Studies on the species level variation of selected coccolithophores

Markus Geisen

University College London, Geological Sciences Department, Gower Street, London WC1E6BT, UK The Natural History Museum, Palaeontology Department, Cromwell Road, London SW7 5BD, UK

Degree: Ph.D. Geology

Supervisors: Dr. Jeremy R. Young, The Natural History Museum Dr. Paul R. Bown, University College London ProQuest Number: U643698

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest.

ProQuest U643698

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 A bstract 2

Abstract

The oceans cover roughly seventy percent of the surface of the earth. With few exceptions minute photosynthetic primary producers like diatoms, dinoflagellates, silicoflagellates and coccolithophores inhabit the upper 200 m of this vast expanse. This phytoplankton forms the base of the marine food chain and plays an important role in geochemical cycles. Knowledge about species level biodiversity and spécia­ tion thus is important to understand marine ecology and biogeochemistry. Amongst the mentioned groups coccolithophores appear to be the ideal test group since their biomineralised periplasts - the coccoliths - provide a rich suite of qualitative and quantitative morphological characters and a uniquely extensive fossil record. In ad­ dition to this extant coccolithophore species can be grown in culture and hence are available for molecular genetic studies and cytological research. Combining data from both palaeontology and biology hence allows for a more detailed interpreta­ tion of coccolithophore biodiversity, evolution and spéciation. For the CODENET project (the EC funded Coccolithophorid Evolutionary Biodiversity and Ecology Network) extant species with seemingly global occur­ rences - spanning the biodiversity of coccolithophores - were selected and different case studies were carried out on them to elucidate their species level biodiversity. Here we present the results and discuss tentative models of spéciation for the coc­ colithophores in general. Following a brief introduction the second chapter (Methods) deals with sampling for both life and fossil coccolithophores and introduces culture techniques as well as biometrical methods used. Chapter 3 (Results) is divided in four parts - in part 1-3 detailed results from three CODENET species will be presented, mainly dealing with their morphological variability and physiology, but also with their evolution­ ary history. Part 4 is a published research paper in which the novel phenomenon of cryptic spéciation is demonstrated for the first time in the coccolithophores. This paper equally highlights the importance of studies on living species to achieve a better understanding and interpretation of coccolithophore evolution, diversity and spéciation. Chapter 4 (Discussion) comprises a comprehensive review about species level variation in all the CODENET key species. Here we tie together the various sources of data obtained during the project and introduce models of spécia­ tion for the coccolithophores. Further published manuscripts - both dealing with methods used in this study, but also supplementary data important for this thesis are given in an appendix. Aknowledgements

Acknowledgements - A narration

Work for this thesis started April 1998 with my move from Bochum University - where I graduated as a geologist - to the Natural History Museum (NHM) in London. This was not my first visit to the NHM - in fact I had been there just a year earlier using the biometrical routines that Jeremy Young had developed for work on my then beloved coccolithophore species Watznaueria barnesae from the Hauterivian. Being in need for a morphometrician for his Training and Mobility of Researchers (TMR) network Jeremy put his faith in me and subsequently offered me a position. This network - the aptly abbreviated CODENET - consisted out of eight European large scale facilities or universities with specialists in the fields of research that were targeted. As each of the institutes had the possibility to hire a post-doc eight so called young visiting researchers formed the core team. This team consisted of Ian Probert (Univ. Caen), Sabrina Renaud (ETHZ Zurich), Mario Cachao (Univ. Lisbon), Kees van Lenning (CSIC Barcelona), Alberto Garcia- Saez (AWI Bremerhaven), Hanno Kinkel (NIOZ), Patrizia Ziveri (Free Univ. Amsterdam) and myself. Some of us were lucky and so we got to know each other on the International Nannoplankton association meeting in Puerto Rico in early 1998. Tales of fabulous, albeit non-alcoholic drinks has it from there... Work really started in April and one of our first tasks was the identification and preparation of a communal sample set. This work was done in the course of several workshops in Zurich by a number of people namely Jorg Bollmann (ETH Zuric), Patrizia Ziveri, Jorijntje Henderiks (ETH Zurich), Sabrina Renaud and Mario Cachao. From there I still vividly recall the time Jorg spent on introducing me to his new SEM. What I learned from Jorg came in handy as the Electron Microscopy and Mineral Analysis (EMMA) unit at the NHM had just purchased a brand-new high specs SEM that soon became my very own toy. Chris Jones and Alex Ball from the EMMA unit did not only maintain this (and other) machines to superb standards, but they proved to be real experts in the field of microscopy whenever a tricky pro­ cedure was required to get the best out of the machine. I think the micrographs in the thesis speak for themselves and are a testimony to their skills. Later this year I went on my first oceanographic cruise around the Canary Islands on the German RV Meteor. In Claudia Sprengel (then Univ. Bremen, now AWI Bremerhaven) I found a specialist, happy to introduce me to the field of coccolitho­ phore sampling at the high seas. We turned out to become close friends over the years and have been a real “Dream Team” working together on cruises and are now sharing an office at the AWI. Aknowledgements

It was in the late summer of 1998 that I first had the chance to participate in one of the famous Blagnac meetings. Although I had a pretty good idea what to expect - from images that is - the beauty of the place, its suitability for scientific and private discussion until the early morning hours, and the hospitality of the owners - Dorine and Jan van Hinte - really blew me away. Blagnac has become an annual experience since then and 1 really miss the windowsill Ian and myself used to sit on in the morning to get over our hang-over. After a nice meeting, organised by Hanno Kinkel on Texel at the NIOZ in early 1999, a series of cruises followed. One was a short test bed trip for sampling techniques for which we hired the RV Andromeda from the Portuguese army and in which nearly all post-docs participated. Although Mario did an expert job on organising the trip he nevertheless refused to partake - allegedly due to his sea­ sickness. Having been on the ship, or boat one should really say, 1 guess Mario had the right approach. The methods and sampling protocol developed there were then used onboard the RV Hesperides in the Alboran Sea by a competent team consisting of Alexandra Broerse (Free Univ. Amsterdam), Kees van Lenning, Andy Howard (UCL) and the author. LIuisa Cros (CSIC Barcelona), whose taxonomic skill has been a great help throughout the project, provided bed and breakfast upon our return and an unforgettable tour around Barcelona late at night. This turned out to be our most successful trip and upon my return with the living samples Ian Probert shared his broad knowledge on culturing phytoplankton with me. Many trips from London to Caen should follow and Ian and myself probably deserve the “Seamiles Gold Card” for travelling with Brittany ferries. The head of lab in Caen, Chantai Billard and her colleague Jaqueline Fresnel have been fantastic hosts providing me with plenty opportunities for splendid working dinners. Later in the year Masanobu Kawachi and Marie-Helene Noel invited Patrizia, Jeremy, Ian and myself over to the National Institute for Environmental Sciences (NIES) in Tsukuba, Japan. He proved to be an excellent host and introduced us to the Japanese way of life. Under his expert guidance we had the chance to obtain samples from a small volcanic island off Tokyo Bay, which blew up later in the year. The year 2000 saw the Mid-term-Review of our project at the NHM - including a rather picturesque Them se boat trip involving loads of drinks... Later in the year 1 had the change to meet with Claudia and with Babette Bockel (Univ. Bremen) on a small meeting organised by Karl-Heinz Baumann (Univ. Bremen). The INA meeting in Bremen followed immediately afterwards and Claudia and Babette had the courage to allow a number a British colleagues to stay in their apartments. These were busy weeks - shortly after the INA meeting Claudia and myself left for Namibia for an oceanographic cruise in the South Atlantic. My culture experiments started late in the project and after my initial success with two of the selected species 1 managed to interest Blair Steel - a friendly, hard working and hard partying Scotsman - to perform further experiments for his mas­ ters thesis at UCL. Sharing my office at the NHM and the same taste as far as bever­ Aknowledgements

ages are concerned we quickly became friends and I am happy that Blair since then took up a PhD studentship with Michal Kucera at Royal Holloway. After the final Blagnac meeting in summer 2002 the project, sadly, came to an end. Luckily Hans Thierstein (ETH Zurich) offered to subsidise me to attend a meet­ ing on the Monte Verita near Ascona where I met Ulf Riebesell (AWI Bremerhaven) who invited me over to the AWI as a guest. To all of you mentioned in the narration I wish to express my thanks in support­ ing me and sharing your time, experience and enthusiasm with me. Further I have to express my thanks to my many colleagues and friends at the NHM for enduring my sometimes rather germanic attitudes and for introducing me to the scientific work done at one of the worlds finest natural history museums! Over the years a number of people have supported me - Herbert Geisen is thanked for specialist advice in English and moral support (you have been there, dad..), Jurgen Li pa and Gabi Lipa have provided additional funding, thankfully without having to apply for a grant. A big thank you goes to Karin Wiesenthal and Roland Schmitt whose flat in Cologne has given me the opportunity during the last year to escape from Bremerhaven ! Paul Bown, my supervisor at the UCL, and Jackie Lees (UCL) are thanked for their friendship and for adding the palaeontological perspective to my work. A very special person I have to thank is my friend and supervisor Jeremy Young. Thanks for placing your trust in me and employing me on this project. Your way of sharing your vast knowledge about coccolithophores has made working with you a pleasure! During the last five years Jeremy would let me stray off track considerably whilst pursuing various strands of research. His expert way of supervising my thesis would be to only gently nudge me back on track if I veered too far off course.

This thesis is dedicated to a person, whose wish to remain anonymous I shall respect. Table o f contents

Table of contents

Studies on the species level variation of selected coccolithophores 1

Abstract

Acknowledgements - A narration.

Introduction

1.1 Overview 9

1.2 Key achievements 10

1.3 References 12

2 Methods 14

2.1 Samples 14 2.1.1 Plankton samples 14 2.1.2 Sediment trap samples 15 2.1.3 Cultures 16 2.1.4 GDP sample selection 19 2.2 Sample preparation 19 2.2.1 Sediment samples 19 2.2.2 Culture samples 20 2.2.3 Sediment trap samples / filtered plankton samples 21 2.2.4 SEM preparations 21 2.3 Biometry 21 2.3.1 Scanning electron microscopy 21 2.3.2 Light microscopy 22 2.3.3 Measurement macros 22 2.3.4 Data processing / statistics 23 2.4 References 23 Table of contents 7

3 R esults 25

3.1 Syracosphaera pulchra - a model species to study fine scale variation and spéciation in coccolithophores? 25 3.1.1 Introduction 25 3.1.2 Taxonomy and description 25 3.1.3 Life cycle (Fig. 2) 31 3.1.4 Cultures and natural samples 33 3.1.5 Sediment trap samples 40 3.1.6 Holocene samples 42 3.1.7 Downcore samples (ODP site 664) 44 3.1.8 First occurrence 46 3.1.9 Conclusions 47 3.1.10 References 52 3.2 Spéciation in the Helicosphaera plexus 55 3.2.1 Introduction 55 3.2.2 Taxonomy and description of Helicosphaera carteri 56 3.2.3 Taxonomy and description of Helicosphaera wallichii 58 3.2.4 Taxonomy and description of Helicosphaera hyalina 59 3.2.5 Remarks on the genus Helicosphaera 62 3.2.6 Life cycle 64 3.2.7Cultures 64 3.2.8 Sediment trap samples 70 3.2.9 Holocene samples 72 3.2.10 Downcore samples (ODP site 664) 74 3.2.11 First occurrence 76 3.2.12 Conclusions 78 3.2.13 References 78 3.3 Umbilicosphaera spp. - A case of spéciation? 82 3.3.1 Introduction 82 3.3.2 Taxonomy and description of Umbilicosphaera sibogae 82 3.3.3 Taxonomy and description of Umbilicosphaera foliosa 83 3.3.4 Further methods 88 3.3.5 Cultures 89 3.3.6 Holocene samples 89 3.3.7 Downcore samples (ODP site 664) 93 3.3.8 First occurrence 99 3.3.9 Conclusions 101 3.3.10 References 101 3.4 Life-cycle observations Involving pairs of holococcolithophorid species : intraspecific variation or cryptic spéciation? ______103 Table of contents

4 Discussion; Species level variation in coccolithophores______[24

Species level variation in coccolithophores______125

Abstract _____ 125

Introduction 126 Morphology 126 Ecological / biogeographical separation 127 Culture studies 127 Genetic separation 127 Life-cycles and holococcolith phase differentiation 128 Summary - integrated methodology 129 Results 129 Umbilicosphaera sibogae and U. foliosa (Plate I) 129 Coccolithus spp. (Plate 2) 134 Helicosphaera spp. (Plate 3) 139 Calcidiscus spp. (Plate 4) 144 Gephyrocapsa oceanica and related species (Plate 2) 148 Syracosphaera pulchra (Plate 5) 152 Synthesis 157 Spéciation and divergences (Fig. 1) 157 Local adaptation - ecological spéciation? 157 Final remarks 159 Acknowledgements 160

References 160

5 A D oendix 166

5.1 Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton______166

5.2 Determination of absolute coccolith abundances in deep-sea sediments by spiking with microbeads and spraying (SMS-method) ______173

5.3 Xenospheres - Associations of coccoliths resembling coccospheres 184

5.4 Coccolithophores for exhibition : A note ______193

Bibliography Markus Geisen ______199 Introduction

1 Introduction

1.1 Overview

Phytoplankton is the base of the marine food web and an essential key to under­ standing both marine ecology and biogeochemistry. Critical to such understanding is knowledge of the extent of and controls on species-level biodiversity. The classi­ cal evolutionary model sees natural selection and most of all geographical isolation (allopatric spéciation) as the key factors driving spéciation. However, the concept of geographical isolation is difficult to apply to marine microplankton with seem­ ingly global occurrences and recent research is suggesting that spéciation is more ecologically driven. Coccolithophores are an ideal test group since their elaborate mineralised peri­ plasts, coccoliths, provide a rich suite of morphological characters and a uniquely ex­ tensive fossil record. The group has been subject to extensive research over the past decade especially in EU funded network project, “Coccolithophorid Evolutionary Biodiversity and Ecology Network”, CODENET (http://www.nhm.ac.uk/hosted_ sites/ina /CODENET/index.html) in which the author was working as a “Young visiting researcher”.

