Navilithus altivelum: a remarkable new genus and species of deep photic coccolithophores

A very distinctive new deep-photic coccolithophore is described from the NE Indian Ocean. The new species is trimorphic with: 200–300 body coccoliths bearing low spines attached by narrow stems to a basal narrow-rimmed placolith structure; up to 18 circum-flagellar coccoliths with tall sail-like spines; and up to 22 coccoliths with moderately elevated spines occurring both around the circum-flagellar coccoliths and antapically. These features make the coccolithophore unique and require placement in a new species and genus. The basal structure, however, shows similarities to a recently recognized group of narrow-rimmed placoliths. Hence, the new coccolithophore provides some support for this grouping as a significant addition to our understanding of coccolithophore biodiversity, and potentially an explanation for a set of anomalous molecular genetic results. In addition the new taxon provides further evidence that the deep-photic coccolithophore community is more diverse than has been assumed.


INTRODUCTION
Extant coccolithophores have been intensively studied over the past two decades and their taxonomy is probably better established than that of any other phytoplankton group. Nevertheless, new taxa continue to be recorded, especially from the deep-photic zone Kleijne et al., 1991;Jordan & Chamberlain, 1993;Hagino & Okada, 1998;Cros & Fortuño, 2002;Young et al., 2003). The number of very distinctive but rare coccolithophores is a curious feature of the deep photic zone. Another such unusual new coccolithophore species was found during examination of samples from the Indian Ocean off SW Indonesia. The ecological study of these samples will be described in detail elsewhere; this paper describes the new species, discusses its affinities and its implications for understanding of coccolithophore diversity.

MATERIAL AND METHODS
The samples were collected during a geological cruise, So139, of the R/V Sonne in the Eastern Indian during January-February 1999 (Beiersdorf et al., 1999). Samples were collected at seven stations southwest of Sumatra and Java, over the Java Trough ( Fig. 1). At each station depth transects of six or seven samples were collected using a rosette sampler with attached CTD device for recording salinity, temperature and depth (Fig. 2). Nutrient and chlorophyll data were not collected on this cruise. Sample depths were varied based on the temperature data in order to ensure good sampling, especially of the thermocline population. For each sample 5 l of seawater were filtered onto fleecesupported regenerated cellulose filters (Sartorius , 50 mm diameter, 0.45 µm pore width). The samples were counted using a standard electron microscope at the BGR, with more than 300 specimens being identified and counted per filter, except in the deepest samples, where a smaller number was counted due to the scarcity of coccospheres. In total 33 coccospheres of the new species were found during the routine counts. Selected filters on which the new species was more common were then re-examined Diagnosis. Cellula coccolithifera polymorpha cum coccosphaera composita ex coccolithis angustimarginatis placolithis cum plano areae centralis convexo proximale et cum spinis robustis suffultis ab caule humili.
Polymorphic coccolithophore with coccosphere formed of narrow-rimmed placolith coccoliths with central area plate which is convex proximally and with robust spines supported by a low stem. Species of Navilithus with coccosphere composed of c. 200-300 body coccoliths with low triangular spines, 10-20 coccoliths with tall sail-like spines and a variable number of coccoliths with tall triangular spines.
Type specimen. Specimen illustrated in NHM SEMs 22 (Pl. 1,Pl. 3,figs 6,7,9,13,14,15;Pl. 4,figs 3,4). This specimen is located on NHM coccolithophore collection SEM stub 483. Description -Coccospheres. All observed coccospheres (Pls 1, 2) are collapsed so the original shape cannot be determined with certainty; however, these scatters have length to width ratios of 1.5 to 1.0, average 1.15 (30 measurements) so it is reasonable to infer that the original coccospheres were sub-spherical in shape. The collapsed coccospheres are 5 to 11 µm across, average 7.8 µm, suggesting an original coccosphere diameter of c. 5-8 µm, assuming some of the variation is due to variable degree of scatter of coccoliths during collapse of the coccospheres. The coccospheres are trimorphic with body coccoliths, moderately elevated coccoliths and sail coccoliths (Pl. 3).
The number of body coccoliths observed range from 100 to 170 (counts on seven specimens, including the largest and smallest observed coccospheres). This suggests that the coccospheres contain 200-300 body coccoliths, on the assumption that about half the coccoliths are visible on a collapsed coccosphere. Tall sail-coccoliths are always clustered together on the coccosphere but the number observed is highly variable, from 18 to 0 (average 7.6, 30 counts). This variation is interpreted as being largely due to variable orientation of collapse of the coccospheres, with minimum values when the sail-like coccoliths are on the concealed side of the coccosphere and maximum values when they are all visible. The typical number of sail coccoliths per coccosphere is thus likely to be about 15. The number of moderately elevated coccoliths observed is highly variable, from 0 to 22 (average 4.1, 30 counts). They mostly occur around the outside of the group of sail-coccoliths but there are also often a few of these coccoliths well-separated from the circum-flagellar area and in at least some cases they appear to form a cluster of antapical coccoliths.
On the collapsed coccospheres, fields of coccoliths seen in proximal view are normally very obscured. This is probably a result of the cell membrane and other organics from the cell coating the coccoliths. No evidence of flagella was seen, but this is not necessarily significant, since they are only occasionally visible in SEMs of coccolithophores.
The rim is narrow (0.08-0.12 µm). Elements in the distal shield are not usually imaged clearly but there seem to be 15-20. Sutures are sub-radial, with weak clockwise obliquity; the elements do not appear to be imbricated. The tube is narrow, height (0.09-0.15 µm) approximately equal to the distal shield width. The proximal shield is similar in width to the distal shield or slightly narrower. Both shields are essentially horizontal. The central area plate is not imaged clearly in any specimens, possibly because of an organic coating, but appears to be composed of numerous irregular elements, often with a small central pore. The proximal surface of the coccolith is convex, i.e. curves toward the cell. This morphology is unusual, as coccoliths normally have concave bases, but it occurs in too many specimens for it to be a result of deformation.
The spine is low (0.4-0.7 µm), triangular in profile. It is attached to the central area plate by a very narrow basal stem. Above this it immediately widens to its maximum width then tapers upward. It is formed of numerous rhombic, vertically elongated elements arranged in irregular tiers, the lowermost tier is higher than the others. The elements show flat rhombic faces and are probably modified calcite rhombohedra. In plan view the spine is approximately diamond-shaped, aligned parallel to the axes of the coccolith.

