Loricate choanoflagellates (Acanthoecida) from warm water seas. VIII. Crinolina

The loricate choano ﬂ agellate genera Diaphanoeca Ellis and Crinolina Thomsen encompass a total of ten species. The majority of these are recorded from the warm water regions reported on here. A distinct morphological dichotomy characterizes the genus Diaphanoeca as currently circumscribed. The species distribute themselves within a ‘ D. grandis subgroup ’ and a ‘ D. pedicellata subgroup ’ distinguished on e.g., the position of the protoplast inside the lorica chamber and the elaboration of the anterior projections. We are, while awaiting in particular further molecular evidence, taking a conservative approach and abstain from dealing with the subgroup issue at the generic level. The examination of material from the warm water regions of the world ’ s oceans has resulted in the description of D. sargassoensis sp.n., D. pseudoundulata sp.n., and D. throndsenii sp.n., and a thorough re-examination of D. undulata . Species of Crinolina share multiple features with in particular the D. grandis species subgroup. It is yet relevant, both in a morphological and molecular perspective, to retain the genus Crinolina which remains unambiguously de ﬁ ned based on the posteriorly open lorica. A high level of agreement is found when contrasting morphological and molecular based phylogenetic schemes. (cid:1)


Introduction
In an ongoing effort (Thomsen and Østergaard 2019a, Thomsen and Østergaard 2019b, Thomsen and Østergaard 2019c, Thomsen and Østergaard 2019d, Thomsen and Østergaard 2019e, Thomsen et al. 2020a, Thomsen et al. 2020b) to provide a first comprehensive overview of warm water loricate choanoflagellate diversity, based on a traditional microscopical approach, we here deal with species of the tectiform genera Diaphanoeca Ellis, 1929, andCrinolina Thomsen, 1976.Species allocated to these genera are fairly large and, in some cases, even exceeding 100 mm in length.The spacious lorica is constructed from open mesh patterns of longitudinal and transverse costae and anteriorly terminated by spines or projections.The main criterion separating these genera is the elaboration of the posterior lorica end which is open in species of Crinolina.While no less than 8 species are currently allocated to the genus Diaphanoeca, only 2 species of Crinolina have been described so far.Apart from providing morphological and biogeographical details on the species encountered in samples from the warm water equatorial belt, the overall emphasis is here in particular on (1) the morphological distinctiveness of two subgroups of species of Diaphanoeca, viz. the 'D.grandis cluster' and the 'D.pedicellata cluster', and (2) the description of three new taxa.The process of re-evaluating, based on morphological evidence, species circumscriptions and the validity of the generic framework established over time is both supported and challenged by molecular evidence of no less than six species (Schiwitza and Nitsche 2020).

Material and Methods
The material that constitutes the background for this and a series of papers on warm water acanthoecid choanoflagellates was collected over a period of 35 years.The geographic origin of samples is recorded in Fig. 1.See Thomsen and Østergaard (2019a) for information on each of the collection sites and sampling campaigns.In order to substantiate the analysis of warm water specimens from the two genera in focus here, we have added material from Danish coastal waters (D. sphaerica/Fig.5a The general protocol for processing water samples for the light microscope (LM) and electron microscope (TEM/SEM) was according to Moestrup and Thomsen (1980) and Thomsen (1982).For details on sample processing, preparational issues and microscopes used see Thomsen and Østergaard (2019a).
The material examined here is dried, which means that the natural 3-D structures have collapsed to become 2-D structures leading to a partial dislocation of costal strips, and an artefactual expansion of in particular the lorica width.While several structures can still be measured with confidence, e.g., lorica height and the length of spines and pedicels, it does imply that certain values such as lorica diameter, typically at the level of the transverse costa(e), cannot be measured directly but only calculated from measurements of the circumference.This approach has been taken in the species descriptions below.
Efforts are made to make use of a concise terminology when describing lorica features and we follow the standards that have developed in the course of dealing with these organisms; see e.g., Leadbeater (2015;loc. cit.chapter 4 and glossary p. 278) and Thomsen and Buck (1991).
The unfortunate existence of an electron microscope specific problem causing negatives to appear horizontally flipped, was discussed in Thomsen and Østergaard (2019d).The evidence in favour of dealing with this problem as a purely technical issue is overwhelming.Scans of the affected negatives have accordingly been flipped Fig. 1.Map showing the approximate sampling sites for material reported here and MODIS sea surface temperatures (2003-2011 average).A circular dot refers to a single spot sampling, while a line or square indicates that samples were collected along extended transects.For further information see the materials and methods section in Thomsen and Østergaard (2019a).
horizontally to produce non-inverted images.In the current publication this applies to: Fig. 3m; Fig. 6i;Fig. 8a,b;Fig. 9b,c.In an attempt to extract more generalized biogeographical information from single point occurrences, we have tabulated species-specific biogeographical information making use of marine biogeographic global realms as defined by Costello et al. (2017).An analysis of 65,000 species of marine animals and plants (Costello et al. 2017) has led to the identification of 18 continent-shelf and 12 offshore deep-sea realms.

