehrenberg (1838) in Kiel and then has been frequently reported in the later literature; the latest comprehensive description was provided by Li et al. (2008). It is similar with A. shini sp. n. in cell size and number of dorsal ridges and membranelles in AZM1. However, the former can be separated from the latter by the number of membranelles in AZM2 (12 or 13 vs. 7–10) and dikinetids in the fourth dorsal kinety (9–11 vs. 6 or 7) ( Figures 7N–P ; Table 3 ). In addition, the two species differs from each other by 95-nucleotide substitutions in the SSU rRNA gene sequences (Li et al., 2008).

Aspidisca polypoda, originally found by Dujardin (1841), was redescribed by Song and Wilbert (2002) and Choi et al. (2020). It is very similar to A. shini sp. n. in body size, number of membranelles in AZM1, and number of dikinetids in the fourth dorsal kinety. However, these two species have a different number of dorsal ridges (7 or 8 in A. polypoda vs. 4 in A. shini sp. n.) and membranelles in AZM2 (10–12 in A. polypoda vs. 7–10 in A. shini sp. n.). Moreover, the frontoventral cirri of the new species are more stout and powerful than cirri of A. polypoda ( Figures 7X–Z , Table 3 ) (Song & Wilbert, 2002; Choi et al., 2020).

Aspidisca quadrilineata was firstly described by Kahl (1932); Song and Wilbert (2002) improved the former work by providing the detailed description and the silverline system. A. quadrilineata shares many morphological features with A. shini sp. n., such as cell size, number of dorsal ridges, and number of membranelles in AZM1. In addition to the thin and delicate frontoventral cirri, A. quadrilineata differs from A. shini sp. n. by the higher number of membranelles in AZM2 (12–14 vs. 7–10) and dikinetids in the fourth dorsal kinety (11–16 vs. 6 or 7) ( Figures 7Z1–Z3 , Table 3 ) (Song & Wilbert, 2002).

As shown in Figure 6 , the Order Euplotida is a well-supported clade that contains five families, namely, Euplotidae, Certesiidae, Aspidiscidae, Gastrociihidae, and Uronychiidae, which has been recognized in previous studies (Yi et al., 2009; Shen et al., 2010; Huang et al., 2011; Jiang et al., 2013; Lian et al., 2018). The internal relationships in the monophyletic family Aspidiscidae are far from robust, as indicated by the low support values. The previously proposed criteria to group members of the genus Aspidisca according to their freshwater or marine habitat, as proposed by Kahl (1932), or on the basis of number and pattern of frontoventral cirri, as proposed by Wu and Curds (1979), are not supported by molecular phylogeny. Indeed, the two new species here described were both collected from the coastal waters of the East China Sea and share the same frontoventral pattern of “polystyla-arrangement” with other species as well; however, they do not cluster together within the same clade of the phylogenetic tree. This finding suggests that the morphology-based classification of Aspidisca can be problematic, particularly when species have few morphological characters and even the use of silver staining or transmission electron microscopy does not add useful information. Therefore, SSU rRNA gene sequence phylogeny may represent an important tool that, together with morphology, could help to construct a more valuable classification system. Since only 12 SSU rRNA gene sequences of the approximately 60 Aspidisca morphospecies are available, the correlation between detailed morphological data and gene sequences for a greater number of taxa is thus required to determine more stable phylogenetic relationships within the family Aspidiscidae.