Description -Moderately elevated coccoliths (Pl. 3, figs 3-6).
These are similar to the body coccoliths but have higher spines (0.7-1.4 µm). The spines have similar structure to those of the body coccoliths, including the presence of a basal cycle of relatively tall elements and attachment by a basal stem. They are diamond-shaped in plan view and elongated parallel to the long axis of the coccolith. The spine structure becomes more regular upwards and cross-shaped in plan view. The basal structure appears to be identical to that of the body coccoliths.

Description -Circum-flagellar sail-coccoliths (Pl. 3, figs 1-2).
These have similar bases to the other coccoliths but the spine is much taller (1.3-2.2 µm, average 1.84 µm, from 15 measurements) and is flattened perpendicular to the long axis of the coccolith. The spine is cross-shaped in plan view but the arms extending parallel to the long axis are very short. The basal stem of the spine is rather higher than in the other coccolith types so that the base of the sail is elevated slightly above the rim. The spine structure is regular, with each blade of the sail being formed of two series of elements originating from either side of the blade and interfingering toward the outer margin of the sail (Fig. 3).
The basal structure is similar to that of the body coccoliths but perhaps slightly smaller, and the central pores are larger and more consistently developed.

Ecological affinities
Specimens of Navilithus were found in four of the seven sites sampled and in a total of six samples (out of 45). All the sites contained low abundance coccolithophore communities, with surface assemblages either dominated by Gephyrocapsa oceanica or Umbellosphaera irregularis or a diverse mixture including the two named species plus Palusphaera vandelii, Ophiaster hydroideus, Syracosphaera spp. and Calciosolenia spp. with no clear dominance of a single species. There was a well-defined thermocline in the area at c. 100m and below this a deep-photic nannoplankton community was developed dominated by Florisphaera profunda, but also with high abundances of Algirosphaera robusta, Gladiolithus flabellatus and Oolithotus antillarum. In addition Gephyrocapsa oceanica persisted in high abundances in the deep samples, although some at least of this population was probably settling out from the upper photic zone, it is particularly noticeable that the deepest samples recovered (from c. 200 m) typically had reduced abundances of Florisphaera profunda.
All the Navilithus specimens observed came from samples within or below the thermocline and with a significant abundance of deep photic zone species. Hence, one can infer confidently that this is a true obligate deep-photic species. The thermocline is usually abrupt in this area, with temperatures dropping from the surface values of 28-30(C to deep values of 11-13(C in an interval of less than 50 m. Consequently thermocline-dwelling species inevitably encounter a wide range of temperatures. This applies to Navilithus and the samples containing the species came from water samples ranging in temperature from 12.8-21.6(C. Similarly the thermocline is marked by a rather abrupt increase in salinity from surface values of sometimes less than 33 PSU to deep values of c. 35 PSU and samples with Navilithus had salinities of 34.35 to 34.72 PSU. The depth of Navilithus-bearing samples varies from 100 m to 160 m, again this seems to be largely controlled by the varying thermocline depth at different stations. There is no obvious difference between the communities present in the sub-thermocline samples with and without Navilithus specimens. In addition, Navilithus specimens occur at sites with rather different and even contrasting upper photic zone communities. It is rather curious that such a distinctive species should be found in what is a very well-studied group of organisms. The coccospheres are small and could be overlooked, especially on cellulose filters as used here, but they would be readily visible on polycarbonate filters. Moreover, there have been several very detailed studies by taxonomically interested workers of sample sets, including deep photic zone taxa (e.g. Okada & Honjo, 1973;Okada & McIntyre, 1979;Hagino et al., 2000;Cortes et al., 2001;Cros & Fortuño, 2002;Andruleit et al., 2005) and this species has never been recorded before, even in open nomenclature. This was confirmed by showing the images to many nannoplankton specialists via the coccoliths listserver (coccoliths@morgan.ucs. mun.ca) and at the International Nannoplankton Association conference in 2004. None of the consulted specialists had ever seen this species. Similarly, several other very distinctive deep photic zone species have been described with very sporadic occurrences, notably including Gladiolithus striatus, Placorhombus ziveriae, Tetralithoides quadrilaminata, Turrilithus latericioides, numerous informally described Papposphaeraceae (Cros & Fortuño, 2002) and several small Rhabdosphaeraceae (Young et al., 2003). The deep photic coccolithophore flora is characteristically dominated by Florisphaera profunda and this has given an impression of homogeneity in the flora. However, as Quinn et al., (2005) have shown, F. profunda itself is morphologically highly variable and it is suspected that it may constitute a plexus of pseudo-cryptic species. This variability in F. profunda, combines with variability in the secondary and rare species constituting the deep-photic assemblage. So, this assemblage may be much more taxonomically heterogeneous than has been assumed and this heterogeneity may be ecologically informative. Taken together, the new evidence of cryptic diversity in Florisphaera, and of a wider diversity of accessory species than has been assumed means that the deep-photic nannoplankton assemblage should be seen as heterogeneous and attempting to discover the controls on this heterogeneity might be highly productive.

