Sullivania a new genus of Palaeogene coccoliths

It is shown that the genus Chiasmolithus as currently understood comprises two discrete lineages. It is proposed that these should be formally recognised as separate genera. A new genus, Sullivania, is described and ten new combinations: S. californica, S. consueta, S. danica, S. edwardsii, S. gigas, S. inconspicua, S. minima, S. nitida, S. titus and Cruciplacolithus oulchyensis are introduced. The speculative evolution in the genera Cruciplacolithus, Sullivania and Chiasmolithus is discussed and illustrated.


I N T R O D U C T I O N
Untilnow thegenus Chiasmolithus has beenused for Palaeogene placoliths with a diagonal cross of X or H shape, regardless of the structure of the arm or the number of tube cycles. Detailed studies on the lineages in and structure of Chiasmolithus were presented by Gartner (1970), Romein (1979), Perch-Nielsen (1985) and van Heck & Prins (1987). Gartner (1970) and Romein (1979) distinguished two groups of Chiasmolithus based on the structure of the diagonal cross. Perch-Nielsen (1985) also distinguished two groups of Chiasrnolithus, but based on the extent of elements surrounding the central area distally; Group A (elements do not reach crest or just reach crest), which included the forms here distinguished as Sullivania and Group B (build the crest), here retained by Chiasmolithus.
In this study, only the species with two distally exposed tube cycles, a central mesh and a complex diagonal cross are retained in the genus Chiasmolithus ( Fig. 1; Gartner, 1970, fig.3). These criteria are consistent with the characteristics of the type species Chiasmolithus oamaruensis. The taxa with a single distally exposed tube cycle and a diagonal cross of which the arms are constructed by transverse elements have been assigned to the new genus Sullivaiiia ( Fig. 1; Gartner, 1970, fig. 5). There is no central mesh in the opening between the arms of the central cross. These criteria are consistent with the characteristics of the type species Sullivania consueta.
The various species of Chiasmolithus and Sullivanin are used as markers throughout the Palaeogene, but with the introduction of further species, the potential exists for a more detailed biostratigraphy than currently available. In the present study the zonation schemes of Martini (1971) and Varol(l989) are applied (Fig. 2).

LINEAGES
The genera Cruciplacolithus, Sullivania and Chiasmolithus constitute a distinctive sub-group within the Palaeogene Coccolithaceae. Lineages within this group have been discussed by Gartner (1970), Romein (1979), Perch-Nielsen (1985), and van Heck & Prins (1987). However, there has been no comprehensive overview for the entire Palaeogene, and new information and the description of additional species means that the previous schemes need revision. Figure 2 presents a revised analysis of the lineages. It is based on the published work as tested and supplemented by the present author's observations. The notes below concentrate on points of disagreement.

SPECULATIONS ON T H E LINEAGES IN SULLlVANlA
Sirllivania edzvardsii is the oldest known species of the genus and developed from Cruciplacolithus iriternzedius in Zone NP2 (NTp3A) by rotation of the central cross. Sullivarzin dnnicn developed from S. edzuardsii at the base of Zone NP3 (NTp3B) by rotation and narrowing of the angle of the central cross. The longer arm became curved and slightly dislocated at the centre whilst the short arm remained straight. Within Zone NP3 (NTp5), S. inconspictin developed from S. dnrzicn by further rotation of the diagonal cross until the gently curved arms became equal in length and by the diagonal cross becoming only weakly birefringent under cross-polarised light. This part of the evolutionary lineage was suggested by van Heck & Prins (1987) and the present author agrees with them.
In Zone NP5 (NTp9) S. consuetn developed from S . inconspicun by an increase in size and by the diagonal cross becoming more distinct. Sullivnnin cnliforrzicn developed from S. coiisiietn by becoming larger within Zone NP9 (NTp20). In ZoneNP12, S. minimn also developed from S. consuetn by slight rotation of its diagonal cross. The arms became straight and unequal in length and the short arm became slightly dislocated with respect to the long one. Sirllizurzin riririirrrn developed into Sullizwnin titirs by the curving of the long arm in Zone NP12, which in turn developed into S. riifida by the development of 'feet' at the end of the diagonal cross. Within Zone "5, S. gigns developed from S. rriiriirrin by enlargement of its shields and reduction of its central area.