Species concepts This thesis integrates a number of different species concepts and hence a brief review of the methods used to define and recognise species is important. The operational species concepts in coccolithophores are traditionally based on using the morphological characters of the coccoliths covering the cell, and especially for heteroccolithophorids crystallographic orientation of the component crystal units (Young et al. 1992). This phenetic taxonomy has been successfully applied to the fossil record and compares well with findings from other characterisation methods such as cell ultrastructure (Inouye and Pienaar 1984) and more recently molecular genetics (Edvardsen et al. 2000; Fujiwara et al. 2001 ; Medlin et al. 1997; Sâez et al. 2003). However, the discovery of haplo-diplontic life cycles (Billard 1994; Fresnel 1994; Gayral and Fresnel 1983; Parke and Adams 1960; Rayns 1962) and therefore sexual reproduction in the coccolithophores makes it important to consider the bio­ logical species concept of Mayr (1963), defining a species as a reproductive entity, as well. Recently the discovery of cryptic species, in which spéciation is apparently uncoupled from morphological evolution, has provided evidence for another level of fine-scale genotypic variation (Geisen et al. 2002; Sâez et al. 2003).

Coccolithophore diversity The conventional interpretation of coccolithophore taxonomy is that there are about 150 well-described species, with, in almost all cases, inter-oceanic distributions 1 Introduction 10

within broad ecological boundaries (Jordan and Chamberlain 1997; Jordan and Kleijne 1994; Young and Bown 1994). A perception of very widely distributed rath­ er homogeneous species is well supported by geological evidence of synchronous apparently sympatric evolution across the world ocean and for near synchronous (on scales of less than a few thousand years) extinction events (Chepstow-Lusty et al. 1992; Wei 1993; Wei and Shilan 1996). Research over the past decade and especially from the CODENET project has greatly refined our knowledge of fine scale diversity in this group. New data has come from several sources. The primary source of data has been morphometric investigation of selected taxa using plankton, sediment trap, and surface sediment samples, and often time series studies of geo­ logical samples. This has lead to the identification of sub-morphotypes and morpho­ logical gradients and consequently to the development of hypotheses of causation in terms of ecotypic or genotypic variation.

1.2 Key achievements

The aim of this thesis was to examine the species level variation of selected coc­ colithophore species using different techniques, including biometry, physiology of monoclonal cultures, life cycle observations and analyses of samples from the geological record, sediment traps and from sea water collected on oceanographic cruises. Below we briefly summarise the results presented in this thesis.

Morphological studies Biometrical analyses of coccoliths were performed on cultures and natural samples (sediment trap and water samples) of three target taxa - Syracosphaera pulchra, Helicosphaera spp. and Umbilicosphaera spp. - and the results were compared with a detailed analysis of samples taken from the recent geological record. A result from the work on cultures is that the morphology of all tested species appears to be stable if tested under varying environmental conditions. Hence it is assumed to be under strong genomic control and gradual - or sudden - changes in morphology through time can be interpreted as evolutionary signals. In both Syracosphaera and Helicosphaera biometrical analysis has provided evi­ dence for both intra- and infraspecific variation in both cultured clones and natural samples (Chapters 3.1, 3.2) In the case of the Umbilicosphaera variants foliosa and sibogae biometrical anal­ ysis has provided sufficient evidence to raise the variants to species level (Chapter 3.3), which can be traced in the geological record. The true species character of the Umbilicosphaera variants was confirmed independently by Sâez et al. (2003) with molecular methods.

Coccolithophore life cycles Another source of data has come from the recognition of alternate life cycle phases (Fig. 1). It is now clear that the typical life cycle of coccolithophores consists of Introduction

independent haploid and diploid phases, both of which are capable of indefinite asexual reproduction (Billard 1994). Both phases usually produce coccoliths but \ ia distinctly different biomineralisation processes resulting in consistent structural differences (Young et al. 1999). The two phases thus potentially pro\ ide independ­ ent morphological ev idence of differentiation. This potential has been realised in several cases, particularly through recognition of rare combination coccospheres produced during phase transitions and bearing both coccolith types (Geisen et al. 2002). Here we demonstrate the phenomenon of cry ptic spéciation - spéciation taking place without any morphological change - for the first time in the coccolithophore Syracosphaera pulchra (C hapter 3.1). In Helicosphaera spp. (recent) spéciation events are reflected in different holococcolithophore life cycle stages (Chapter 3.2). A detailed discussion of the taxonomical and ev olutionary implications can be found in chapter 3.4 which is reproduced from Geisen et al (2002).

Fig. 1. False coloured award winning photograph of a life cycle change in Calcidiscus quad- riperforatus. Coccolithophores have a haplo-diplontic life cycle and produce coccoliths in both phases, but via different biomineralisation modes. During the haploid phase the green coloured holococcoliths are produced outside the cell membrane, whereas in the diploid phase the pink coloured heterococcoliths are produced in coccolith forming vesicles inside the cell and pushed outwards. This specimen is a rare example of a so called combination cell bearing coccoliths of both types. Combination cells occur when syngamy of two haploid holococcolith bearing cells takes place and the resulting diploid cell starts producing het­ erococcoliths again and sheds the outer cover of holococcoliths. This single specimen has changed the interpretation of diversity in the genus Calcidiscus considerably. For further dis­ cussion refer also to Geisen et al. (2002 - chapter 3.4) and Sâez et al. (2003) and chapter 4. 1 Introduction 12

These strands of evidence have produced significant evidence of different levels of genotypic variation within conventional species. Combining data from both palaeontology and biology has provided us with new insights into coccolithophore biodiversity, evolution and spéciation (Chapter 4).

1.3 References

Billard C (1994) Life cycles. In; Green J C, Leadbeater B S C (eds) The Haptophyte Algae. Clarendon Press, Oxford, pp 167-186 Chepstow-Lusty A, Shackleton N J, Backman J ( 1992) Upper Pliocene D iscoaster abundance variations from the Atlantic, Pacific and Indian Oceans: the significance of productivity pressure at low latitudes. Memorie Sci geol 43: 357-373. Edvardsen B, Eikrem W, Green J C, Andersen R A, Yeo Moon-van der Staay S, Medlin L K (2000) Phylogenetic reconstructions of the Haptophyta inferred from IBS ribosomal DNA sequences and available morphological data. Phycologia 39: 19-35 Fresnel J (1994) A heteromorphic life cycle in two coastal coccolithophorids, Hymenomonas lacuna and Hymenomonas coronata (Prymnesiophyceae). Can J Bot 72: 1455-1462 Fujiwara S, Tsuzuki M, Kawachi M, Minaka N, Inouye I (2001) Molecular phylogeny of the haptophyta based on the rhcL gene and sequence variation in the spacer region of the RÜBISCO operon. J Phycol 37: 121 -129 Gayral P. Fresnel J (1983) Description, sexualité et cycle développement d’une nouvelle coccolithophoracée (Prymnesiophiceae): Pleurochrysis pseudoroscojfensis sp. nov. Protistologica 19: 245-261 Geisen M, Billard C, Broerse A T C, Cros L, Probert I, Young J R (2002) Life-cycle associa­ tions involving pairs of holococcolithophorid species: intraspecific variation or cryptic spéciation? Eur J Phycol 37: 531-550 Inouye I, Pienaar R N (1984) New observations on the coccolithophorid Umhilicosphaera sihof>ae \ar. foliosa (Prymnesiophyceae) with reference to cell covering, cell structure and flagellar apparatus. Br phycol J 19: 357-369 Jordan R W, Chamberlain A H L ( 1997) Biodiversity among haptophyte algae. Biodiversity Conserv 6: 131-152 Jordan R W, Kleijne A (1994) A classification system for living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 83-105 Mayr E (1963) Animal species and evolution. The Belknap Press of Harvard University Press, Cambridge, Massachusetts Medlin L K, Kooistra W H C F, Potter D, Saunders J B, Andersen R A (1997) Phylogenetic relationships of the “golden algae” (haptophytes, heterokont chromophytes) and their plastids. PI Syst Evol Suppl 11: 187-219 Parke M, Adams I ( 1960) The motile {Crystallolithus hyalinus Gaarder & Markali) and non- motile phases in the life history of Coccolithus pelagicus (Wallich) Schiller. J mar biol Ass U K 39: 263-274 1 Introduction 13

Rayns D G (1962) Alternation of generations in a coccolithophorid, Cricosphaera carîerae (Braarud & Fragerl.) Braarud. J mar biol Ass U K 42: 481-484 Sâez A G, Probert I. Geisen M, Quinn F, Young J R, Medlin L K (2003) Pseudo-cryptic spé­ ciation in coccolithophores. Proc natn Acad Sci USA 100: 7163-7168 Wei W (1993) Calibration of Upper Pliocene - Lower Pleistocene nannofossil events with oxygen isotope stratigraphy. Paleoceanography 8: 85-99 Wei W, Shilan Z (1996) Taxonomy and magnetobiochronology of Trihrachiatus and Rhomhoaster, two genera of calcareous nannofossils. J Paleontol 70: 7-22 Young J R, Bown PR (1994) Palaeontological perspectives. In: Green J C, Leadbeater B S C (eds) The Haptophyte Algae. Clarendon Press, Oxford Young J R, Davis S A, Bown P R, Mann S ( 1999) Coccolith ultrastructure and biomineralisa­ tion. J struct Biol 126: 195-215 Young J R, Didymus J M, Bown P R, Prins B, Mann S (1992) Crystal assembly and phylo­ genetic evolution in heterococcoliths. Nature 356: 516-518 2 M ethods 14

2 Methods

In this chapter the different preparation and observation methods which were used to collect data will be explained in brief. This chapter will focus on the preparation of samples for scanning electron microscopy (SEM) and light microscopy (LM) as well as on culture work and morphometric analyses. A detailed account of the methods used can be found in the respective chapters.

2.1 Samples

For this study a number of different samples, including Holocene sediments, downcore sediments from various ODP cores, plankton samples, and sediment trap samples, were used. In this chapter the specific methodology for sampling and preparation of the samples will be given. Lists of all the samples used are given in the respective chapters.

2.1.1 Plankton samples

A large number of plankton samples were used for this study. These samples were collected mainly during oceanographic cruises. Water for filtration was normally taken from the onboard membrane pump at a depth of approximatel) 5 m to sample the surface water association and from a w inched rosette sampler with attached con­ ductivity, temperature, depth (CTD) probe (Fig. I) to sample the deep associations (down to 200 in). Up to ten depths per profile were sampled, the specific depths were selected using - where possible - the fluorometer log or alternativ ely the tem­ perature log of the CTD probe.

Fig. 1. Typical setup of a CTD depth probe. In the left image the trigger mechanism to close the bottle can be clearly seen. In the image on the right Claudia Sprengel (AWI Bremerhaven, Germany ) demonstrates a misfired and hence open bottle with no water recovery. 2 M ethods 15

Seawater was filtered with vacuum filtration ramps (Fig. 2) equipped with filter funnels capable of holding either a 25 mm or 47 mm diameter filter. Two types of filters have been used I) cellulose nitrate (CN) filters with a retention of 0.45 /vm and 0.8 /^m and 2) polycarbonate filters (PC) with a pore size of 0.4 pim. D ep en d in g on the abundance of coccolithophores and the diameter of the filter used up to 5 li­ tres of water were filtered. Salt w as remov ed by rinsing w ith buffered mineral water (0.5 mol NH^FICO,). After rinsing the samples were left to dry in an oven for 12 h at 50° C and stored in airtight petri dishes.

k,_ , 0

Fig. 2. Two different filtration ramps as used on various oceanographic cruises by the author. This image was taken abord R/V Meteor, off Namibia. The ramp on the right is a lightweight version, w hich proved very useful for travelling. 2.1.2 Sediment trap samples

Sediment - or better - particle traps are a rather new tool to quantify time-varying export fluxes of carbon and associated biogenic elements in modern oceans. A prin­ ciple goal is a mechanistic understanding of the processes controlling these biogeo­ chemical fluxes and to evaluate the related exchanges w ith the atmosphere and the sea floor, fhey usually consist of a moored array of collecting funnels, anchored at various depths in the water column. The funnels passively collect all particles set­ tling down through the water column. An array of stained cups can be set to collect particles over a giv en period of time. After the recovery of the trap array this mate­ rial is usually split (on a rotary splitter) and samples for either SEM or LM observa­ tion can be prepared in the usual way. For a detailed discussion of the methodology and its limitations refer to Sprengel et al. (2000) and Broerse (2000). 2 M ethods 16

2.1.3 Cultures

Monoclonal laboratory cultures of coccolithophores are essential to analyse the morphological variation of both cells and coccoliths as well as their physiological reaction to varying environmental conditions such as light, temperature and nutri­ ent levels. Equall} cultures are essential to investigate the molecular biology of coccolithophores, as single cell PCR - which is used on a regular basis in other marine protists like foraminifera - is still in the experimental stage of develop­ ment for coccolithophores. Also it does not allow for high-resolution microscopy for morphological identification prior to DNA extraction. The disadvantage of using monoclonal cultures for study of genetic variation is that it requires labour intensive culture isolation and growth w hich limits the number of analyses that can be carried out. I he advantage is that results from single cultures can be replicated and that molecular genetic analyses can be tightly integrated with quantitativ e mor­ phometric analyses, fhe cultures used in this thesis have not been grown as axenic - meaning virus and bacteria free - cultures, but ev er> care was taken to prevent contamination especially with bacteria. Great care was taken during isolation to ensure that all cultures were monoclonal. For a list of all culture available visit http: //vvvvw.nhm.ac.uk/hosted sites/ina/CODFNFT.