Taxonomic affinities
Extant coccolithophore taxonomy is comprehensively reviewed in Young et al. (2003). The majority of known coccolithophores can be assigned confidently to a few major groupings based on coccolith structure and these groupings have been wellsupported by available molecular genetic data (Saez et al., 2004). There are, however, a number of anomalous genera which do not fit into any of the major groupings. Young et al., (2003) argued that these could be subdivided into two groups. First, narrow-rimmed muroliths, including Vexillarius cancellifer Jordan & Chamberlain, 1993, Wigwamma Manton et al., 1977 (several species), Picarola margalefii Cros & Estrada, 2004, and an undescribed genus. These forms show some affinity to the Papposphaeraceae. Secondly, narrow-rimmed placoliths, including Tetralithoides quadrilaminata (Okada & McIntyre, 1977, Turrilithus latericioides Jordan et al., 1991, Placorhombus ziveriae Young & Geisen, in Young et al. (2003, and the heterococcolith phase of Calyptrosphaera sphaeroidea Schiller, 1913, which was recently described by Noel et al., (2004. Navilithus certainly does not show the characteristic features of any of the main groups, such as the Syracosphaeraceae, Rhabdosphaeraceae or Coccolithaceae, in terms of either rim or central area structure. It does, however, show the basic narrowrimmed placolith rim morphology which Young et al. (2003) used as the primary basis for grouping this set of coccoliths Explanation of Plate 4. Comparison of N. altivelum and other narrow-rimmed placolith coccolithophores. Scale bars 1 µm, except fig. 1. fig. 1. Placorhombus ziveriae coccosphere, scale bar 10 µm. SEM NHM JY 144-04. fig. 2. P. ziveriae, detail of coccoliths SEM NHM JY127-03. fig. 3. Navilithus altivelum, detail of coccoliths SEM NHM JY205-21g. fig. 4. Navilithus altivelum, detail of coccoliths SEM NHM JY205-21h. fig. 5. Turrilithus latericiodes, detail of coccoliths SEM NHM JY200-13a. fig. 6. Tetralithoides quadrilaminata, detail of coccoliths (image courtesy of R. Jordan). fig. 7. 'Calyptrosphaera' sphaeroidea HET, detail of coccoliths (image courtesy of M-H Noel). fig. 8. 'Calyptrosphaera' sphaeroidea HET coccosphere (image of M-H Noel). fig. 9. Turrilithus latericiodes coccosphere SEM NHM JY SEM NHM JY200-13. fig. 10. Tetralithoides quadrilaminata coccosphere SEM NHM JY128-16. incertae sedis. So, it is worth considering in detail the degree to which it shows affinity with this group, and to review the degree to which this group shows sufficient common characteristics to be likely to be monophyletic. The characteristics of this group are discussed below and the other species of the group are illustrated in Pl. 4.
Rim structure. In all the species considered, the rim has a simple morphology with narrow distal and proximal shields connected by a thin sub-vertical tube. It is not possible to be sure of the exact structure but it appears in each case to be formed of a single cycle of elements, rather than two interlocked series of elements, as in the Coccolithaceae, Pleurochrysidaceae or Syracosphaeraceae. The sutures are radial or show weak clockwise obliquity on the distal shield, are vertical in the tube, and less certainly are radial in the proximal shield. Light microscope observations are only available for Tetralithoides, which appears to be formed of V-units (J. R. Young, personal observation).
Rim structure is normally the best single morphological indicator of affinity for heterococcoliths (e.g. Perch-Nielsen, 1985;Bown & Young, 1998;Young et al., 2004), and these rims certainly appear similar to each other and unlike those of other coccoliths, but the absence of detailed structural or crystallographic observations means that this rim similarity cannot be considered decisive evidence of affinity.