SPECULATIONS ON T H E LINEAGES IN CHlASMOLllHUS
Chinsmolithus ederitulirs is the oldest known species of the genus. The view that Ch. edcntulirs (partly Chinsiirolitlius bideris of others) evolved from S. dnriico as suggested by Gartner (1970), Romein (1979) and Perch-Nielsen (1985) is not shared here; instead it is believed to be somehow related to Ericsonia sparsa and Cruciplacolithus? Sp.1 (sensu Perch-Nielsen, 1971; P1.1, Fig.9). Although the genera Sullivania and Chiasmolithus are superficially similar, there are major detailed structural differences, which are discussed here. There is considered to be insufficient time between the evolutionary appearances of S. danica and Ch. edentulus to accommodate the obvious major structural changes. The most important difference is the number of tube cycles. In Chiasmolithus edentulus the distally exposed tube cycle is double, whilst in Sullivania danicn a single distally exposed tube cycle is present (Fig. 1). Other differences are noted in the central area; Ch. edentulus has a diagonal cross with complex structure and a central mesh, whereas S. danica has a simply constructed central cross and lacks a central mesh. In particular, the development of the extra tube cycle in such a very short time span seems unlikely.
Moreover, the different opinions on the lineages of Chiasmolithus and Sullivania are also the result of the varying conceptsof S. danica. In the present study, the lectotype designated by van Heck & Perch-Nielsen (1987) is followed. Their description and figures clearly indicate that S. danica S.S. has a single, distally exposed, tube cycle and a simple diagonal cross. Ch. danicus of Gartner (1970), Romein (1979), Perch-Nielsen (1985) and van Heck & Prins (1987) includes S. danica S . S . and so-called 'intermediate' or 'late' forms which have two, distally exposed, tube cycles and a complex diagonal cross (divided into two halves under cross-polarised light). These latter forms are variations of Ch. edentulus and are not considered to be related to S. danica. The above authors, therefore, assigned these so-called 'intermediate' or 'late' forms to their Ch. danicus. The application of a wider species concept, subsequently led these authors to suggest different lineages in Chiasmolithus and Sullivania than those presented here.
Ch. edentulus gave rise to nine species in which constitute a coherent Palaeogene lineage of welldocumented species, characterised mainly by central area features.
In Ch. edentulus which has its first occurrence in Zone NP4 (NTp7B), one arm is straight whilst the other is slightly curved and is dislocated in the centre along the straight arm. Ch. bidens developed from Ch. edentulus in Zone NP6 (NTplOC) by the growth of two teeth protruding from the side of the inner tube cycle. Chiasmolithusgrandis developed from Ch. bidens in Zone NPll by the growth of two more teeth at the poles in addition to those on the sides and by an increase in the curvature of both arms. In Zone NPlO Chiasmolithus solitus developed from Ch. edentulus by increased curvature of the longer arm and its rotation towards the long axis, becoming almost parallel to it. Both arms of the cross also became more delicate.
Chiasmolithus eograndis also developed from Ch. edentulus in Zone NP10, by rotation of both arms and merging of their central part to form an oblique common arm giving an overall 'H' shape to the cross. One of the arms is usually more curved than the other. Chiasmolithus expunsus developed from Ch. eogrundis by increased curvature of the arms. The arms are equal in length and symmetrical. The angles between them are greater along the short axis. Chiasmolithus medius developed from Ch. eograndis in Zone NP15 by reduction of the curvature of one of the arms such that both arms form a short longitudinal common arm in the centre. Both arms are equal and symmetrical. Chiasmolithus modestus developed from Ch. medius in Zone NP16 by straightening and slight rotation of the arms until they made a right angle. Chiasmolithus oamaruensis developed from Ch. modestus at the base of Zone NP18 by further rotation of the arms until they made a small angle along the short axis of the ellipse.

Varol
Chiasrnolithus altus also developed from Ch. modestus, by reduction of its central opening and central cross. Perch-Nielsen (1981) suggested that Cruciplacolithus primus developed from Cruciplacolithus inseadus which, in turn, may have developed from the Cretaceous genus. In Zone NP3 (NTp5) Cruciplacolithus subrotundus developed from Cri primus by becoming circular. In Zone NP3 (NTp5) Cr. primus also developed into Cruciplacolithus latipons by an increase in size and an expansion of the axial cross to fill almost the whole centralarea. Inturn,assuggested byRomein(1979) Cr. latipons developed into Clausicoccus spp. by the almost complete disappearance of the axial cross into a perforated or non-perforated central plate, in Zone NP9 (NTpl8).