Seawater collection and culture isolation

To obtain concentrated seawater for isolation of coccolithophores small hand op­ erated plankton nets were used with a mesh size of 5 and 10 /vm. The nets were deployed from ships on station at depths between 5 to 15 m and were left in water for up to two hours (Fig. 3). During this time the dynamic positioning of the ship in the ambient current would ensure continuous pumping of water through the net. Sampling underway is not possible since the net would rapidly clog and tear. Additionally the nets were used to collect water from the CTD rosette sampler to

Fig. 3. The author and a colleague fishing for coccolithophores off Namibia, whilst on bord R/V M eteor. Research on coccolithophores is critically dependant on time consuming culture work. 2 M ethods 17 sample for deep dwelling species occurring in the deep photic zone. After collection the concentrated seawater was filtered through a 64 f4vn mesh sieve to remove large foraminifera and decanted into translucent storage containers. Usually two containers were used per sample and GeO, (to inhibit the reproduction of diatoms) and nutrients were added to one aliquot. The containers were stored as near as possible the ambient water temperature either in an incubator set to a 16 h light, 8 h dark (16L/8D) cycle or in a room with the light switched on. Every day the containers were opened to allow for air exchange. The samples were transported back to the laboratory in a cool box as soon as possible. Maintenance of water tem­ perature and rapid transport back to the laboratory were found to be vital in ensur­ ing obtaining healthy samples for culture isolation. However, with this care it was found possible to store samples for up to a week prior to isolations and to transport them thousands of miles back to the laboratory. This was a significant change from conventional wisdom, which suggested that isolations should be carried out within a few hours of water collection. Culture isolation was performed on an inverted microscope (Olympus BH 2) using 80x magnification and a glass micropipette. A single cell was captured, trans­ ferred in fresh medium, picked up again and finally transferred into sterile poly­ styrene tissue culture microplates with the wells filled with a media series ranging from K/2 to K/10 (Keller et al. 1987). Normally a microplate with only 24 wells was used to prevent the media from getting too warm whilst isolation. After completion the lid of the microplate was sealed with parafilm to prevent evaporation and the microplate was stored in an incubator. The microplates were checked on regularly, and growing cultures were than transferred into sterile 75 ml tissue culture flasks filled with ca. 40 ml of medium. Culture isolation was carried out in association with Ian Probert (U. Caen) as part of the CODENET project, approximately 40% of the cultures used in CODENET are from a number of oceanographic cruises where the author participated.

Culture maintenance

Cultures were maintained in exponential growth in an incubator (fig. 4) set to 17° C on a 16L/8D cycle. Typically every two weeks the cultures were checked with an inverted microscope and reinoculated in fresh medium using a laminar flow cabinet to prevent contamination. The main CODENET culture collection was maintained at U. Caen with a duplicate collection of key strains, including all those used in this study, maintained at the NHM. All the cultures used were clonal (i.e. produced from a single cell and so with only one genotype in a culture), but not axenic (i.e. viral and bacterial contamination was not prevented). However, whenever contamination by bacteria was evident under the light microscope the cultures were reisolated. All experiments were conducted with low bacteria cultures only. 2 M ethods 18

Fig. 4. Typical incubator as used for biological research. Physical parameters like light lev els and temperature can be adjusted. Cultures are either stored in glass flasks or sterile plastic tissue culture flasks

Medium was prepared from seawater collected from the French coast of the Hngiish Channel. The seawater was prefiltered with an ordinary paper filter circle and atitoclaved at 120° C for 15 min. After cooling nutrients - nitrate, phosphate, trace metals and vitamins - were added under a laminar flow cabinet. For the de­ tailed chemical composition of K medium, which is similar to f medium (Gtiillard 1975), refer to figure 5 and Keller et al. ( 1987).

Composition of K medium. Stock solutions are numbered 1-5. Each stock is made such that the addition of 1 ml/litre yields the final concentration in the medium. ______Additions Finai concentration in medium Comments ______OtM)______( 1 ) K N 0 j 8 8 4 sam e as f/2 (2 ) NH ,CI 10 addition to f/2 (3) Naj ortho-P 0 4 3 6 sam e as f/2 (K uses organic form) (4) Trace metals: FeEDTA * 11.7 f/2 uses FeCI] MnCl2.4H20 0.9 sam e as f/2 ZnSO^.TH^O 0.03 same as f/2 C 0 SO 4.7 H 2O 0.05 same as f/2 N a 2M o 0 4 .2 H2 0 0 .0 8 sam e as f/2 CUSO 4.5 H 2O 0 .0 1 o n e h a lf f/2 lev el Na2EDTA.2H;0 100 order of magnitude higher than f/2 N a jS e O j 0.01 addition to f/2 (5) Vitamins Thiamin-HCI 0.3 same as f/2 Biotin 0.0021 same as f/2 8 1 2 0 .0 0 0 3 7 sam e as f/2 S e a w a te r to 1 litre 1 -5 made with reagent grade chem icals and HPLC grade water. All solutions filter-sterilised through 0.2pm membrane filters. 1-4 stored at 4°C. 5 stored frozen at -20'’C. * ethylenediamine tetra-acetic acid. ______

Fig. 5. Composition of the medium used for culture maintenance and culture experiments. Nutrients were either filter sterilised or autoclaved with the seawater. Vitamins were filter sterilised and kept frozen until prior to usage. 2 Methods 19

Temperature / light gradient table

To test the physiology, morphology and growth rate of various strains of selected species a purpose build temperature / light gradient table (X T/L table) was used at the NHM. The table consists of a 15 mm thick aluminium plate with the temperature on both ends being controlled by circulating water from a temperature controlled bath through a boring in the plate. The temperature on the table ranged from 5° C on the cold side to 30° C on the warm side. To achieve a good temperature control the experiments on the table were conducted in sterile 75 ml borosilicate glass flasks.

2.1.4 ODP sample selection

To study the basic evolutionary history of the selected species samples were taken from seven DSDP / ODP cores covering the last 4 Ma at a 500 ka interval, with cold/warm climate stage sample pairs taken for each interval. For all cores but core 664 high resolution oxygen isotope stratigraphy is available and was used to identify the samples (e.g. Hodell and Venz 1992; Mix et al. 1995; Shackleton et al. 1990; Tiedemann et al. 1994), the age model for core 664 is based on few magneto- chronological datums (Brunhes/Matuyama and Jamarillo-top and bottom) plus nan- nobiochronological data mostly from LAD of Discoaster forms (Su 1996). Tables with sample information are given in the respective chapters.

2.2 Sample preparation

2.2.1 Sediment samples

A range of techniques was used to prepare the Holocene and ODP samples. Initially it was tried to eliminate a possible bias in the homogeneity of the distribution of particles on standard smear slides. It was decided that a range of quantitative meth­ ods should be investigated and compared. This was conducted as collaborative ex­ ercise with Jorijntje Henderiks, Jorg Bollmann and Sabrina Renaud of ETH-Zurich. Methods tested on a standard sample included the random settling (Fig. 6) method first described by Williams and Bralower (1995) and Beaufort (1991), filtration (Andruleit 1996; Backman and Shackleton 1983) and spiking with microbeads and spraying - SMS method (Bollmann et al. 1999). Subsequent comparison of the count data from the same standard sample however revealed a methodological error in the settling method (Geisen et al. 1999) caused by elevating the slide over the bottom of the settling tank as described in Beaufort (1991). Geisen et al. (1999) con­ clude that the sinking of particles in a settling tank is not only controlled by Stokes Law but is also influenced by random currents caused by thermodynamic effects during the prolonged settling time of up to 24 hours. These flows should lead to a 2 Methods 20

Fig. 6. Laboratory setup at the ETH in Zurich whilst the author performed the settling experi­ ments. To the left the infamous airguns used for the SMS experiment can be seen. random dispersion of the particles left in the water causing similar settling rates on the elevated slide and on the bottom of the settling tank. For the communal sample set it was decided to use the novel SMS-method. Unfortunately the SMS-method has a number of disadvantages:

1. fhe density of particles on the slide is usually very low, which proves problem­ atic for rare species. 2. fhe preparation is \ery time consuming and involves precision balances and expensive microbeads. 3. The method will only work with tine fraction material, which may represent an altered flora. 4. fhere is no control on the dynamics of the spraying process. 5. To gain quantitative data both coccoliths and beads have to be counted which reduces the total number of samples that can be analysed in a given time.

For a review of quantitative sample preparation techniques and for a detailed de­ scription of the settling and SMS-method refer to the reprints of Bollmann et al. ( 1999) and Geisen et al. {1999) in the appendix. Mainly due to problems with the abundance of the selected species in the slides prepared with the SMS method it was decided to prepare simple smear slides from the flne fraction (<38 /

2.2.2 Culture samples

Culture material for light- or scanning microscope observation was usually harv est- ed in the late exponential growth phase. To remove salt from the medium - which would make light microscopic observ ation impossible -a wash and centrifuge series as follow s was applied: In a first step 1.5 ml Eppendorf centrifuge tubes were used to collect and centrifuge the samples for 8 min. The supernatant was removed, alcohol was added, the samples were ultrasonicated for 10 seconds and centrifuged again for 2 minutes. Afterwards the supernatant was removed again and the samples were 2 Methods 21 left to dry in an oven at 40° C overnight. To produce simple smear slides the pellet was suspended in alcohol again pipetted on a cover slip and left to dry. Alcohol was used rather than buffered distilled water to ensure dissolution did not occur.

2.2.3 Sediment trap samples / filtered plankton samples

Like the wild plankton samples the samples from sediment traps used in this study were splits from the original sample filtered on CN filters. For light microscopic ob­ servation a portion of the filter was mounted in microscope immersion oil on a glass slide with a cover slip. Although this preparation method would generally be seen as a non-permanent mounting method, our own experience has shown that, given correct horizontal storage in a slide cabinet and use of good quality immersion oil, the slides can be used for years without obvious deterioration.

2.2.4 SEM preparations

To avoid contamination of the high vacuum SEM column conventional mounting media such as double side tape or silver dag cannot be used in the field emission SEM. Instead a piece of photographic film was glued emulsion side upwards on a standard stub with Araldite and a replicate of the filter was then mounted on this with water. Afterwards the stub was coated with 20 nm of 95.5 % gold palladium using a Cressington 208HR sputter coater. A quartz crystal thickness monitor con­ trolled the thickness.

2.3 Biometry

2.3.1 Scanning electron microscopy

Scanning electron microscopy was performed on a digital field emission scanning electron microscope (Philips XL-30 PEG) at the NHM. As a standard an accelera­ tion voltage of 5 Kv was used and working distances ranged between 5 mm and the eucentric stage position at 10 mm. All imaging was performed using a built-in digital framegrabber system producing 1424 x 968 pixel images. Calibration on standards, however, revealed a spatial distortion of the digital images in the process of capturing and the short axis of the pictures had to be stretched in FhotoShop by 9.5%, yielding images of 1424 x 1064 pixels. Subsequent measurements of coc- cosphere and coccolith size were performed manually on-screen using the image analysis package Scion Image (Rasband 1998). 2 Methods 22

2.3.2 Light microscopy

The system used for biometry consists of four parts:

1. Microscope. For coccolith biometry a Zeiss Axioplan photomicroscope with a 1.4 NA oil immersion condenser lens and a 1.3 Na x 100 Neofluar oil immersion objective lens and a x 1.6 Optovar intermediate magnification was used. Young et al. (1996) have noted the advantages of using an oil immersion condenser for biometrical work on low biréfringent coccoliths of Emiliania Imxleyi, however due to the optical characteristics of the species selected in this study this proved unnecessary. 2. Camera. A Hamamatsu charged coupled device (CCD) camera with an analogue output, which is mounted to a C-mount on the microscope. 3. Framegrabber. The framegrabber converts the analogue signal from the camera into a digital image. With a grey level camera this typically results in an 8 bit image with 256 grey levels. Two different framegrabbers have been used, firstly a Perceptics Pixel Buffer board as described in Young et al. ( 1996) which pro­ vides a fast video refresh rate but is rather unstable and secondly a Scion Image LG-3 PCI based framegrabber with a slower video refresh rate, but drastically improved stability. Together with the specific camera both systems deliver an image of 768x521 pixels and at the magnification typically used this represents a pixel scale of 15.4 pixels//im (I pixel = 0.065 ^m). Both framegrabbers pro­ duced images with 5% spatial distortion. As with the SEM images the images were stretched digitally, immediately after capture to remove this distortion. 4. Computer & software. All the image capturing, storing and analysis have been performed on Apple Macintosh computers running the image analysis package NIH-image (Rasband 1998). Unlike many commercial image analysis packages NIH-lmage is public domain software and features a macro language, similar to Pascal, which can be programmed to perform user-defined tasks.

2.3.3 Measurement macros

A set of macro measurement routines had previously been developed for work on Emiliania huxleyi (Young et al. 1996). These macros treat image acquisition and image analysis as two separate processes. In the first step composite mosaics of separate coccolith images are assembled and stored. In this work phase contrast and crossed polarised light (CPU, #-polars) images have been captured, depending on the species. The capturing macros have been modified slightly to allow for captur­ ing of both CPU and phase images of a single specimen. The measurement macros written by Young et al. (1996) focused on automated measurements on clean culture material with veiy good preservation. For most of the geological material used in this study the automated measurement macros however proved to be too unreliable for routine use. Instead the author of this study devel­ oped a set of macros for fast manual measurements. This macro uses horizontal and 2 Methods 23 vertical lines, which can be positioned with the mouse at the respective characters to be measured. The macro then automatically calculates species specific features like coccolith length, width, central opening diameter (abbreviated with co in the figures) and stores the data in a file. These macros are very robust in that even dam­ aged or partially broken specimens can be measured and they are easy to change to allow measurements on images of coccoliths of other coccolithophore species and other microplankton. However due to the nature of some graphic routines used they will only work on Apple Macintosh computers. These macros are supplementary to the set of macros developed by Young et al. (1996) and have been included together with a working copy of NIH-lmage on a compact disc. The Coccobiom macros are further documented on the CODENET web site: www.nhm.ac.uk/hosted sites/ina/ CODENET/.