Results and Discussion
Crinolina Thomsen, 1976 The genus Crinolina is uniquely defined (Thomsen 1976) by a spacious lorica that is open posteriorly.There are two interior transverse costae and more than 10 longitudinal costae.Anterior projections comprise two longitudinal costal strips.Posteriorly the lorica is terminated by either short spines or projections that encompass 2-4 costal strips.The protoplast is suspended anteriorly in close affinity with the anterior transverse costa.A lorica mesh has, to the best of our knowledge, not been reported from species of Crinolina.Considering that such a lorica lining has been repeatedly observed in D. grandis (see e.g., Manton et al. 1981) and the fact that C. isefiordensis is in essence a posteriorly truncated version of the D. grandis lorica (see Fig. 2a, f) it is surprising, despite having access to numerous highresolution images, that we have been unable to pinpoint a lorica membrane in either species of Crinolina.The flagellum often assumes a characteristic curly position ('pig-tail'; Fig. 3k, see also e.g.Thomsen 1976, loc. cit Figs 12, 17) in species of Crinolina, a feature not regularly observed in any other loricate choanoflagellate species.The apparent absence of a lorica lining membrane, and the flagellar anomaly perhaps indicates that feeding behaviour in species of Crinolina somehow deviates from what is meticulously described from D. grandis (Nielsen et al. 2017).
C. isefiordensis Thomsen, 1976 (Fig. 3a-e) The material examined here (Fig. 3a-e) is largely identical to the Danish type material (Thomsen 1976) with reference to morphometric details (Table 1).However, there are marginally fewer longitudinal costae (13-14 versus 15-16), and a tendency in the warm water material to show hardly any sign of posterior lorica spines.The overall variability with reference to the number of longitudinal costae is thus 11-16 when including also observations made by Hara et al. (1997; Japanese and Taiwanese coastal waters; 11 long.costae; lorica height 19 mm) and Tong et al. (1998;Sydney Harbour, Australia;12-14 long. costae).Each longitudinal costa comprises 6-7 costal strips in parallel to what was observed in the Danish type material (Thomsen 1976).In specimens observed by Hara et al. (1997;loc. cit. Fig. 12) a longitudinal costa comprises only five longitudinal costal strips.
In summary the distribution pattern of C. isefiordensis (Table 2) is that of a quasi-cosmopolitan species which is currently only absent from low salinity areas (e.g., the inner Baltic Sea) and very high latitude Arctic and Antarctic habitats.
C. aperta (Leadbeater in Manton et al., 1975) Thomsen, 1976 (Fig. 3f-m) (Syn.: Diaphanoeca aperta Leadbeater in Manton et al., 1975) Crinolina aperta is differentiated from C. isefiordensis based on the presence of an extended skirt of posterior projections (Fig. 3i).The C. aperta lorica is overall less flaring and less constricted at the level of the anterior transverse costa when compared with C. isefiordensis.In the current material the posterior transverse costa is in C. isefiordensis on the average 172% larger than the anterior transverse costa as opposed to only 138% in C. aperta.
Warm water specimens of C. aperta are in comparison with high-arctic specimens (Manton et al. 1975) smaller and with fewer longitudinal costae and longitudinal costal strips (Table 1).The heavily silicified L-shaped transverse costal strips, which were a characteristic feature of the Resolute Bay (Arctic Canada) type material (Manton et al. 1975) and material from Antarctic realms (e.g., Buck 1981), are clearly absent in warm water populations of C. aperta (Fig. 3m).In specimens examined here all costal strips appear to be narrow rods.Manton et al. (1975) commented on that terminal longitudinal costal strips at both lorica ends are exceptionally slender.This is obviously also the case with reference to warm water specimens of the species (e.g., Fig. 3i, j).
Small colonies of C. aperta specimens were occasionally observed (Fig. 3g) as previously reported for this species by Buck and Garrison (1983); loc.cit.Fig. 32), Thomsen andØstergaard (2017), andEscalera et al. (2019;loc.cit.Fig. 1G-I).Neighbouring cells adhere alongside opposing longitudinal costae.A cell-to-cell connexion by filopodia has not been observed.Under ambient environmental circumstances and considering the overall geometry of the C. aperta lorica, this species has the potential to develop spherical colonies similar to what has been observed in the closely related species Diaphanoeca sphaerica (Thomsen 1976).Notice also that assembling the fringe of posterior lorica projections in C. aperta to a single point and thus closing the lorica posteriorly virtually converts C. aperta to D. sphaerica (see e.g., Fig. 2b, g).It can be further emphasized that these two species also share the Lshaped configuration of transverse costal strips (Thomsen 1976(Thomsen , 1982;;Manton et al. 1975).