SPECULATION O N THE LINEAGES IN CRUClPLACOLlTHUS
Cruciplacolithus intermedius developed from Cr. primus in Zone NP2 (NTp2) by an increase in size and by the proximal shield and tube cycle becoming more strongly birefringent under cross-polarised light. Cruciplacolithus tenuis developed from Cr. intermedius by the growth of 'feet' at the end of the axial cross, in Zone NP3 (NTp3). Cruciplacolithus frequens developed from Cr. tenuis by anti-clockwise rotation of the central cross in Zone NP3 (NTp5). Campylosphaera eodela developed from Cr. intermedius by an increase in curvature of the shields about the short axis, in Zone NP9 (NTplS). Due to the curvature of the shields, Ca. eodela appears sub-rectangular. In Zone NPlO Campylosphaera dela developed from Ca. eodela by further curving of the shields along the short axis, giving it a rectangular appearance.
In Zone NP9 (NTpl8) Cruciplacolithus mutatus developed from Cr. intermedius by becoming sub-circular and by reduction of the axial cross to a more delicate form. The central area in the early form of Cr. mutatus is smaller than in later forms.
Cruciplacolithus oulchyensis, in Zone NP12, and Cruciplacolithus staurion, in Zone NP14, developed from Cr. mutatus. The differences are not distinct in these species. Perch-Nielsen (1971) placed Cr. staurion within Birkelundia because of the possession of a single proximal shield. There is, however, no convincing evidence that this species has a single proximal shield. Cr. staurion shows all the optical characteristics of Cruciplacolithus under cross-polarised light, having a birefringent proximal shield, tube cycle and axial cross, while Birkelundia has a non-birefringent proximal shield.
In Zone NP12 Cruciplacolithus flavius developed from Cr. mutatus by becoming smaller and more elliptical. In turn, Cruciplacolithus cruciatus developed from Cr. flavius by a further reduction in overall size, but with the central opening becoming larger in relation to the width of the shields. In Cr. cruciatus the extremely narrow tube cycle appears serrated under cross-polarised light. In Zone NP12 Cr. flavius also developed into Bramletteius serraculoides by the growth of a paddle-shaped distal process. Within Zone NP21 Cruciplacolithus tarquinius developed from Cr. flavius by the reduction of the central area, whilst in Zone NP22 Cruciplacolithus quader developed from Cr. flavius also by reduction of the central area and by growth of a quadrate distal structure in the centre of the axial cross. Diagnosis. An elliptical placolith with a distal shield, a double cycled proximal shield, one distally exposed tube cycle and a central area which is spanned by a diagonal cross. SEM description. The elliptical placolith consists of a distal shield made up of dextrally imbricated elements and a smaller proximal shield with two cycles. The radial elements of the tube, of which only one cycle is exposed distally, do not   overlap the inner margin of the distal shield. The central area is spanned by a diagonal cross composed of transverse elements (Fig. 1). So far, no central mesh has been observed in the opening between the arms of the central cross in any species assigned to this genus. L.M. description. The double cycled proximal shield, the tube (Pl. 1, fig. 5) and the diagonal cross are birefringent whilst the distal shield is non-birefringent under cross-polarised light. The arms of the diagonal cross appear as single units under cross-polarised light. Remarks. The species of this new genus are distinguished by the shape, angle of rotation and angle between the arms of the diagonal cross, as described by Perch-Nielsen (1985, Fig. 22) and also illustrated here (Fig. 3). The present interpretation is, however, slightly different from that of the former author.

SYSTEMATIC DESCRIPTIONS
Sullivaniu differs from Chiusmolithus under the scanning electron microscope by having a single distally exposed tube cycle, a diagonal central cross in which the arms are constructed of transverse elements, and absence of a central mesh (Fig.  1). Sullivunia differs from Chiusrnolithus under the light microscope by having a central cross in which the arms appear as a single unit and a strongly birefringent tube cycle without any overlap with the non-birefringent distal shield under cross-polarised light. By contrast in Chiasrnolithus the arms of the diagonal cross appear to be divided into two halves longitudinally, and the wide birefringent tube cycles (outer and inner tube cycles cannot be distinguished under the light microscope) overlap the inner margin of the nonbirefringent distal shield. Sullivunia differs from Cruciplacolithus by having a diagonal cross whereas the latter has an axial cross. Occurrence. The species of this genus occur throughout the world in low and high latitude areas. They become increasingly abundant, however, towards the low latitude areas whereas, conversely, the abundance of Chiasrnolithus increases towards the high latitude areas in many studied Eocene sections. The ratio of Sullivunia to Chiasmolithus may be a useful guide for the detection of palaeolatitudinal or climatic changes.
In the Eocene sediments of the North Sea area, Irish Sea, North Atlantic, West Africa and Falklands Plateau, species of Chiusmolithus are dominant over Sullivaniu whilst in the sediments of Turkey, Middle Eastern Countries, India and Pakistan, Sullivaniu is dominant over Chiasrnolithus. In Libya, however, an exceptionally high number of Chiasmolithus was observed in a certain horizon of Eocene age (within Zone NP13) where Sullivunia is extremely rare.
(Pl. 2, figs 8-10) Basionym: Cruciplacolithus edwardsii Romein, 1979 & Fert, 1954. SEM description. The elliptical placolith consists of a distal shield, a double cycled proximal shield, double tube cycles, both being exposed distally and a central area occupied by a complex diagonal cross and central mesh. The double cycled proximal shield is smaller than the distal shield which is composed of dextrally imbricated elements. The inner tube cycle is made up of strongly imbricated elements (Pl. 1, fig. 1) but due to diagenetic alteration its elements are usually fused together and appear as a smooth 'blanket' on the outer tube cycle. The overlap of the elements of the inner tube cycle is very strong, appearing as stacks of elements one on top of another in cross section (Pl. 1, fig. 2). The outer tube cycle is made up of radial elements which overlap the inner margin of the distal shield (Pl. 1, figs 1,3). The central cross is distally made up of longitudinal elements, and proximally of transverse elements divided into two halves longitudinally. There are net-like structures in the openings between the arms of the central cross in well preserved specimens (eg. P1. 1, fig. 1).
L.M. description. The arms of the diagonal cross are longitudinally divided into two halves under cross-polarised light.
The proximal shield and tube cycles are birefringent, while the distal shield and the diagonal cross are non or weakly birefringent, respectively, under cross-polarised light. The tube overlaps the inner margin of the distal shield.