2.3.4 Data processing / statistics

Biometrical data from coccoliths obtained by light or scanning electron microscopy normally consists of a relatively low number of variables. Typically these variables include length, width, central area diameter and proximal or distal shield diameter for a given number of specimens per sample. Statistical parameters that describe a single sample include the minimum, maximum, statistical mean and the standard deviation (abbreviated with st. dev. in the figures). For a more detailed analysis of the morphospace described by the aforementioned parameters it was decided to grid bivariate datasets to analyse the frequency distribution of measurements in a sample. The size of grid cells for Syracosphaera spp. and Helicosphaera spp. was set to 0.5 pm for coccolith length and width and for Umbilicosphaera spp. 0.5 pm (proximal shield diameter) and 0.25 pm (distal shield rim width) was used. For the gridding process a public domain program (Knappertsbusch 1999) was used, the output matrix was then imported into Transform (Spyglass 1995), a data visualisa­ tion program.

2.4 References

Andruleit H (1996) A filtration technique for quantitative studies of coccoliths. Micropaleontology 42: 403-406 Backman J, Shackleton N J (1983) Quantitative biochronology of Pliocene and Early Pleistocene calcareous nannofossil from the Atlantic, Indian and Pacific Oceans. Mar Micropaleontol 8: 141-170 Beaufort L (1991) Adaptation of the random settling method for quantitative studies of cal­ careous nannofossils. Micropaleontology 37: 415-418 Bollmann J, Brabec B, Cortés M Y, Geisen M (1999) Determination of absolute coccolith abundances in deep-sea sediments by spiking with microbeads and spraying (SMS- method). Mar Micropaleontol 38: 29-38 2 Methods 24

Bown P R, Young J R (1998) Techniques. In: Bown P R (ed) Calcareous nannofossil bios­ tratigraphy. Chapman & Hall, pp 16-28 Broerse A T C (2000) Coccolithophore export production in selected ocean enviroments. PhD thesis. Free University Amsterdam (ISBN 90-9013381-X), p. 185 Geisen M, Bollmann J, Herrle J O, Mutterlose J, Young J R (1999) Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton. Micropaleontology 45: 437-442 Guillard R R L (1975) Culture of phytoplankton for feeding marine invertebrates. In: Smith W L, Chanley M H (eds) Culture of marine invertebrate animals. Plenum, New York, pp 26-60 Hodell D A, Venz K (1992) Toward a high-resolution stable isotope record of the Southern Ocean during the Pliocene-Pleistocene (4.8 to 0.8 Ma). Antarctic Res Ser Washington 5 6 :2 6 5 -3 1 0 Keller M D, Selvin R C, Claus W, Guillard R R L (1987) Media for the culture of oceanic ultraphytoplankton. J Phycol 23: 633-637 Knappertsbusch M (1999) Grid 2.1, public domain data analysis program. Mix A C, Pisias N G, Rugh W, Wilson J, Morey A, Hagelberg T K (1995) Benthic foramini- fer stable isotope record from site 849 (0-5 MA): Local and global climate cycles. In: Pisias N G, Mayer L A, Janecek T R, Palmer-Julson A, Andel T H v (eds) Proc ODP, Sci Results, College Station, pp 371-412 Rasband W (1998) Scion Image, US National Institute of Health / Scion Corporation. Shackleton N J, Berger A, Peltier W R (1990) An alternative astronomical calibration of the lower Pleistocene timescale based on ODP Site 677. Trans RSoc Edinb Earth Sci 81: 251-261 Sprengel C, Baumann K-H, Neuer S (2000) Seasonal and interannual variation of coccol­ ithophore fluxes and species composition in sediment traps north of Gran Canaria (29°N 15°W). Mar Micropaleontol 39: 157-178 Spyglass (1995) Transform 3.02: Savoy, Illinois, Spyglass Inc. Su X (1996) Development of late Tertiary and Quaternary coccolith assemblages in the Northeast Atlantic. Geomar Report 48: 1-120 Tiedemann R, Samtheim M, Shackleton N J (1994) Astronomical timescale for the Pliocene d 180 and dust flux records of the Ocean Drilling Program site 659. Paleoceanography 9: 619-638 Williams J R, Bralower T J (1995) Nannofossil assemblages, fine fraction stable isotopes, and the paleoceanography of the Valanginian-Barremian (Early Cretaceous) North Sea Basin. Paleoceanography 10: 815-839 Young J R, Kucera M, Chung H-W (1996) Automated biometrics on captured light micro­ scope images of coccoliths of Emiliania huxleyi. In: M oguilevsky A , W hatley R (eds) Microfossils and oceanic environments. Aberystwyth Press, pp 261-277 3 R esults 25

3 Results

3.1 Syracosphaera pulchra - a model species to study fine scale variation and spéciation in coccolithophores?

3.1.1 Introduction

The extant species Syracosphaera pulchra is the best known, largest and most abundant of all Syracosphaera species, a genus of about 40 extant species. It occurs globally in all Oceans (Okada and Honjo 1973; Okada and McIntyre 1977) and is common in sediments. As S. pulchra coccospheres and coccoliths are very rich in morphological characters the species was regarded to be very well defined and consequently was chosen in the CODENET project to act as a control species to quantify the degree of intraspecific variation. Despite the character richness of S. pulchra so far only a few morphometric stud­ ies were conducted (Baumann and Sprengel 2000). Here we present new findings on the morphology and physiology of 5. pulchra, from culture samples, maintained at different temperatures, from a sediment trap time series off Somalia, from the Holocene of selected ODP cores and downcore samples of ODP core 664 in the central equatorial Atlantic. Equally the life cycle of S. pulchra is reviewed in the light of new observations.

3.1.2 Taxonomy and description

Syracosphaera pulchra Lohmann (Plate 1, figs 1-5) Lohmann (1902), pp. 133-134, pi. 4, figs 33, 36, 36a, 36b, 37; Kamptner (1937), p. 301, pi. 15, figs 12-14; Kamptner (1941), pp. 85-86,105-106, pi. 7, figs 77-78, pi. 8, figs 79-84; Halldal et Markali (1955), p. 12, pi. 10, figs 1-4, p l.ll, figs 1-4; Loeblich et Tappan (1963), p. 193; Borsetti and Cati (1972), p. 402, pi. 46, figs 2a, 2b; Okada et McIntyre (1977), p. 27, pi. 10, figs 11-12; Gaarder et Heimdal (1977) pp. 54-55, pi. 1, figs 1-8; Nishida (1979), pi. 6, fig. 3; Inouye et Pienaar (1988), pp. 206-213, figs 1-14; Heimdal (1993), pp. 227-228, pi. 7; Kleijne (1993), p. 241, pi. 5, fig. 10; Cros et al. (2000), pl. 1, fig. 1; Gros (2002), pp. 52-53, pl. 26, figs 1-5. 3 Results 26

Syracosphaera pulchra - morphology

%

C) schematic plan view and cross section through an endothecal caneolith A) scanning electron micrograph of a complete displaying the corrugated wall with three flanges. coccosphere with exothecai and endothecal coccoliths.

B) scanning electron micrograph of a circumflagella endotfiecal caneolith with bifurcate spine and radial central area laths.

D) light micrograph of the endothecal caneoliths of a coccosphere. Cross polarised light was used to identify the crystallographlc orientation of the Calcite C-axis. The lower part of the rim and the E) scanning electron micrograph ol a basin flanges display a vertical C-axis, the upper part of shaped exothecai coccolith in sideview, displaying the rim shows a radial C-axis. The central area laths the slitted walls and the central depression. have a tangential C-axis orientation.

Fig. 1. Morphology of S. pulchra in the SEM and LM. Heterotypic synonyms

Daktylethra pirus (Kamptner) Norris (Plate I, figs 6-8) Norris (1985), p. 631, figs 38-39 Calyptrosphaera pirus Kamptner (1937), pp. 304-305, pi. 16, figs 21-23; Borsetti et Cati (1976), p. 211, pi. 13, figs 1-3; Kleijne (1991), pp. 28-29, pi. 3, figs 5-6; Heimdal (1993), p. 176, pi. 1; Cros (2002), p. 88, pi. 64, figs 1-3.

Calyptrosphaera oblonga Lohmann (Plate 1, figs 9-11) Lohmann (1902), p. 135, pi. 5, figs 43-46; Halldal et Markali (1954), p. 332, figs 1-4; Halldal et Markali (1955), p. 8, pi. 1, figs 1-3; Norris (1985), pp. 626-628, fig. 36; Kleijne (1991), p. 28, pi. 3, figs 3-4; Heimdal (1993), pp. 172-174, pi. 1; Cros (2002), p. 85, pi. 61, figs 3-4. 3 R esults 27

Description of the heterococcolithophore

Coccosphere ellipsoidal and dithecate (Fig. 1 - A; Plate 1, figs 1,2) with (031) dimorphic endothecal caneoliths (muroliths) (Plate 1; figs 3,4). The cells bear two flagellae and a haptonema and the flagellar opening is surrounded by (usually 4) modified coccoliths possessing a spine, which is forked at the end (Fig. 2 - B, Plate 1, fig. 4). The elliptical body caneoliths have a corrugated wall with three external flanges (Fig. 1 - B, C; Plate 1, fig. 3). The lower part of the rim, including the flang­ es, is composed out of elements with a vertical crystallographlc c-axis, whereas the elements in the upper part of the rim display a radial C-axis (Fig. 1 - D). The central area is formed out of numerous radial laths - with tangential C-axis orientation - which extend towards the centre of the coccolith and are partly joining (Fig. 1 - B, C). The monomorphic basin-shaped exothecai coccoliths (0l5) are elliptical with a narrow central depression and slitted walls (Fig. 1 - E; Plate 1, fig. 5).

Remarks

The extant heterococcolithophore species Syracosphaera pulchra is motile and the drawing of the type specimen taken from a water sample off Syracuse (hence the genus name) by Lohmann (1902) features a cell with two chloroplasts and a single flagella. Schiller (1925) however records the typical two flagellae. Lohmann (1902) in the type description fails to realise that the coccosphere is dithecate which is later recorded by Kamptner (1941). Kamptner (1941) calls the exothecai coccoliths Calyptroliths (due to their similarity with holococcoliths of the genus Calyptrosphaera) which has caused a considerable amount of confusion as com­ bination cells bearing both holococcoliths of Calyptrosphaera oblonga and hetero- coccoliths of S. pulchra have been reported (Cros et al. 2000; Geisen et al. 2002; Saugestad and Heimdal 2002). All the early observations, including the type species, were conducted using light microscopy and documented with drawings or, later in the century, with photo documentation. Lohmann (1902) used 1-2% formalin and later alcohol to conserve his water samples. These methods will dissolve coccolithophores in the long run and so there is no chance to re-examine the type material. In accordance with the ICBN article 8.3 drawings or images are considered to be the type if a specimen cannot be preserved and hence the figures drawn by Lohmann (1902) are valid type illustrations. Later taxonomic work on S. pulchra uses TEM and SEM observations (Borsetti and Cati 1972; Halldal and Markali 1954) and Inouye and Pienaar (1988) provide a cytological description including the flagellar apparatus. Recently molecular data (A. Garcia-Saez, unpubl. data) and pigment data (K. van Lenning, unpubl. data) has become available for a number ofS. pulchra cultures. S. pulchra was designated as the type species of the genus Syracosphaera by Loeblich and Tappan (1963). S. pulchra has a life cycle independently producing the characteristic heterococ- coliths in the diploid phase and either holococcoliths of C. oblonga (Fig. 3 A) or D. 3 R esults 28 pirus (Fig. 3 B) in the haploid stage (Cros et al. 2000; Geisen et al. 2002). Lohmann (1902) (pl. 6, fig. 67) records a combination cell of S. pulchra with C. oblonga, which he decided not to include in the figured type material. Although both C. ob­ longa and D. pirus become heterotypic junior synonyms for a life cycle stage of S. pulchra it appears that the morphology of the heterococcolith S. pulchra has not undergone any morphologically detectable change to reflect this (cryptic-) spécia­ tion event. Subdividing 5. pulchra on the species level seems to be indicated, but is pending further molecular research. Geisen et al. (2002; submitted, refer to chapter 3.4 and 4) discuss the problem in detail and propose name changes. 3 R esults 29

Plate 1 (next page) Syracosphaera spp. Fig. 1: Scanning electron micrograph of a S. pulchra coccosphere. This typical specimen displays endothecal and exothecai coccoliths. Coccoliths surrounding the flagellar pole are spine bearing. Water sample, N. Atlantic, off the Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 2: Scanning electron micrograph of a S. pulchra coccosphere. Note the lack of exothecai coccoliths. Water sample, N. Atlantic, R/V Meteor cruise 42-4B, station US IB. Fig. 3: Scanning electron micrograph of S. pulchra endothecal coccoliths. The specimen in lateral view shows the typical wall structure with three flanges. Note the lack of exothecai coccoliths. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 4: Scanning electron micrograph of a S. pulchra circumflagellar endothecal coccolith. Note the typical central spine with forked end. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 5: Scanning electron micrograph of a S. pulchra exothecai coccolith. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 6: Scanning electron micrograph of the holococcolithophore stage of S. pulchra. This stage was previously described as “Daktylethra pirus” and is referred to as S. pulchra HO pirus-type. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 7: Scanning electron micrograph of a S. pulchra HO pirus-type circumflagellar holococ- colith in lateral view. The circumflagellar coccoliths typically have a pointed hood. Note the clear offset between the hood and the base and the presence of perforations in the hood. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 8: Scanning electron micrograph of a S. pulchra HO pirus-type holococcoliths. The cir­ cumflagellar coccoliths typically have a pointed hood. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 9: Scanning electron micrograph of the holococcolithophore stage of S. pulchra. This stage was previously described as “Calyptrosphaera oblonga” and is referred to as S. pulchra HO oblonga-type. Water sample, N. Atlantic, off the Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 10: Scanning electron micrograph oi a S. pulchra HO oblonga-type holococcolith. Note the hexagonal arrangement of the calcite rhombohedra and the absence of an offset between the hood and the base. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 11: Scanning electron micrograph of a S. pulchra HO oblonga-type circumflagellar holococcoliths. The circumflagellar coccoliths typically have a pointed hood Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69.