Central area structures. All the included species have central area structures which are clearly disjunct from the rim, i.e. formed of separate crystal units. In Placorhombus and Tetralithoides the central structure is four large plates, striate in the case of Placorhombus. In C. sphaeroidea HET, Navilithus and probably in Turrilithus, the central area is floored by a mass of small irregular crystallites. This floor is remarkably similar in appearance in Navilithus and C. sphaeroidea HET, notably in proximal view, with distinct convexity being shown. This convexity of the proximal surface is very unusual and is also shown, to a lesser extent, by Placorhombus. Only Navilithus and Turrilithus show processes and these are different in shape but exhibit some similarities in being tall and massive in relation to a rather small coccolith base. It is noteworthy that each side of the quadrate tower spines of Turrilithus is formed of two offset series of sub-horizontally elongated elements, thus in total the structure is formed of eight series of offset elements, as are the sail spines of Navilithus, possibly indicating some structural relationship. No other coccolithophores have processes with similar structure.
Coccosphere/cell. All the species form coccospheres formed of numerous coccoliths, from 60 to 500. They show few other common characteristics, however. Most are spherical but Placorhombus coccospheres are tapering cylinders. Young et al. (2003) included monomorphism as a feature of the group, even though Placorhombus is weakly dimorphic (small coccoliths occur at the tapered end of the coccospheres). Navilithus is very different being trimorphic, with the sail coccoliths having a very different form to the body coccoliths. Related to this, it is likely that Navilithus is motile since polymorphism is strongly related to motility (Young, 1994;Young et al., 2005) and the sail coccoliths are grouped in the characteristic manner of circum flagellar coccoliths. By contrast, the Calyptrosphaera sphaeroidea HET phase is known to be non-motile from culture observations (Noel et al., 2004) whilst Turrilithus and Tetralithoides form spherical monomorphic coccospheres with no obvious flagellar opening and so are probably non-motile. As noted by Young et al. (2003), one end of Placorhombus coccospheres is typically bluntly terminated, leaving a possible flagellar opening, so these are likely to be motile. In most coccolithophore families the heterococcolith phase is either consistently motile (e.g. Pleurochrysidaceae, Syracosphaeraceae, Papposphaeraceae) or consistently non-motile (e.g. Noelaerhabdaceae, Calcidiscaceae). However, there are also families with a mixture of motile and non-motile heterococcolithophores (e.g. Rhabdosphaeraceae, Pontosphaeraceae), so the variable motility in the narrow-rimmed placolith group does not prove it is an artificial grouping.
Ecology. All members of the narrow-rimmed placolith group are rare deep-dwelling species; they have been recorded only very sporadically and exclusively from deep photic zone samples. The only likely exception is the Calyptrosphaera sphaeroidea HET phase. This form has been observed only in culture so it is not know if it is a deep photic phase, or, as suggested by Noel et al., (2004), a coastal semi-benthic phase. It also is a less distinctive form than the other species so may conceivably have been overlooked. This ecological similarity provides weak support for the grouping.