There are consistent minor morphological differences between geographical clusters of C. aperta which are most likely driven by abiotic or biotic environmental factors.Evidence in favour of this possibility is provided by e.g.Woznica et al. (2016) and Leadbeater (2015, loc. cit. p. 133-34) discussing conspecificity between the loricate choanoflagellate species Savillea parva and S. micropora.In Arctic material there is consistently only a small overlap between costal strips from anterior projections (Manton et al. 1975;loc. cit. Fig. 28), whereas in Antarctic material Table 1.Size characteristics and morphological features in species of Crinolina and Diaphanoeca.The transverse costae are consecutively numbered starting from the anterior lorica end.In addition to the information directly provided in cited publications, the table information above is occasionally supplemented by measurements of illustrations provided.Numbers in round brackets are rare occurrences.Numbers in curly brackets are mean values.The generic type species are marked by asterisks.Type material of individual species are marked by double asterisks.(Thomsen et al. 1990; loc.cit.Fig. 28) the overlap is almost complete to the extent that the thin terminal costal strip is often concealed (see also Takahashi 1981a).The overall uniformity of costal strips in warm water material of C. aperta sets these specimens apart from both high latitude northern and southern specimens where in particular the transverse costal strips are deviant due to the presence of L-shaped terminal tips where the strip adjoins a longitudinal costa.
Crinolina aperta is repeatedly recorded (Table 2) from high latitude northern (N.American boreal realm) and high latitude southern sites (Southern Ocean realm).The species is here for the first time reported from multiple warm water sites.
Diaphanoeca fiordensis (Scagel and Stein, 1961) Norris, 1965 (syn. Microsportella fiordensis Scagel andStein, 1961) is described from light microscopy only (Fig. 2i).Considering the pronounced constriction at the level of the anterior transverse costa, the clear association between the collar and the anterior transverse costa, and the limited number of mid-lorica and posterior transverse costae, it is plausible that M. fiordensis can be relegated to a later synonym of D. grandis.
It is obvious when reviewing species of Diaphanoeca as presented in Fig. 2 that the genus comprises two quite distinct subgroups, i.e., a D. grandis cluster (Fig. 2c-i) and a D. pedicellata cluster (Fig. 2j-r).The main morphological characteristics that define the D. grandis subgroup (Table 1; Fig. 2) are the presence of anterior projections that always comprise two longitudinal costal strips, more than seven costal strips in each longitudinal costa, a separation of the anterior transverse costae by at least three longitudinal coastal strips, and a protoplast that is suspended centrally in the lorica chamber and with the tips of collar microvilli in contact with the anterior transverse costa (upper mid-lorica transverse costa in D. multiannulata).It must be emphasized that the true position and volume occupied by the protoplast inside a lorica can only be verified from high power light microscopy of living cells (see e.g., Fig. 5a).The manipulation of material in preparation for microscopy inevitably disturbs the appearance of the protoplast.It shrinks, and the collar microvilli snap away from costal elements (e.g., Fig. 4k).
It applies to members of the D. pedicellata cluster (Table 1; Fig. 2) that (1) the anterior projection comprises just a single costal strip, (2) the longitudinal costae typically encompass around 5 costal strips, (3) anterior transverse costae are separated by a maximum of two longitudinal costal strips, and further (4) that the protoplast is posteriorly positioned in the lorica.The morphological differences between the 'grandis' and 'pedicellata' subgroups are distinct (Fig. 2) and likely reflecting phylogenetic discrepancies across the matrix of species currently allocated to the genus Diaphanoeca that will eventually lead to the redistribution of species within two or more genera.
It is apparent that the Crinolina species (Fig. 2a, b) deviate from the D. grandis cluster only due to the fact that the lorica is here open at the posterior end, and also the 'pig-tail' configuration of the flagellum referred to above.Crinolina spp.forms a distinct unity in close association with the D. grandis sub-group.
The 'grandis' sub-group D. cylindrica Leadbeater, 1974 (Fig. 4a-e).This species was first described from the Mediterranean Sea (Leadbeater, 1973(Leadbeater, , 1974;; former Jugoslavia (Kastel Bay and Bay of Kotor) and Bay of Algiers) and has not been reported since then.The main morphological characteristics of the Mediterranean type material are the extended cylindrical lorica (65-110 mm) that has only two transverse costae and terminates in a simple pedicel.Cells examined here from West Australia (Fig. 4a-e) are slightly smaller (ca.50 mm) but otherwise identical (Table 1) to the type material (Leadbeater 1974).
D. grandis Ellis, 1929 (Fig. 4f-k).Observed in samples from all sampling sites.The material is characterized by a pronounced size variability (Fig. 4f-k; Table 1) and also by variation in the number of both longitudinal and transverse costae (Table 1).Diaphanoeca grandis is widely distributed (Table 2) and experimentally observed to be capable of surviving within significant ranges of major environmental stressors (Misiak et al. 2008).