Scale bars represent: Figs 1, 2, 9: 5pm, fig. 6: 10/

m

g # . '

M y #

3. S, pulchra endothecal coccoliths 4, S. pulchra circumflagellar 5. S. pulchra circumflagellar exothecai in distal and sideview endothecal coccolith coccolith

6 . S. pulchra holococcolithophorid 7. S. pulchra circumflagellar 8. S. pulchra pirus-type holococcolith pirus-type holococcoliths pirus-type

10. s. pulchra holococcolith oblonga-type

11. S. pulchra circumflagellar holococcoliths oblonga-type 9. S. pulchra holococcolithophorid oblonga-type 3 R esults 3 1

3.1.3 Life cycle (Fig. 2)

First observations that the life cycle of S. pulchra involves a holococcolitho­ phore and a heterococcolithophore species come from combination cells bearing both holo- and heterococcoliths. Cros et al. (2000) pictures one clear example and one tentative example of S. pulchra in combination with the holococcolith Calyptrosphaera oblonga. A re-examination of older literature revealed another four examples of this type of combination (Kamptner 1941; Lecal-Schlauder 1961; Lohmann 1902). During routine examination of water samples from the North Atlantic we have found two more clear examples of this combination, taking the total of observations to eight. In addition to that we have found five coccospheres of S. pulchra in combination with the holococcolith Daktylethra pirus in water sam­ ples from the Alboran Sea (western Mediterranean) (Geisen et al. 2002). Literature research provided tentative evidence for another example of this type of combina­ tion (Lecal-Schlauder 1961) and we were informed of a further two observations from the western Mediterranean (Saugestad and Heimdal 2002) and an additional observation from the Adriatic Sea (B. Balestra, unpubl. data). In addition to this one of the cultures used for the morphometric study of S. pulchra (NAP 10, off Naples, Italy) has undergone a transition to the holococcolithophorid phase. The resulting motile phase bears holococcoccoliths but they are - as it is often the case in cultures - only poorly formed and usually disintegrate during preparation for scanning elec­ tron microscopy. Recently these coccoliths have been confirmed with light micros­ copy and scanning electron microscopy as D. pirus. Although D. pirus and C. oblonga are conventionally placed in different holoco­ ccolithophorid genera and do show significant differences. Nonetheless, they show enough morphological similarities, for instance in the unique form of their circum flagellar coccoliths with pyramidal bosses, to allow a very close affinity to be hy­ pothesised (see discussion in 2002 and chapter 3.4 of this thesis). Intergradational morphotypes between the two holococcoliths have not been observed and the com­ bination cells are either with C. oblonga or D. pirus. The S. pulchra coccoliths on both types of the combination cells show normal morphology including both endo- and exothecai coccoliths, indicating that their formation takes place at the same time. Despite the high number of morphological characters of S. pulchra coccoliths it has not been possible to detect any consistent differences between the heterococ­ coliths on the two types of combination cells. As heterococcoliths are formed inside the cell and holococcoliths calcify outside the cell (Young et al. 1999) the disposi­ tion of coccoliths on combination cells can reveal the mode of transition between the two phases and there is evidence for both transitions from heterococcolitho­ phore to holococcolithophore (including culture observations on culture NAP 10) and vice versa for both types of combinations.

Fig. 2 (next page). Life cycles of S. pulchra. The predominant mode of reproduction for coc- colithophorids is simple mitotic division in the heterococcolith bearing diploid stage (left part of both A and B). If the cells undergo Meiosis they can change to a haploid, holococcolith bearing stage, again capable of mitotic division. The conditions causing a life cycle change 3 Results 32

Syracosphaera life cycles

Meiosis Mitosis Mitosis r

evidence from wild combination cells

heterococcolith Syngamy holococcolith stage stage

diploid phase morphology haploid phase morphology^- "C oblonga" r •

M eiosis Mitosis Mitosis

evidence from wild combination cells and cultures

heterococcolith Syngamy holococcolith stage stage

haploid phase morphology - "D. pirus"

remain unknown. Syngamy of two daughter cells may be seen either in culture or in rare combination cells displaying both holo- and heterococcolith morphologies. This life cycle is well demonstrated in other coccolithophorid species - as a novelty S. pulchra seem s to consist out of two biological species with a cryptic heterococcolithophore. The morphologies of the holococcoliths involved (A. C. oblonga, B. D. pirus) are displayed here. 3 Results 33

3.1.4 Cultures and natural samples

T he tw o S. p u lc h ra strains (GK7, G K 17) used for the temperature experiment were isolated from the Delta de 1 Ebro (Spain, western Mediten anean). To compare mor­ phologies 12 further strains maintained at 17° C only were analysed. Data for all the cultures used is summarised in figure 3.

s. pulchra culture GK1 7 (Delta de I'Ebro, Spain) coccolith size temp, in °C n mean length (pm) st. dev. mean width (pm) st. dev. 13.7 32 4.6 0.6 3.4 0.4 15.5 49 4.8 0.6 3.5 0.4 16.7 49 4.8 0.5 3.3 0.4 19.6 50 4.5 0.5 3.2 0.5 23.1 49 4.7 0.4 3.3 0.3 all te m p e r a tu r e s 2 2 9 4 .7 0 .5 3 .3 0 .4

S. pulchra coccolith size of cultures at 1 7°C cu ltu re mean length (pm) St. dev. mean width (pm) St. d ev . TW2 (Blanes, W. Mediterranean) 5 0 5 .5 0.6 4 .2 0.6 NAP4 (off Naples, Italy) 50 6.1 0 .5 4 .6 0 .5 NAP2 (off Naples, Italy) 50 5.9 0.6 4 .3 0.6 NAP1 (off Naples, Italy) 50 5.3 0 .5 3 .8 0 .5 NAP10 (off Naples, Italy) 50 5.3 0 .5 3 .8 0 .5 GK7 (Delta de I'Ebro, Spam) 50 5.5 0 .5 4.1 0 .5 GK24 (Delta de I'Ebro, Spam) 50 5 .4 0 .4 3 .9 0 .4 GK15 (Delta de I'Ebro, Spam) 50 5.9 0.4 4 .5 0 .4 GK11 (Delta de I'Ebro, Spam) 50 5.5 0 .5 4 .0 0 .5 8082 (Blanes, W. Mediterranean) 5 0 5 .6 0 .4 3 .9 0 .4 TW11 (Blanes, W. Mediterranean) 50 5.6 0 .5 4 .2 0 .5 GK1 7 all te m p e r a tu r e s 2 2 9 4 .7 0 .5 3 .3 0 .3 all cu ltu res 7 7 9 5 .5 0 .5 4 .0 0 .5

S. pulchra coccolith size of natural samples sa m p le n mean length (pm) St. dev. mean width (pm) St. dev. FBI 6 (N. Atlantic) 50 5.6 0.6 4 .2 0 .5 MATER_1 5_12_B (W. Mediterranean) 99 5.7 0 .7 4 .2 0.6 all wild sa m p le s 1 4 9 5 .7 0.6 4 .2 0 .5

S. pulchra culture GK7 (Delta de 1 Ebro, Spam) coccosphere size after 6 days te m p , in °C mean length (pm) St. dev. mean width (pm) St. d e v . 1 3 .7 20 1 6 .8 2 .4 1 5 .8 2.2 1 5 .5 20 1 7 .6 4 .6 1 6 .8 4.1 1 6 .7 20 1 6 .2 3 .5 1 4 .5 1 .5 1 9 .6 20 16 .1 3 .6 1 3 .8 1.6 all te m p e r a tu r e s 8 0 1 6 .7 3 .5 1 5 .2 2 .4

S. pulchra culture GK17 (Delta de I'Ebro, Spam) coccosphere size after 18 days te m p , in °C n mean length (pm) St. dev. mean width (pm) St. dev. 1 3 .7 20 17 .1 2.6 1 5 .9 2 .5 1 5 .5 20 15.8 2.4 14.4 1.8 1 6 .7 20 16 .1 2 .9 1 5 .1 2 .7 1 9 .6 20 15.6 2.4 13.0 0.6 2 3 .1 20 1 4 .3 2.0 1 3 .0 1.1 all te m p e r a tu r e s 100 1 5 .8 2 .5 1 4 .3 1 .7

Fig. 3. Sample location and statistics for all the cultures and natural samples of S. pulchra studied. 3 R esults 34

The cultures were grown under sterile but not axenic conditions on a tempera­ ture-light gradient table. With the water temperature in the Delta de TEbro showing an annual variation between 13 to 25° C (data from Picco 1990) six temperatures covering a range from 14 to 27° C were selected. The mean irradiance level was between 42 (±3) /^mol/mVsec and a 16L/8D cycle was used. All the cultures were grown in 100 ml glass flasks with 75 ml of K/2 medium. Prior to the experiment the cultures were acclimatised with the first inoculation growing for 37 and the second for another 19 days. During this time temperatures were adjusted and inoculations for batches with no growth were made from the nearest temperature that yielded enough cells. For the final experiment the cultures were grown at six temperatures (13.7, 15.5, 16.7, 19.6, 23.1, 26.7° C) starting off with an initial concentration of 1500 cells per ml (c/ml). Subsequently the cultures were counted every 2-5 days using a solu­ tion of 0.5% Glutaraldehyde in alcohol to immobilise cells and a Sedgwick-Rafter counting cell. With the initial low cell concentrations 50 fields of view (FOV) were counted, later on a count of more than 100 cells was seen to be sufficient. The growth rate k of the cultures in divisions per day was then calculated using the following formula A=log(Ay / Ng)*(3.22l t), with Nj and being the cell concentra­ tions at the end and the beginning of a period time t in days (Guillard 1973). In addition to the cell counts coccosphere size of 20 cells per temperature was measured using an inverted microscope (Olympus BH-2) with an eyepiece grati­ cule shortly after inoculation (day 6) and in the stationary phase (day 18). As the stationary phase was reached cells were harvested, centrifuged and suspended in ethanol before mounting as permanent light microscope mounts with Lakeside. Morphometrical analyses on the slides were then performed under cross polar­ ised light using a Zeiss Axioplan microscope with a Hamamatsu video camera attached.

Physiology (Fig. 4-6)

Both cultures tested (GK7, GK17) grew over a temperature range from 13.7 to 23.1° C (Fig. 4). Cultures repeatedly inoculated at 26.7° C and at 12° C did not grow so the cut-off temperatures will be somewhere between 23.1° and 26.7° C and between 12° and 13.7° C, respectively. Both cultures had their optimum growth temperature at 19.6° C (Fig. 5,6). In both cultures the stationary phase was reached after approximately 13 days (Fig. 4,5) with a growth rate in the exponential phase in culture GK17 ranging from 0.15 divisions per day for the 23.1° C culture up to 0.26 at 19.6° C (Fig. 6). Final cell density in the stationary phase ranged from ca. 9000 cells per ml to ca. 20000 cells per ml (Fig. 3,4). Counts in the stationary phase varied slightly as the cultures had the tendency to form clumps of cells with higher densities in the later stages of the experiment. Due to this in culture GK7 only two temperatures (16.7° and 19.6° C) could be used for calculation of growth rates and the results compare well with culture GK17. 3 Results 35

T in «C day 13.7 0

00 4

for day 0-30 0,05 for day 0-13 0.17 stationary phase not reached for day 13-29 0.01

cAnI * T in "C day Cfhll 15.5 0 0 21

lor day 0-30 lor day 0-13 stationary phase not reached for day 13-29

T in “C day c m * 16.7 0

0 0 0 1 06 0 01 for day 0-30 1240 0 12 for day 0-13 0.22 stationary phase not reached for day 13-29 0.02

T .n -C day c/ml N2 IN I k T .n -C day *

for day 0-30 for day 0-13 stationary phase not reached lor day 13-29

T in “C day c/mi N2 IN I T in "C day c m A 23.1 0 23.1 0

8

for day 0-30 lor day 0-13 13 stationary phase not reached for day 13-29 16

Fig. 4. S. pulchra growth rates for culture GK7 and GK17. For the calculation of growth rates see text.

W ith S. p u lc h ra being one of the few flagellate heterococcolithophores held in culture so far qualitative observation on further aspects of the physiology could be gathered. S. p u lc h ra cultures stored in the incubator exhibited a strong phototaxis, if the cultures were left undisturbed the cells would accumulate in the part of the culture flask closest to the light source in the incubator. Although this seems to be a common phenomenon in other algae (Kawai and Kreimer 2000) this is a novel observation for haptophytes. Cultures kept on the temperature gradient table, how­ ever, showed the tendency to accumulate in the part of the culture flask that was closest to the warm end of the table. This phenomenon - thermotaxis - was tested by repeatedly homogenising the cultures by shaking and waiting until the patch of 3 Results 36

s. p u lc h ra culture GK17 S. p u lc h ra culture GK7

j G'C

23 r c R ? = 0 98 l |2 3 r C R 2 = 0 61 I 15 5° C R 2 = 1 I 15 5' C R 2 = 0 7 4 ( 13 7- C R 2 = 0 86 13 7' C R 2 = 0 99

days

Fig. 5. Logarithmic plot of the cell densities in both cultures. The lines for each tempeature represent best fit, the confidence level is is given as r (left: exponential growth, right: staion- ary phase). Exponential growth lastet between 9-13 days.