Molecular genetic evidence.
Of the narrow-rimmed placolith group, only Calyptrosphaera sphaeroidea has been cultured, so this is the only species for which there is molecular genetic data. However, C. sphaeroidea fell in a somewhat surprising place in the molecular tree, occurring within the Coccolithaceae-Calcidiscaceae clade, in an intermediate position between those two families, together with a few other holococcolithophores known only from cultures (Fujiwara et al., 2001;Saez et al., 2004). An obvious problem of this result was that these holococcolithophores should have heterococcolith phases and it was difficult to predict which species these should be. That is, the molecular genetic data identified a group of holococcolithophores falling outside the main groups of heterococcolithophores and so predicted that there should be an associated group of heterococcolithophores. The Calyptrosphaera sphaeroidea HET phase described by Noel et al. (2004) is a narrowrimmed placolith, so an attractive possibility is that the other narrow-rimmed placoliths are heterococcolith phases of the additional holococcolithophores associated with C. sphaeroidea.
The morphological evidence in support of the narrow-rimmed placolith grouping is plainly suggestive but not overwhelming or conclusive. The group is diverse and the features which unite it are all somewhat disputable. The strongest features uniting the group are the form of the rim and the central plate. In this respect it is remarkable that the two most similar forms are C. sphaeroidea HET and Navilithus altivelum. The C. sphaeroidea heterococcoliths are almost identical to N. altivelum body coccoliths which have lost their spines. This similarity is so strong that it seems very good evidence of close taxonomic affinity, especially since basal features have consistently proven much more reliable for taxonomy than spines and other appendages. Given this, it can be rather strongly predicted that Navilithus will fall close to C. sphaeroidea in the molecular phylogeny and, from the tree topology (Saez et al., 2004), that this group diverged from the Coccolithaceae early in the evolution of the Coccolithaceae, i.e. probably in the early Palaeogene.
Of the narrow-rimmed placoliths, only Tetralithoides has a documented fossil record, extending back to the Early Miocene (Theodoridis, 1984;Young, 1998). There are, however, many poorly characterized fossil coccoliths in the Cenozoic and, as these are revised, possible affinity with the narrow-rimmed placolith group should be considered.

CONCLUSIONS
This paper has described a new, highly distinctive, deep photic coccolithophore species, Navilithus altivelum. The continued discovery of rare and patchily distributed deep photic species suggests that this environment is more heterogeneous than has been assumed and that systematic study of this assemblage may enable an understanding of the ecology of this key environment.
The unique feature of Navilithus coccoliths are the spines, which are supported by narrow basal stems and, in the case of the sail coccoliths, are formed of eight rows of interdigitating elements. No obviously similar spines occur in living or fossil coccoliths. However, it is well established that basal and especially rim features of coccoliths are the most useful phylogenetic characters of coccoliths. In this respect the Navilithus coccoliths show striking similarity to the coccoliths of Calyptrosphaera sphaeroidea HET and arguable similarities to the three other species included in the informal narrow-rimmed placolith grouping of Young et al. (2003). Since C. sphaeroidea is known from molecular genetic data to belong to a group of holococcolithophore species nesting within the Coccolithales clade, it can be predicted that Navilithus is a member of this group.