D. multiannulata Buck, 1981 (Fig. 4l, m).A few specimens (Fig. 4l, m) that morphometrically match previous findings were observed in samples from West Australia (#8 on the Broome transect; sample collected from 100 m depth at a temperature of 23 °C; max.depth at station ca.675 m).Diaphanoeca multiannulata is of considerable importance in Southern Ocean pelagic ecosystems (see e.g.Escalera  2).More recently (Thomsen et al. 1995;Thomsen and Østergaard 2017) the species was also sparingly observed in samples from the Arctic.However, the apparent stenothermal nature of the species (with a clear preference for cold water habitats) is contradicted by our observations here but also by findings of D. multiannulata in association with stromatolites from Shark Bay, W. Australia (Al-Qassab et al. 2002;67 PSU;17 °C), and from the Pettaquamscutt River, Rhode Island, USA (Menezes 2005; 30 PSU; 24 °C).
D. spiralifurca Hara, 1996 (Fig. 5b-h).This species has 10-13 longitudinal costae and two transverse costae (Fig. 5b, f,  g).The tip of the anterior longitudinal costal strips is notched (Fig. 5h).The lorica is posteriorly terminated by an extended aggregated pedicel, that consists of intertwined longitudinal costal strips terminating in a claw-like hold-fast that may comprise at least four costal strips (Fig. 5b, d, g).When examining dried material, the lorica chamber longitudinal costae are often taking up helical paths (Fig. 5c, g).This is interpreted (Leadbeater 2015;Nitsche 2016) as an artefact caused by inherent tension from the aggregated stalk during cell preparation.Nitsche (2016; loc.cit.Fig. 1A) illustrates a living cell with a distinct perpendicular arrangement of costae in support of the fact that the D. spiralifurca lorica is in the living state much similar to D. grandis.Hara et al. (1996) refers to the presence in this species of a simple pedicel, and not the aggregated pedicel as illustrated here (Fig. 5d).However, when scrutinizing the type specimen (Hara et al. 1996; loc.cit.Fig. 7), the only image provided that shows a posterior lorica extension, it is apparent that the pedicel is not formed by a simple strand of costal strips, but rather costal strips running in parallel and thus basically identical to what is reported here.Diaphanoeca spiralifurca is sparingly recorded at low latitude regions of the world's oceans (Table 2).
Synonym: Campanoeca pedicellata (Leadbeater, 1972) Throndsen, 1974 This species was described from Danish coastal waters (Leadbeater 1972a) based on surface water samples collected late June 1971 in the vicinity of Frederikshavn, Northern Kattegat.Average values for sea water temperature at the time of sample collection is ca.15 °C and the salinity ca. 25 PSU.Morphometric details as listed for the type material is supplemented (Table 1) by measurements of the single illustration that accompanied the original description (Leadbeater 1972a;loc. cit. Fig. 20).The species was typified by a schematic drawing (Leadbeater 1972a; loc.cit.Fig. 1d).Details to be emphasized are the number of longitudinal costae ( 14), the fact that the anterior transverse costa is marginally smaller than the mid-lorica transverse costa, and further that the mid-lorica transverse costa crosses anterior to (ca. 1 mm) the junction between longitudinal costal strip number three and four counted from the anterior lorica end.Extensive nanoflagellate surveys have been carried out in Danish coastal waters during a period of four decades (Thomsen et al. 2016).The species was reported by Thomsen (1973) from the Isefjord but is otherwise rarely encountered.One of the few cells observed (from Southern Kattegat; collected late July 1989 at a depth of 14 m) is reproduced here (Fig. 6i) to supplement the information that can be extracted from the single Kattegat image that was part of the first description of this taxon.Morphometric details of this specimen are also included as a separate line in Table 1 to underpin the variability of this taxon as observed from the 'type locality'.Apart from adding some variability to the number of longitudinal costae (12-14) this specimen is similar to the type material also with reference to the positioning of the mid-lorica transverse costa.
Material examined here from West Australia, the equatorial Pacific Ocean, the Gulf of California, and the Sargasso Sea (Fig. 6a-h; Table 1), differ from the type material only with reference to the number of longitudinal costae (10-11).Although less pronounced than in both the type material and the Kattegat specimen (Fig. 6h), the anterior transverse costa is slightly more constricted than the mid-lorica transverse costa.The mid-lorica transverse costa crosses, also in specimens from the warm-water belt, shortly anterior to the junctions between longitudinal costal strip number three and four (counted from the anterior lorica end).
However, the material that globally most convincingly mirrors the D. pedicellata type material clearly originates from the North American Boreal realm (Table 2).In order to complete the survey of D. pedicellata variability and ren-der possible an easy inter comparison of subpopulations, examples of such specimens are provided in Fig. 6j (NE Greenland), Fig. 6k (Davis Strait/Disko Bay), and Fig. 6l (Baffin Bay/North Water Polynya), and morphometric details (Thomsen 1982) included in Table 2. D pedicellata cfr.(Fig. 7).Specimens obviously resembling D. pedicellata, however, differing in some minor features and with an overall diverging appearance when comparing in a bird's-eye view the collection of specimens of D. pedicellata cfr.(Fig. 7) with specimens of D. pedicellata (Fig. 6).