S pulchra - culture GK17

8 I- 1I1- I 1 I

Fig. 6. Cell yield and growth rate as a function of temperature. Both parameters show a nar­ row temperature tolerance. cells assembled again. With dense cultures it would normally take less than an hour for the patch of cell to be visible again on the warm side. It was then tested if the cells would accumulate at their optimum growth temperature. This was achieved by using a glass tube laid along the length of the table and so spanning a temperature range of ca. 20 - 30° C. This, however, was not the case, the cells inoculated at the cold end of the table would swim until they reached the warmest part of the tube which happened to be beyond their viable temperature range. As a consequence a dead accumulation of cells was found in this position after 24 hours. This interesting novel phenomenon - and particularly its underlying principle (infrared detection?) - certainly warrants more future research work on S. p u lc h ra and other flagellate haptophytes. 3 R esults 37

Morphology

Coccosphere morphology (Fig. 3) For culture GK17 coccosphere length after 18 days growth ranges from 14.3 - 17.1 pim with mean of 15.8 pim over the whole temperature range. Coccosphere width ranges from 13 - 15.9 pim with a mean of 14.3 pim. For culture GK7 coccosphere length after 6 days growth ranges from 16.1 - 17.6 pim with mean of 16.7 pim over the whole temperature range. Coccosphere width ranges from 13.7 - 16.7 pim with a mean of 15.2 pim. Even given the relatively low precision of the measurements and the relatively vari­ able shape of the S. pulchra cell it is still noteworthy that smallest cell sizes seem to coincide with high growth rates around the temperature optimum.

Coccolith morphology (Figs 3,7-9) For culture GK17 mean coccolith length after 29 days growth varies from 4.5 - 4.8 pim with a mean of 4.7 p m over the whole temperature range. Mean coccolith width ranges from 3.2 - 3.5 p m with an overall mean of 3.3 pm . The distribution of both length and width appears to be unimodal and shows no correlation with temperature/growth rate (Fig. 3,7). 12 further 5. pulchra clones grown at 17° C only revealed a mean coccolith length of 5.3 - 6.1 pm (with an overall mean of 5.6 pm ) and a mean coccolith width varying from 3.8 - 4.6 p m (overall mean 4.1 pm ). Note that unlike in figure 3 these values exclude measurements from culture GK17. Although the distribution of length and width is unimodal there appears to be a variability of morphospace in between cultures (Fig. 3,8). Generally the morphology of the clonal strains com­ pares well with measurements (Fig. 9) taken from two natural samples from the N. Atlantic (FBI6,53° 30’ N 20° 30’ W) and the W. Mediterranean (MATER II, station 15, 35° 54’ N 1° 20’ W). Mean coccolith length observed here was 5.7 pm , mean width 4.2 pm . 3 Results 38

Syracosphaera pulchra culture GK 17

1 3 .7 ° C 1 9 .6 ° C

‘ _ '

1 5 .5 ° C 2 3 .1 ° C

n=49

1 6 .7 ° C all temperatures

Contour lines 2 • max ® 2 intervals Contours < 5 are dasned n=229 30 35 40 45 50 55 60 65 70 75 8 0 lengthin pm

Fig. 7. Density plot for coccolith length and width for S. pulchra culture GK 17 at various temperatures. 3 R esults 39

Syracosphaera pulchra (cultures grown at 17°C)

Coniouf Unes 2 • ma* ® 1 imervai* Contoun < 4 are dashed

GK 17 - all temperatures

Contour lines 2 • ma* Q 2 tntervals Contours < 5 are dashed

all cultures plus G K 17

Contour inas 2 - ma* Q 5 intervals Contours < 5 are dashed

Fig. 8. Density plot for coccolith length and width for selected cultures of S. pu lch ra at 17° C. For comparison the measurements from the temperature experiment were included. 3 Results 40

Syracosphaera pulchra ; natural samples

MATER 15_12B FB 16

5 0

I iÏ

20

3 0 3 5 4 0 4 5 5 0 5 5 8 0 6 5 7 0 7 5 8 0

Fig. 9. Density plot for coccolith length and width for two natural samples of S. pulchra. 3.1.5 Sediment trap samples

To investigate the temporal variation of S. p u lc h ra coccolith size subsamples were taken from a sediment trap mooring off Somalia (MST-9, 10°45’ N 53°34’ E). Water depth was 4050 m with the specific trap used collecting at a depth of 3050 m. For details about the specific trap location and coccolithophore fluxes in a higher trap in the array (MST9-E) refer to Broerse et al. (2000). 14 samples spanning 133 days from the 07/06/1992 until 25/10/1992 were analysed (Fig. 10).

S. pulchra coccolith size in trap MST-9G off Somalia sam ple starting day duration (days) aliquot used SST in "C n mean length (pm) St. dev. mean width (pm) St. dev. G2 0 7 /0 6 /1 9 9 2 14 1 /3 2 0 >27 17 4.8 0.7 3.5 0.5 G3 2 1 /0 6 /1 9 9 2 14 1/640 >22 <27 30 4.9 0.8 3.4 0.5 G4 0 5 /0 7 /1 9 9 2 14 1/1280 >22 <25 30 4.5 0.7 3.3 0.5 G5 1 9 /0 7 /1 9 9 2 7 1 /3 2 0 >22 <25 13 4.4 0.7 3.3 0.5 G6 2 6 /0 7 /1 9 9 2 7 1/320 >22 <25 50 4.5 0.9 3.3 0.6 G7 0 2 /0 8 /1 9 9 2 7 1/320 >22 <25 50 4.7 0.8 3.4 0.5 G8 0 9 /0 8 /1 9 9 2 7 1/320 >22 <25 50 4.6 0.6 3.3 0.4 G9 1 6 /0 8 /1 9 9 2 7 1/320 >22 <25 50 4.9 0.7 3.6 0.6 G11 3 0 /0 8 /1 9 9 2 7 1 /6 4 0 >22 <25 50 5.1 1.0 3.7 0.7 G12 0 6 /0 9 /1 9 9 2 7 1/640 >22 <25 50 5.0 0.8 3.7 0.6 G13 1 3 /0 9 /1 9 9 2 14 1/1280 >22 <25 50 5.0 0.9 3.6 0.6 G14 2 7 /0 9 /1 9 9 2 14 1/1280 >25 50 5.5 0.8 4.0 0.5 G15 11/10/1992 14 1/1280 >25 50 4.9 0.7 3.6 0.5 G16 25/10/1992 14 1 /1 2 8 0 >25 50 5.5 0.8 4.0 0.6 III samples 133 590 4.9 0.8 3.5 0.5

Fig. 10. Statistics for all the sediment trap samples of S. pulchra studied. The sea surface data (SST) is taken from Broerse (2000).

Morphology

Mean coccolith length in the samples varies from 4.9 - 5.5 /im with mean of 4.9 p m over the whole sample period. Mean coccolith width in the samples varies from 3.3 - 4 p m with an overall mean of 3.5 p m (Fig. 10). The morphospace occupied by the measurements covering the whole interval is variable, unimodal samples with low variation in length and width (G3, G4) are followed by bi- to multimodal distribu­ tions with a higher variation in coccolith length and width (Fig. 11). 3 Results 41

Syracosphaera pulchra (sediment trap time series)

G 1 2

G t 5

G i e all samples

1

Fig. 11. Density plot for coccolith length and width for all sediment trap samples (including G2 and G5) o f 5. pulchra studied. Most of the samples reveal a bi- or multimodality, which can be clearly seen if all the measurements are plotted together (bottom right). 3 Results 42

3.1.6 Holocene samples

Seven Holocene coretop samples taken from DSDP/ODP sites covering the North and South Atlantic Ocean, the Indian Ocean and the eastern Pacific Ocean and a boxcore (BC3) from the Meditenanean were analysed (Fig. 12,13). Two more cores have been checked, but for both ODP site 704 (S. Atlantic Ocean) and ODP site 552 (N. Atlantic Ocean) the abundance of S. p u lc h ra coccoliths proved to be too low to allow for a statistical analysis. Note that the sample from ODP site 872 is of early Pleistocene age and was included as a comparison for the Holocene samples.

Fig. 12. Map of the DSD? / ODP cores studied to de­ termine the morphological variation in Holocene S. pu l­ chra. Sample BC 3 is from a boxcore in the Mediterranean and was provided by Utrecht University.

s. pulchra coccolith size in Holocene samples site code location mean length (pm) S t. dev. mean width (pm ) St. dev. ODP 664 108, 664D. 1 HI, 0-1 cm 00.1 ON 023.21 W 30 4,9 0,7 3.5 0.5 ODP 806 130, 8068, IHl, 2-4cm 00.31 S 159.35 E 30 4.9 0.6 3.4 0.5 ODP 872 144, 872A, 1H2, 127cm 10.09 N 162.85 E 20 5.0 0.6 3.6 0.5 BC 3 BC3, EE0019, 6.8-10.4cm 33.37 N 024.76 E 120 5.5 0.8 4,1 0.7 ODP659 108, 659A, IHl, 2-4cm 18.08 N 021.02 W 30 5.2 0.7 3.6 0,4 ODP 709 1 15, 709C, IHl, 3-5cm 03.90 S 060.55 E 29 5.1 0.6 3.6 0.4 ODP 552 81, 552, 1 H 1 ,9 -Ilc m 56.04 M 023.23 W 30 6.0 0.5 4,2 0,5 all sam ples 289 5.2 0.6 3.7 0,5

Fig. 13. DSDP / ODP site location and statistics for all Holocene samples of S. pulchra studied. Morphology

Mean coccolith length in the samples varies from 4.9 - 6 p m with mean of 4.2 p m for all Holocene locations. Mean coccolith width in the samples varies from 3.4 - 4 .2 p m with an overall mean of 3.7 p m (Fig. 13). By analogy with the natural samples the morphospace occupied by the measurements reveals a complex pattern ranging from unimodal distributions in ODP site 664 to bimodal distributions in cores ODP 709, 552, 806 and multimodal distributions in cores BC3 and ODP site 659 (Fig. 14). 3 Results 43

Syracosphaera pulchra Holocene:

ODP 659

ODP 806

OOP 552

8C3 all Holocene samples

3.00 3 50 4 00 4 50 5 00 5 50 oOC (i 50 7 00 7 50 8.00

Fig. 14. Density plot for coccolith length and width for all Holocene samples of S. pulchra studied. 3 Results 44

3.1.7 Downcore samples (ODP site 664)

A series of 14 samples (Fig. 15) taken from ODP site 664 was analysed covering a time interval from 0 to 3.482 Ma. However sample ODP664D 35.82 was ex­ cluded from the analysis due to the low abundance of S. p u lc h ra coccoliths. The age model used is based on few magnetochronological datums (Brunhes/Matuyama and Jaramillo-top and bottom) plus nannobiochronological data mostly from LAD of forms (Su 1996).

S. pulchra coccolith size of ODP site 664 sam ples sample core location sub4jottom depth (m) age in Ma n mean length (pm) S t. dev. mean width (pm) St. dev. ODP664D0.01 0.1 N 23.21 W 0.0 0.0 30 4.9 0.7 3.4 0.5 ODP664D1.52 1.5 0.0 30 5.1 0.6 3.7 0.4 ODP664D3.32 3.3 0.1 30 4.8 0.6 3.4 0.5 ODP664D8.52 8.5 0.3 30 4.7 0.7 3.4 0.6 OOP664D12.02 12.0 0.4 30 4.7 0.6 3.5 0.5 ODP664D15.02 15.0 0.4 25 4.6 0.6 3.5 0.5 ODP664D35.82 35.8 1.0 2 no analysis ODP664D45.92 45.9 1.2 30 4.7 0.5 3.2 0.3 ODP664D60.44 60.4 1.6 30 5.9 1.0 4.3 0.7 ODP664D104.42 104.4 2.4 30 4.8 0.5 3.4 0.4 ODP664D109.34 109.3 2.5 30 4.9 0.9 3.5 0.7 ODP664D1 25.84 125.8 2.9 30 5.1 0.5 3.6 0.4 ODP664D1 39.02 139.0 3.2 26 5.1 0.8 3.7 0.6 ODP664D1 49.23 149.2 3.5 16 5.3 0.5 3.7 0.5 all sam ples 367 5.0 0.6 3.6 0.5

Fig. 15. ODP site 664 location and statistics for all samples of S. pulchra studied. Morphology

Mean coccolith length in the samples varies from 4.7 - 5.9 p m with mean of 5 p m for all time slices. Mean coccolith width in the samples varies from 3.2 - 4.3 //m with an overall mean of 3.6 p m (Fig. 15). The time series reveals no clear (gradu- alistic) morphological evolution, most of the measurements occupy a bimodal mor­ phospace whereas the two youngest samples show a unimodal distribution of the measurements (Fig. 16). 3 R esults 45

Syracosphaera pulchra 0.3 5 7 Ma 2.5 2 9 Ma morphology of coccoliths In ODP site 664

Ctrlouf inw 0 5 • max @0 5 mtwvas Contours < t 5 ar« dashed 0 44 6 Ma 2 .8 5 2 Ma

1 242 Ma 3 .207 Ma

0 .253 Ma

I 4 I

3,0 35 4.0 4.5 50 5,5 6,0 6 5 7.0 7 5 8,0

Fig. 16. Density plot for coccolith length and width for ail samples of S. pu lch ra studied from ODP site 664. 3 Results 46

3.1.8 First occurrence

The first occunence of S. p u lc h ra in the fossil record was dated at approximately 13 Ma at two ODP sites (ETH Neptune database, DSDP, leg 74, site 525, 34cc in Jiang and Gartner 1984) and (Takayama and Sato 1987). The timescale used for the Neptune database and for the nannofossil zones in figure 17 is that of Berggren et al. (1995). The published biostratigraphic tables for the two sites with the oldest first occurrence of S. p u lc h ra show a continuous record from the Holocene down to Early Pliocene nannofossil zone NN14 (site 608) and Late Miocene zone NNl 1 (site 525B) with singular observation of S. p u lc h ra in NN7 (site 525B) and NN5 (site 608). However re-examination of the DSDP/ODP reference collection of site 525 at the NHM revealed the first unambiguous S. p u lc h ra coccolith in a sample (DSDP, leg 74, hole 525, 12-1,9-10), which is of early Pliocene age (4.8 Ma). These findings compare well with the frequency data obtained from the 900 data entries for S. p u lc h ra from 28 DSDP/ODP sites in the Neptune database (Eig. 17), so we believe that the oldest records are probably erroneous possibly due to contamination or species misidentification.