The most immediate factual morphological difference between D. pedicellata and D. pedicellata cfr.relates to mutual size differences between the anteriormost transverse costae.While the lorica is anteriorly expanding in size in D. pedicellata cfr.(Fig. 7; Table 1) it is slightly constricted in D. pedicellata (Fig. 6; Table 1).The number of longitudinal costae is furthermore smaller in D. pedicellata cfr.(9-10) in comparison with both D. pedicellata s. str.(12)(13)(14) and the warm water version of this taxon (Fig. 6a-h ; 10-11), and the distance between the anteriormost transverse costae somewhat reduced (roughly corresponding to the length of one longitudinal costal strip while in D. pedicellata s. str. the distance almost equals the length of two longitudinal costal strips.Molecular evidence will obviously be needed to further evaluate the taxonomic distance between D. pedicellata s. str.and the material presented here in Fig. 7. Diaphanoeca sargassoensis sp.nov.(Fig. 8).
Diagnosis: Lorica conical, 19-22 mm in length, comprising 14-16 longitudinal costae and three transverse costae with posteriorly decreasing diameters (12-15; 9-10, and 3-4 mm respectively).The distance between the two anterior transverse costae almost equals two longitudinal costal strips.The lorica is anteriorly terminated by projections each one costal strip long.Longitudinal costae comprise 5 costal strips and adjoin posteriorly with a pedicel (9-18 mm) that decrease in width towards the posterior tip.The protoplast is posteriorly positioned in the lorica chamber and the flagellum (c.20 mm) reaches out of the lorica.Division is tectiform.
Holotype: The specimen illustrated in Fig. 8d of the present work is fixed as holotype (ICZN 1999, Article 73.1.4).
Type locality: A 30 m depth water sample from the fluorescence maximum layer (17.9 °C, 34.4 PSU) collected 9 April 2014 from the Central Sargasso Sea (27.3°N, 59.3°W).
Etymology: The species-group name is chosen to reflect the massive occurrence of this species in Sargasso Sea waters.
The species described above was a loricate choanoflagellate community dominant form in the vast majority of the Sargasso Sea samples analysed.While there are obvious similarities between this form and D. pedicellata sensu stricto (Fig. 6i-l), D. sargassoensis is yet clearly distinguished by a consistently larger number of longitudinal costae, and the fact that the D. sargassoensis lorica is flaring and has its maximum diameter at the level of the anterior transverse costa.
Fig. 6. a-l.Diaphanoeca pedicellata LM (a-h, j-l; phase contrast) and TEM (i) whole mount micrographs from West Australia (a, d, h), the equatorial Pacific Ocean (b, c, e), the Gulf of California (f), the Sargasso Sea (g), the Kattegat, Denmark (i), North East Greenland (j), Disko Bay, West Greenland (k), and the North Water Polynya, Greenland (l).(a-h) Selected warm water specimens with 10-11 longitudinal costae; notice the extended pedicel (h); (i) Danish specimen with 12 longitudinal costae; (j-l) Greenlandic specimens with 15, 14 and 16 longitudinal costae respectively.The scale bar (h) applies to all light micrographs.While D. sargassoensis was not reported from any other of the warm water sites visited, it is obviously closely related to material from the Southern Ocean that has previously been ascribed to D. pedicellata (e.g., Thomsen et al. 1990;loc. cit. Figs. 33-36).See also Buck and Garrison (1983); loc.cit.Fig. 30), Marchant (1985), and Marchant (2005;loc. cit. Fig. 13.4.c).In order to render possible an easy morphological comparison between Sargasso Sea and Southern Ocean specimens, we have included here also specimens from the Weddell Sea, Antarctica (Fig. 8a, b).Morphometric details of the Southern Ocean material are also included in a separate line in Table 1.
The flaring nature of the anterior lorica chamber is a shared feature between D. sargassoensis (Fig. 2o) and D. dilatanda (Fig. 2n).The main difference between these two taxa is overall lorica size (lorica chamber length: 30-35 mm in D. dilatanda versus 19-22 mm in D. sargassoensis; anterior transverse costa diameter: 18 mm in D. dilatanda versus 12-15 mm in D. sargassoensis).While the anterior part of the lorica chamber is virtually identical in both species the addition of one extra costal strip to each longitudinal costa in the posterior part of the lorica chamber creates significant overall morphological differences across the two species.
D. dilatanda Thomsen and Østergaard, 1917 (Fig. 2n).This species, which is so far only recorded sparingly from Greenlandic waters, differs in overall size and by having a conspicuously flaring lorica chamber (Fig. 2n).See Thomsen and Østergaard (2017) for further details.D. undulata Thomsen, 1982 (Fig. 8).The circumscription of, or rather the subsequent established tradition for applying the species epithet 'undulata', is not entirely satisfactory.The West Greenland type material (Thomsen 1982) is characterized by a middle transverse costa that is displaced from the anterior transverse costa by less than the length of a longitudinal costal strip, and also by comprising less heavily silicified strips that tend to be organized in an undulating band (hence the specific name).Anterior projections are tapering and terminate anteriorly in a sharp-pointed spine.Specimens of D. undulata s. str.have 11 or rarely 12 longitudinal costae that each comprises five costal strips.The pedicel comprises costal strips that are more heavily silicified than costal strips elsewhere.
Material that acceptably matches the type material has been later reported from California (Thomsen et al. 1991;loc. cit. Fig. 20), Southampton, UK (Tong 1997a; loc.cit.Fig. 7B, C), Pettaquamscutt River estuary, Rhode Island, USA (Menezes 2005; loc.cit.Pl. 6B), NE Greenland, NW Greenland (Thomsen and Østergaard 2017;loc. cit. Fig. 11 H, I; see also Fig. 9d in the current publication).Notice that Menezes (2005) reports the presence of 10-11 longitudinal costae in the Rhode Island material.Mediterranean Sea material examined by Leadbeater (1973;loc. cit. Pl. 20e, f) and referred to as D. pedicellata is most likely better classified as D. undulata based on morphometric details and also an overall similarity to cells observed in e.g., Californian offshore samples (Thomsen et al. 1991).In order to further substantiate the circumscription of D. undulata, we have included here images of Danish costal water specimens (Fig. 9a-c, e-g; see also Table 1) and a single specimen from the Baffin Bay north water polynya (Fig. 9d), that will help future identification of D. undulata sensu stricto and provide here a relevant  frame of reference for the description of D. pseudoundulata (see below).
Cells identified as D. undulata and originating from lower latitudes (Tong 1997b, loc. cit. Fig. 3b;Hara et al. 1997, loc. cit. Fig. 14;Lee et al. 2003, loc. cit. Fig. 2a) are similar to the material presented here as D. pseudoundulata sp.nov.(Fig. 10), i.e., with anterior transverse costae that are interspaced the full length of a longitudinal costal strip, and with no apparent differences in the appearance of the anterior transverse costae.
Syn.: Diaphanoeca undulata Thomsen, 1982, in Tong 1997b (loc.cit.Fig. 3b), Hara et al. 1997 (loc. cit.Etymology: The species-group name is chosen to reflect the previous taxonomic history of this taxon. Diaphanoeca pseudoundulata displays a certain size differentiation across the four regions where it has been found (Fig. 11).In general, the Indian Ocean specimens (Andaman Sea and West Australia) are markedly larger than specimens from the Pacific Ocean and the Gulf of California.
While the new taxon described here is morphometrically very similar to D. undulata, there are still a number of subtle differences, that in our opinion justifies the attempt to separate the two taxa.The apparent lack of costal strip differentiation in D. pseudoundulata and fairly consistent differences in both the number of longitudinal costae (D. undulata: 11/D.pseudoundulata: 10) and the number of longitudinal costal strips in each costa (D. undulata: 5/D.pseudoundulata: 4-( 5)) are among the lorica characteristics that contribute to creating fairly obvious differences in overall appearance (e.g., lorica overall shape and the distinct and accurate layout of the anterior lorica end in D. pseudoundulata) when comparing e.g., the light micrographs provided of the two taxa (D. undulata: Fig. 9d-g/D.pseudoundulata: Fig. 10c-h).Molecular tools will obviously be needed to more conclusively underpin (or reject) the approach taken here based on morphometric details only.The possibility exists that D. undulata and D. pseudoundulata are con-specific yet markedly shaped by their preferred habitats.Diaphanoeca pseudoundulata is thus biogeographically restricted to warm water habitats (Table 2), while D. undulata is more widespread yet with a preference for low temperature realms (Table 2).This species is closely resembling D. pseudoundulata yet differing because of a consistently lower number of longitudinal costae (8 versus 10) and the fact that each of these unchangingly comprises 4 costal strips.While the D. pseudoundulata middle transverse costa (Fig. 10a) is exactly in line with the junctions between longitudinal costal strip number two and three (counted from the anterior lorica end), this costa is in D. throndsenii shifted forwards by c. 2 mm equal to one-fifth costal strip length (Fig. 12a, b/arrows).The D. throndsenii lorica (Fig. 12) is markedly larger than D. pseudoundulata (Fig. 10c-h) occasioned by the fact that costal strips are twice as long in D. throndsenii.
The morphometric analysis of material examined here in our opinion supports the inclusion of D. throndsenii within the D. pedicellata subgroup of species of Diaphanoeca through stepwise similarities across a string of taxa (i.e., D. throndsenii > D. pseudoundulata > D. undulata > D. pedicellata).It is yet obvious that the differences between D. throndsenii and D. pedicellata when directly compared (e.g., Fig. 2j and Fig. 2r) are quite significant and point to that once molecular evidence becomes available, the taxonomic framework as here established will likely have to be thoroughly revised.