Cumulative occurences of Spulchra in 0 5 IVIa time intervals in ODP / DSDP sites (data from ETH Neptune database with 900 observations from 28 sites)

0 S tag e s Piacenzian | Zanciian Messinian | Totlonian Serravalian (pars) Epochs Pleistocene Pliocene Late Miocene Middle Miocene jge Ma 0

Fig. 17. Frequency plot of occurrences of S. pulchra in 29 DSDP / ODP cores with a total of 900 observations (from the BTH Neptune database). The Middle Miocene occurrences are probably erroneous, own observations on material from DSDP site 525 point to 4.8 Ma for the first occurrence o f unambiguous S. pulchra coccoliths. However continuous records have been reported for a number of other DSDP / ODP sites, dating back to 7 Ma. This interval has been highlighted by the green lines. The timescale used for the Neptune database and for this figure is that o f Berggren et al. (1995). 3 R esults 47

3.1.9 Conclusions

The key observations on S. pulchra are:

1. Measurements of coccoliths from cultured 5. pulchra reveal a unimodal, well defined morphospace which is contrasted by a bi- or multimodal distribution in natural, sediment trap, Holocene and time series samples. 2. There is evidence from life cycle observations, both in culture and in natural samples that S. pulchra comprises at least two cryptic biological species. 3. In the monoclonal cultures used there was no temperature control on either cell or coccolith size. 4. Compared with other coccolithophore species previously held in culture growth rates are relatively low, and optimum growth is at a well defined temperature. 5. The presence of photo- and especially thermotaxis is a novel phenomenon in the haptophytes.

At first the results from the morphological work on S. pulchra seem confusing as physiological and morphological results from the culture work indicate a morpho­ logically very well defined species, with stable morphology in clonal culture and a marked temperature preference which is contrasted by the morphological varia­ tion observed in the sediment trap samples, the natural samples and the Holocene sediments. However the life cycle observations indicate the presence of (at least) two different cryptic species, which had previously been lumped together into S. pulchra. However if the means of coccolith length and width for the one culture with a known life cycle phase (NAP 10 with D. pirus) and a natural sample (FBI6 from the North Atlantic Ocean) with 5. pulchra and C. oblonga (Fig. 18) are su­ perimposed on the density plot for the other investigated cultures it seems to be possible to divide the cultures up into two clusters which tentatively represent the two cryptic species (Fig. 19). Culture GK 17, which was used for the temperature experiments fits in neither of the modes and so there is tentative evidence of even more cryptic species nesting in S. pulchra. This agrees well with measurements of coccoliths on combination cells, using a SEM (Fig. 20). If the same method is applied to the sediment trap samples the two species can be identified as well although additionally there is a clear peak for a smaller (morpho-) species which can be tentatively interpreted as a third cryptic species (Fig. 21). Currently we are trying to test these findings with molecular data and culture samples of each studied strain were grown. This is done in collaboration with Colomban De Vargas (Rutgers University). Preliminary results using the fast evolving TufA gene indicate clear genetic differences between strains. The interesting topic of the timing of the divergence of the cryptic species in S. pulchra remains under discussion. Due to preservation problems the fossil record of the two holococcolith types involved is too sporadic and the few records are of Quaternary age only. Equally the morphological variation in the heterococcolith stage is too slight and tentative to be applied to the fossil record. It can however 3 Results 48 be assumed that unless parallel evolution has occurred divergence must post date acquisition of morphological characters which allow to recognise S. p u lc h ra and hence the maximum divergence time is that of the FO of S. p u lc h ra, which is dated here at 4.8 Ma (see 3.1.8). A working hypothesis therefore is that S. p u lc h ra consists of set of locally adapt­ ed, recently evolved genotypes with very similar, strongly overlapping coccolith morphologies, each adapted to specific ecological conditions. Heterococcolith mor­ phology of a single genotype therefore appears to be less plastic than previously as­ sumed. The use of morphometrical analyses to reconstruct the evolutionary history of this species reveals the same pattern observed in recent samples. Due to the poor fossil record of the holococcolithophores associated with S. p u lc h ra their FO cannot be dated precisely and unless specific (morphological) markers to distinguish differ­ ent genotypes are found the morphology of S. p u lc h ra coccoliths is of little use to reconstruct the evolutionary history of S. pulchra. s. pulchra A phase change to D pirvs observed in culture

natural sam ple - MATER 15_12B

B natural sample • FB 16 -’pie with cn y C oblonga present

E ' all cultures plus GK17

I

C natural sam ple - MATER 15_12B natural sample witft both D pirus and C oblonga present

I

Fig. 18. Measurements on samples of the heterococcolith stage of S. pulchra. A. Culture NAP 10 in which a life cycle change to the holococcolithophore D. piru.s was observed. B. Natural sample FB 16 from the North Atlantic. Only the holococcolithophore C. oblonpa was present in this sample, so it has been assumed that no 5. pulchra forming life cycle associa­ tions with D. pirus are present. C. This sample contains both holococcolithophores. In D. The means taken from A and B are superimposed on the natural sample with both holococcolitho­ phores. The means fall exactly in the two maxima on the density plot, allowing for a tentative discrimination of the two cryptic species in the available cultures (E). 3 R esults 49

Syracosphaera pulchra NAP 4 NAP 2 (cultures grown at 17°C)

' I G K 15 TW 11

f

G K 24 G K 7

BOB 2 1 T W 2

% 1

...... ~50|

NAP 1 GK 11 GK 17 (all lem peralures)

Fig. 19. Means taken from ligure 18 were used for a tentative discrimination of the two cryp­ tic species in all the measured cultures of S. pu lch ra. Note that culture GK 17 lits into neither node and might represent a third cryptic species. 3 Results 50

Coccolith length & width for S. pulchra combinations

pirus oblonga

6.5 -

i 5.5

5 -

4.5

n = 44 (S. pulchra / O. pirus) n = 13 (S. pulchra i C. oblonga)

2.5 3.5 4.5 width in |im X -

Fig. 20. Length and width of endothecal heterococcoliths of Syracosphaera pulchra m eas­ ured on scanning electron micrographs of seven combination cells. A total of 44 coccoliths on combinations with Daktylethra pirus and 13 on combinations with Calyptrosphaera oblonga were measured. All measurements occupy the same morphospace, with the ones taken on combination cells with C. oblonga showing slightly higher mean coccolith lengths and widths. The means and the standard deviation for the measurements on the two types of combination cells are displayed near the axis.

Fig. 21 (next page). Means taken from figure 18 were used for a tentative discrimination of the two cryptic species in all the measured sediment trap samples of S. pulchra. 3 Results 51

Syracosphaera pulchra 0 3 0 4 trap MST-9 off Somalia

0 8 0 9 011

0 6 0 7 012

0 13 0 1 6 0 1 5

0 1 4 all samples

I 4 3 R esults 52

3.1.10 References

Baumann K-H, Sprengel C (2000) Morphological variations of selected coccolith species in a sediment trap north of the Canary Islands. J nannoplankton Res 22; 185-193 Berggren W A, Kent D V, Swisher C C, Aubry M-P ( 1995) A revised Cenozoic Geochronology and Chronostratigraphy. Geochronology Time Scales and Global Stratigraphie Correlation, SEPM Special Publication 54: 129-212 Borsetti A M, Cati F (1972) 11 Nannoplancton caicareo vivente nel Tirreno Centro-meridion- ale. G Geol 38: 395-452 Borsetti A M, Cati F (1976) 11 Nannoplancton caicareo vivente nel Tirreno Centro-meridion- ale. Parte 11. G Geol 40: 209-240 Broerse ATC, B rummer G-J A, Hinte J E v (2000) Coccolithophore export production in response to monsoonal upwelling off Somalia (northwestern Indian Ocean). Deep-Sea Res 11 47: 2179-2205 Cros L (2000) Variety of exothecal coccoliths of Syracosphaera. J nannoplankton Res 22: 41-51 Cros L (2002) Planktic coccolithophores of the NW Mediterranean. PhD thesis, Universitat de Barcelona (ISBN 84-475-2680-1), p. 181 Cros L, Kleijne A, Zeltner A, Billard C, Young J R (2000) New examples of holococcolith- heterococcolith combination coccospheres and their implications for coccolithophorid biology. Mar Micropaleontol 39: 1-34 Gaarder K R, Heimdal B R (1977) A revision of the genus Syracosphaera Lohmann (Coccolithineae). Meteor ForschErgebn D 24: 54-71 Geisen M, Billard C, Broerse ATC, Cros L, Probert 1, Young J R (2002) Lifecycle associa­ tions involving pairs of holococcolithophorid species: intraspecific variation or cryptic spéciation? Eur J Phycol 37: 531-550 Geisen M, Young J R, Probert 1, Saez A G, Baumann K-H, Bollmann J, Cros L, De Vargas C, Medlin L K, Sprengel C (submitted) Species level variation in coccolithophores. In: Young J R (ed) Cocoolithophores - From molecular processes to global impact. Springer, pp XXX XXX Guillard R R L (1973) Division rates. In: Stein J R (ed) Handbook of phycological methods Culture methods & growth measurements. Cambridge University Press, Cambridge, pp 289-311 Halldal P, Markali J (1954) Observations on Coccoliths of Syracosphaera mediterranea Lohm ., S. pulchra Lohm. and S. molischii in the Electron Microscope. J Cons perm int Explor Mer 19: 329-336 Halldal P, Markali J (1955) Electron microscope studies on coccolithophorids from the Norwegian Sea, the Gulf Stream and the Mediterranean. Avh Norske VidenskAkad Oslo 1: 1-30 Heimdal B R (1993) Modem Coccolithophorids. A Guide to Naked Flagellates and Coccolithophorids. In: Tomas C R (ed) Marine Phytoplankton. Academic Press, London, pp 147-243 Inouye 1, Pienaar R N (1988) Light and electron microscope observations of the type species o f Syracosphaera, S. pulchra (Prymnesiophyceae). Br phycol J 23: 205-217 3 R esults 53

Jiang M-J, Gartner S (1984) Neogene and Quartemary calcareous nannofossil biostratig­ raphy of the Walvis Ridge. In: Moore T C, Rabinowitz P D, Boersma A, Borella P E, Chave A D, Duée G, Fütterer D K, Jiang M-J, Kleinert K, Lever A, Manivit H, O Connell S, Richardson S H, Shackleton N J (eds) Initial Rep deep Sea Drilling Proj U.S. Government Printing Office, Washington, pp 561-595 Kamptner E (1937) Neue und bemerkenswerte Coccolithineen au s dem Mittelmeer. Arch Protistenkd 89: 279-316 Kamptner E (1941) Die Coccolithineen der Südwestküste von Istrien. Annln naturh Mus Wien 51: 54-149 Kawai H, Kreimer G (2000) Sensory mechanisms. Phototaxis and light perception in algae. In: Leadbeater BSC, Green J C (eds) The flagellates: unity, diversity and evolution. Clarendon Press, Oxford, pp 124-146 Kleijne A (1991) Holococcolithophorids from the Indian Ocean, Red Sea, Mediterranean Sea and North Atlantic Ocean. Mar Micropaleontol 17: 1-76 Kleijne A (1993) Morphology, taxonomy and distribution of extant coccolithophorids (Calcareous nannoplankton). PhD thesis. Free University Amsterdam (ISBN 90- 9006161-4), p. 321 Lecal-Schlauder J (1961) Anomalies dans la composition des coques de flagelles calcaires. Bull Soc Hist nat Afr N 52: 63-66 Loeblich A R, Tappan H (1963) Type fixation and validation of certain calcareous nanno­ plankton genera. Proc Biol Soc Wash 76: 191-198 Lohmann H (1902) Die Coccolithophoridae, eine Monographie der Coccolithen bildenden Flagellaten, zugleich ein Beitrag zur Kenntnis des Mittelmeerauftriebs. Arch Protistenkd 1: 89-165 Nishida S (1979) Atlas of Pacific Nannoplanktons. News Osaka Micropaleontol Special Paper: 1-31 Norris R E (1985) Indian Ocean nanoplankton. II. Holococcolithophorids (Calyptrosphaeraceae, Prymnesiophyceae) with a review of extant genera. J Phycol 21: 619-641 Okada H, Honjo S (1973) The distribution of oceanic coccolithophorids in the Pacific. Deep- Sea Res 20: 355-374 Okada H, McIntyre A (1977) Modem coccolithophores of the Pacific and North Atlantic Oceans. Micropaleontology 23: 1-55 Picco P (1990) Climatological atlas of the Western Mediterranean. Italian commission for nuclear and alternative energy sources, Rome Saugestad A H, Heimdal B R (2002) Light microscope studies on coccolithophorids from the western Mediterranean Sea, with notes on combination cells of Daktylethra pirus and Syracosphaera pulchra. Plant biosyst 136: 3-28 Schiller J (1925) Die planktonischen Vegetationen des adriatischen Meeres. A. Die Coccolithophoriden-Vegetation in den Jahren 1911-14. Arch Protistenkd 51: 1-130 Su X (1996) Development of late Tertiary and Quaternary coccolith assemblages in the Northeast Atlantic. Geomar Report 48: 1-120 3 R esults 54