Conclusions
The morphometric analyses of the genera Crinolina and Diaphanoeca, centered around the warm water species encountered, has on the one hand corroborated the validity and relevance of separating species with an open lorica end (Crinolina) from the remainders.On the other hand, it has also been pointed out that each of the Crinolina species show particular morphological affinity to specific species of Diaphanoeca (i.e., C. aperta/D.sphaerica and C. isefiordensis/D.grandis respectively).The molecular evidence available (D. grandis, D. sphaerica, C. isefiordensis) closely groups D. grandis and D. sphaerica while C. isefiordensis is more remotely positioned and thus emphasizing the integrity of the genus Crinolina (Nitsche et al. 2017).
The morphological distinction between the D. grandis and D. pedicellata subgroups of species, based on a suite of morphological features (i.e., single costal strip projections, two or fewer longitudinal costal strips separating the anterior transverse costae, and a posterior positioning of the protoplast in the lorica chamber in the D. pedicellata subgroup) is in our perspective quite substantial.However, the positioning of D. pedicellata (HQ237460) in recent phylogenetic diagrams (Schiwitza et al. 2019, Schiwitza andNitsche 2020), where it clusters in close connection with D. grandis and D. sphaerica and relegates D. spiralifurca, which is otherwise morphologically closely related to in particular D. grandis, to a more remote positioning, does none the less at first sight leave significant doubt with respect to the phylogenetic validity of the Diaphanoeca subgroups identified here.The strain (HQ237460) was a new sequence generated for the Nitsche et al. (2011) paper on higher level choanoflagellate phylogeny.The species was here identified as Diaphanoeca sp. and an illustration of the organism provided (Nitsche et al. 2011;loc. cit. Fig. 5).It is evident beyond doubt that the illustration does not depict D. pedicellata but rather a member of the D. grandis cluster of species, most likely D. spiralifurca.The anterior projections comprise two costal strips, the protoplast is suspended anteriorly, the general outline of the lorica is highly reminiscent of D. grandis, and the stalk is solid (aggregated?) as is known only in D. spiralifurca.The fact that the sequence (HQ237460) is listed in GenBank under the name D. pedicellata must, based on the evidence available (Nitsche et al. 2011), be considered an error.
Ignoring the HQ237460 sequence when examining recent phylogenetic trees (Schiwitza et al., 2019, Schiwitza andNitsche 2020) leaves D. undulata (Nitsche et al. 2017) as the only D. pedicellata cluster representative with trustworthy sequence data.It takes up an isolated position and branches out prior to C. isefiordensis and the D. grandis cluster species, thus basically confirming the species matrix as resolved by the morphometric analysis.However, before jumping to far-reaching taxonomic conclusions, it is important to emphasize that the D. pedicellata cluster of species can be further divided into two subgroups (i.e., D. pedicellata, D. dilatanda, D. sargassoensis/D.undulata, D. pseudoundulata, D. throndsenii) where the latter group comprising D. undulata is obviously morphometrically furthest away from the generic type species (D. grandis).It therefore remains relevant, despite the clarification with reference to D. pedicellata sequence data, to take a taxonomic conservative approach and refrain from splitting prematurely the entire Diaphanoeca species matrix, while waiting for trustworthy molecular sequence data for in particular D. pedicellata and D. sargassoensis.

Fig. 3 .
Fig. 3. a-m.TEM (m) and LM whole mounts (phase contrast) of Crinolina isefiordensis (a-e), and C. aperta (f-m) from West Australia (a, h, i), the Gulf of California (b-d), the Andaman Sea, Thailand (e, l, m), the Caribbean Sea (f), the Sargasso Sea (g), and the equatorial Pacific Ocean (j, k).(a-l) Selection of specimens from various geographic regions; the thin anteriormost longitudinal costal strip is visible in most images; notice colony formation (g) and the 'pig's tail' curly flagellum (k); (m) Detail of posterior transverse costa to show the complete absence of 'L-shaped' elaborations of transverse costal strips where these join the longitudinal costae.The scale bar (a) applies to all LM images except (g).
Fig. 4. a-m.LM whole mounts (phase contrast and DIC (j)) of Diaphanoeca cylindrica (a-e), D. grandis (f-k), and D. multiannulata (l, m) from West Australia (a-e, g, h, k-m), the Gulf of California (f, i), and the Andaman Sea, Thailand (j).(a-m) Selection of specimens from various geographic regions.The scale bar (a) applies to all LM images.

Fig. 5 .
Fig. 5. a-h.LM (a-c, g; phase contrast and DIC (a)) and TEM (d-f, h) whole mounts of Diaphanoeca sphaerica (a) and D. spiralifurca (b-h) from the Kattegat, Denmark (a), West Australia (b), and the Andaman Sea, Thailand (c-h).(a) Living cell to show the actual position of the protoplast; (b, c, g) Complete loricae; (d) Detail of twisted aggregated pedicel; notice the holdfast comprising four costal strips; (e) Detail from (f) of anterior transverse costa; (f) Complete lorica; (h) Bifurcated anterior tip of longitudinal costa; detail from (f); The scale bar (b) also applies to (c, g).

Fig. 7 .
Fig.7.a-h.Diaphanoeca pedicellata cfr.LM whole mounts (phase contrast and DIC (e)) from the Sargasso Sea (a-d), the Mediterranean Sea (e), the equatorial Pacific Ocean (f, g), and West Australia (h). (a-h) Selected specimens to show the variability encountered across all regions sampled.Notice the extended pedicel in (h).The scale bar (c) applies to all images.

Fig. 8 .
Fig. 8. a-h.Diaphanoeca sargassoensis TEM (a, b) and LM (c-h; phase contrast) whole mount micrographs from the Weddell Sea, Antarctica (a, b) and the Sargasso Sea (c-h).(a, b) Complete cells with 16 (a) and 15 (b) longitudinal costae; (c-h) Selected Sargasso Sea specimens to show the morphological variability encountered.The scale bar (e) applies to all LM images.

Fig. 9 .
Fig. 9. a-g.Diaphanoeca undulata SEM (a), TEM (b-c) and LM (d-g; phase contrast) whole mounts from the Isefjord, Denmark (a-c, e-g), and the North Water Polynya, Greenland (d).(a, b) Complete specimens; the exterior position of longitudinal costae is evident in (a); (c) High magnification to show the anterior spiny tips of longitudinal costal strips; notice that also the penultimate longitudinal costal strips have pointed tips (arrows); (d-g) Specimens with eleven longitudinal costae and a species-specific wavy appearance of the mid-lorica transverse costa.The scale bar (d) applies to all light micrographs.

Fig. 10
Fig. 10.a-h.Diaphanoeca pseudoundulata TEM (a, b) and LM whole mounts (c-h; phase contrast and DIC (d)) from the Andaman Sea, Thailand (a-d), the equatorial Pacific Ocean (e), West Australia (f, h), and the Gulf of California, Mexico (g).(a) Type specimen; notice the accumulation of costal strips in the collar region; (b) Empty lorica; (c-h) Selected specimens to show the variability encountered across all regions sampled.The scale bar (f) applies to all light micrographs.
Fig. 14) and Lee et al. 2003 (loc.cit.Fig. 2a) Diagnosis: Lorica conical, 18-23 mm in length, comprising 10 longitudinal costae and three transverse costae.The two anterior transverse costae are of almost the same size (diam.10-15 mm), while the posterior transverse costa is much smaller (diam.4-6 mm).The distance between the two anterior transverse costae exactly equals the length of one longitudinal costal strip.Longitudinal costal strips and transverse costal strips from the anterior ring join together forming E-joints.The lorica is anteriorly terminated by projections each one costal strip long.Longitudinal costae comprise 4-(5) costal strips and adjoin posteriorly with a pedicel (9-27 mm) that decrease in width towards the posterior tip.All costal strips are narrow rods.The protoplast is posteriorly positioned in the lorica chamber and the flagellum (c. 12 mm) reaches out of the lorica.Division is tectiform.Holotype: The specimen illustrated in Fig. 10a of the present work is fixed as holotype (ICZN 1999, Article 73.1.4).Type locality: A surface water sample (c.28 °C, 35 PSU) collected 22 September 1981 from the PMBC (Phuket Marine Biological Center) pier (7°48 0 05 00 N, 98°24 0 21 00 E).

Fig. 12 .
Fig. 12. (a-f) Diaphanoeca throndsenii LM whole mounts (phase contrast) from the Andaman Sea, Thailand (a-c), West Australia (d-f), the Sargasso Sea (g, h), and the Gulf of California (i).(a) Type specimen; (b-i) Selected specimens to show the variability encountered across all regions sampled.The scale bar (d) applies to all LM images.
Diaphanoeca throndsenii sp.nov.(Fig.12).Diagnosis: Lorica conical, 27-30 mm in length, comprising 8 longitudinal costae and three transverse costae.The two anterior transverse costae are of almost the same size (diam.16.7-17.8mm and 14.5-16.7 mm), while the posterior transverse costa is much smaller (diam.5.6-7.5 mm).Longitudinal costal strips and transverse costal strips from the anterior ring join together forming E-joints.The distance between the two anterior transverse costae is slightly less than the length of one longitudinal costal strip.The lorica is anteriorly terminated by projections each one costal strip long.Longitudinal costae comprise 4 costal strips and adjoin posteriorly with a pedicel formed by a single longitudinal costal strip.Longitudinal costal strips are 7-8 mm in length.All costal strips are narrow rods.Both the anterior projections and the pedicel are tapering from base to tip.The protoplast is posteriorly positioned in the lorica chamber.Division is tectiform.Holotype: The specimen illustrated in Fig. 12a of the present work is fixed as holotype (ICZN 1999, Article 73.1.4).The species-group name is chosen to acknowledge contributions by Dr. Jahn Throndsen, University of Oslo, Norway, to choanoflagellate systematics.