Takayama T, Sato T (1987) Coccolith biostratigraphy of the North Atlantic Ocean, Deep Sea Drilling Project Leg 94. In: Ruddiman W F, Kidd R B, Baldauf J G, Clem ent B M , Dolan J F, Eggers M R, Hill P R, Keigwin J D j, Mitchell M, Philipps I, Robinson F, Salehipour S A, TakayamaT, Thomas E, Unsold G, Weaver P PE (eds) Initial Rep deep Sea Drilling Proj U.S. Government Printing Office, Washington, pp 651-702 Young J R, Davis S A, Bown P R, Mann S (1999) Coccolith ultrastructure and biomineralisa­ tion. J struct Biol 126: 195-215 3 R esults 55

3.2 Spéciation in the HeUcosphaera plexus

3.2.1 Introduction

HeUcosphaera carteri was described by in 1877 by Wallich (1877) a retired surgeon of H.M. Indian Army who developed considerable skill in microscopy. Rice et al. (1976) give an interesting account of Wallich's life and his contribution to oceanography. Wallich named the species after Mr. Carter, who in 1871 supports Huxley's earlier view (in notes on his observations from the H.M.S. Cyclops cruise in 1857) that coccoliths are indeed produced by unicellular algae (Carter 1871) as opposed to forming part of the infamous Bathybius (Huxley 1868), Hackel's “Urschleim”. Equally, in a personal communication to Wallich as cited in Wallich (1877), Carter correctly interprets Wallich's drawings of textulariid foraminifera covered with coccoliths as coincidental (Young and Geisen 2002). According to Jordan and Young (1990) the extant cosmopolitan genus HeUcosphaera comprises two species {H. carteri and H. pavimentum) with three varieties in H. carteri (Jordan and Green 1994; Jordan and Kleijne 1994). The three described varieties in HeUcosphaera carteri - H. carteri var. carteri, H. carteri var. wallichii and H. carteri var. hyalina - have been considered either as separate taxa (Jordan and Young 1990) or as morphological extremes of intergradational morphotypes (Nishida 1979). However - with our new results from morphological work (this study), molecular data (Saez et al. 2003) and life cycle studies (Cros et al. 2000; Geisen et al. 2002) on culture and natural material - there seems to be sufficient evidence that the three varieties are in fact three separate, albeit closely related, species. Following the normal rules of nomenclature this would result in four extant species, H. carteri, H. hyalina, H. walllichii and H. pavimentum (see Taxonomy section). We will use the new species names in this manuscript and will only refer to the previous names where necessary in historical context. Equally the species names for the invalidated holococcolithophore life cycle stages are given as heterotypic synonyms in the taxonomy section, but for the plates the nomenclature suggested by Cros et al. (2000) and Geisen et al. (2002) is used. HeUcosphaera is present in all extant oceans (Okada and McIntyre 1977; Winter et al. 1994) and common in the sedimentary record - in a review of the genus HeUcosphaera Jafar and Martini (1975) list 25 extant and fossil species. For accounts of the fossil biodiversity of HeUcosphaera see also Theodoridis (1984) and Aubry (1990). According to Young (1998) the most common variant H. carteri (as H. carteri var. carteri) has a first occurrence of Late Oligocene age and H. wallichii (as H. carteri var. wallichii) of Late Miocene (Tortonian) age. These ages correlate well with our own review of data from the ETH Neptune database (see 3.2.11). Although HeUcosphaera had been successfully cultured before and cytological sections had been cut and analysed to evaluate the microanatomy of the organism this work by Inouye (pers. comm.) - until now - remains unpublished. Re­ 3 R esults 56 examination of a sample from this culture material however proved to be essential for our interpretation of biodiversity in the genus HeUcosphaera. With a good fossil record, a demonstrated ability to be held in culture (Inouye, unpublished data) the HeUcosphaera carteri variants, as representatives of the family Helicolithaceae, were an obvious choice to act as a keystone species in CODENET. Along with other studies on HeUcosphaera in CODENET, focusing on molecular biology (Saez et al. 2003), cytological work and photosynthetic pigments, this study deals mainly with morphological variation of HeUcosphaera coccoliths, with the physiology of the organism and provides new insight in HeUcosphaera life cycles and spéciation.

3.2.2 Taxonomy and description ofHeUcosphaera carteri

HeUcosphaera carteri (Wallich) Kamptner (Plate 1, figs 1-3) Coccosphere Huxley (1868), p. 209, figs 6 (d, e) non 6 (a, b, c). Coccosphaera carterii sp. nov. Wallich (1877), p. 348, pi. 17, figs 3, 4 non 6, 7, 17. Coccosphaera pelagica var. carterii Ostenfeld (1899), p. 436. Coccolithophora pelagica Lohmann (1902), pp. 138-139. Coccolithophora pelagica Lohmann (1919), pp. 13-14, fig. 21b. Coccolithus pelagicus Schiller (1930), pp. 246-247, figs. 124a non 123 (a-d). Coccolithus carteri (Wallich) Kamptner in Kamptner (1941), pp. 93-94, 111-112, pi. 12, fig. 134, pi. 13, figs 135, 136. HeUcosphaera carteri (Wallich) comb. nov. Kamptner (1954), pp. 21-23, figs 17-19; Gaarder (1962), pp. 114-119, figs 1 (h), 2 (e, f); Borsetti et Cati (1972), p. 505, pi. 52, figs 1,2; Jafar et Martini (1975), pp. 381-397, pi. 1, figs 1,4, 5; Nishida (1979), pi. 9, figs 4 (a-d); Heimdal (1993), p. 215, pi. 5; Kleijne (1993), pp. 232-233, pi. 1, fig 7 non fig. 8; Winter et Siesser (1994), pp. 121-122, figs 23 (a, b); Young (1998), p. 236, pi. 8.1 figs 1, 2.; Geisen et al. (2002), p. 537, fig. 2. Helicopontosphaera kamptneri gen. nov. sp. nov. Hay et Mohler in Hay et al. (1967), p. 448, pi. 10-11, fig. 5. HeUcosphaera sp. Nishida (1979), pi. 9, fig. 3. HeUcosphaera carteri var. carteri (Wallich) nov. var. Theodoridis (1984), pp. 132- 133, pi. 23, fig. 5; Jordan et Young (1990), pp. 15-16. HeUcosphaera kamptneri Perch-Nielsen (1985), pp. 485-494, figs 45 (25, 27, 28), 46 (4).

Heterotypic synonyms

Syracolithus catilliferus (Kamptner) Deflandre (as H. carteri solid type - Plate 1, figs 10, 11) Syracosphaera catillifera sp. nov. Kamptner (1937), p. 301, pi. 14, figs 10, 11. Syracosphaera {Syracolithus) catillifera Kamptner (1941), pp. 81, 101-103, pi. 4, 3 Results 57 figs 43-45. Syracolithus catillifera (Kamptner) comb. nov. Deflandre (1952), p. 453, figs 351 (c, d). Calyptrosphaera catillifera (Kamptner) comb. nov. Gaarder (1962), pp. 36-38, pi. 1, figs. a, b. Calyptrolithophora catillifera (Kamptner) comb. nov. Norris (1985), p. 626, fig. 33. Syracolithus catilliferus (Kamptner) Deflandre (1952), p. 453, fig. 351 c, d; Kleijne (1991), p. 34, pi. 6, figs 1,2; Geisen et al. (2002), p. 537, fig. 3.

Syracolithus confusus Kleijne (as H. carteri perforate type - Plate 1, fig. 12) Syracolithus confusus sp. nov. Kleijne (1991), pp. 52,55, pi. 6, figs 3-5.

Note: The 5. confusus - S. catilliferus combination cells (p. 537, figs 4-6 in Geisen et al. 2002) would belong to both heterotypic synonyms.

HeUcosphaera spp. - heterococcolith morphology

I I G

Fig. 1. Morphology of the three HeUcosphaera spp. in the SEM and LM. The central area features - like the absence or presence of pores or slits and their orientation - are stable char­ acters that can be used for species identification and discrimination. 3 Results 58

Description of the heterococcolithophorid H. carteri

The coccosphere is ellipsoidal with (020) spirally arranged asymmetrical helicoliths (Fig. 1-A; Plate 1, fig. 1). The helicoliths possess a bar, separating 1-2 openings (pores or slits) in the central area aligned with the long axis. The helical flange is well developed, ending in a wing (Fig, 1-C; Plate 1, figs 2,3). The cell bears two flagellae and a haptonema and the flagellar pole is surrounded by modified coccoliths, these are usually larger, with a larger wing bearing small, tooth-like protrusions (Plate 1, fig 1). The outer rim of the coccolith consists of V-units whilst R-units form the central part of the proximal shield. Outgrowths from the R-unit extend across the distal shield to form a blanket that resembles a large number of rhombic laths (Fig. 1-B), but is in fact made up of relatively few elements characterised by numerous steps (Henriksen et al. 2004).

3.2,3 Taxonomy and description ofHeUcosphaera wallichii

HeUcosphaera wallichii (Lohmann) Okada et McIntyre (Plate 1, figs 4-6) Coccolithophora wallichii sp. nov. Lohmann (1902), p. 138, pi. 5, figs 58, 58b, 59 non 60. Coccolithus wallichii Schiller (1930), pp. 247-248, fig. 124c non 124b. HeUcosphaera wallichii (Lohmann) comb. nov. Okada et McIntyre (1977), pp. 14- 15, pi. 4, fig. 8; Perch-Nielsen (1985), pp. 485-494, figs 9, 10. HeUcosphaera carteri var. wallichii (Lohmann) nov. var. Theodoridis (1984), pp. 133-134, pi. 23, figs 8, 9, pi. 27, fig. 7; Jordan et Young (1990), pp. 15-16; Young (1998), pi. 8.1, fig. 6.

Heterotypic synonyms

Syracolithus dalmaticus (Kamptner) Loeblich et Tappan (Plate 1, figs 14-16) Syracosphaera dalmatica sp. nov. Kamptner (1927), p. 187, fig. 2. Syracosphaera {Syracolithus) dalmatica Kamptner (1941), pp. 81, 104, pi. 4, figs 46-48. Syracolithus dalmaticus (Kamptner) Loeblich et Tappan (1966); Kleijne (1991), p. 37, pi. 7, fig. 1; Winter et Siesser (1994), p. 147, fig. 160.

Description of the heterococcolithophorid H. wallichii

The coccosphere is ellipsoidal with (020) spirally arranged asymmetrical helicoliths (Plate 1, fig 3). The helicoliths possess a bar spanning the short axis of the central area, separating 2 dextrally (as seen from distal view) angled aligned slits. The bar can be angled dextrally with respect to the long axis and the ends of the slits are often kinked further dextrally (Fig. 1-D; Plate 1, figs 5-6). The helical flange is well developed, ending in a wing. The cell bears two flagellae and a haptonema and the flagellar pole is surrounded by modified coccoliths, these are usually larger, with a 3 R esults 59 larger wing bearing small, tooth-like protrusions (Plate 1, fig. 3). The outer rim of the coccolith consists of V-units whilst R-units form the baseplate and extend to form a blanket of small elements.

3.2.4 Taxonomy and description ofHeUcosphaera hyalina

HeUcosphaera hyalina Gaarder (Plate 1, figs 7-9) HeUcosphaera hyalina sp. nov. Gaarder (1970), pp. 113-119, figs 1 (a-g), 2 (a-d), 3 a; Borsetti et Cati (1972), p. 406, pi. 53, figs 3, 4; Nishida (1979), pi. 9, fig. 1; Theodoridis (1984), p. 134; Heimdal (1993), p. 215, pi. 5. HeUcosphaera carteri var. hyalina (Gaarder) nov. var. Jordan et Young (1990), pp. 15-16; Young (1998), pi. 8.1, fig. 5. HeUcosphaera carteri (Wallich) Kamptner; Kleijne (1993), pp. 232-233, pi. 1, fig.

Description of the heterococcolithophorid H. hyalina

The coccosphere is ellipsoidal with (017) spirally arranged asymmetrical helicoliths (Plate 1, fig 7). The helicoliths possess thin, tangentially arranged needle shaped elements in the central area - a character unknown in H. carteri and H. wallichii. Openings are always absent. The helical flange is well developed, ending in a wing (Fig. 1-D; Plate 1, figs 8-9). The cell bears two flagellae and a haptonema and the flagellar pole is surrounded by modified coccoliths, these are usually larger, with a larger wing bearing small, tooth-like protrusions. The outer rim of the coccolith consists of V-units and R-units form the baseplate and extend to form a blanket of small elements. 3 R esults 60

Plate 1 (next page) HeUcosphaera spp. Fig. 1: Scanning electron micrograph of a H. carteri coccosphere. The helicoliths show the typical spiral arrangement and possess enlarged flanges in the circumflagellar coccoliths. The central area of this specimen shows the typical morphology with two aligned slits, which are separated by a bar. Morphotypes with 1-2 pores and intermediate central area morphologies have also been observed. Note the little triangular protrusions on the flange. Water sample, N. Atlantic, Portuguese shelf, RA^ Andromeda cruise CODENET 2, station 6. Fig. 2: Scanning electron micrograph of a H. carteri coccolith in proximal view. Sediment trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 3: Scanning electron micrograph of a H. carteri coccolith in distal view. Sedim ent trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 4: Scanning electron micrograph of a coccosphere of H. wallichii. The central area shows the typical morphology with two angled slits with kinked ends which are separated by a bar. This fine morphological feature is stable in culture. Note the little triangular protrusions on the flange. Water sample, western Pacific, Miyake-jima island, Miike Port, Japan. Fig. 5: Scanning electron micrograph of a H. wallichii coccolith in proximal view. Note the kinked ends of the aligned slits. Water sample, western Pacific, Miyake-jima island, Miike Port, Japan. Fig. 6: Scanning electron micrograph of a H. wallichii coccolith in distal view. Sediment trap sample, Indian Ocean, off Somalia. Fig. 7; Scanning electron micrograph of a H. hyalina coccosphere. The central area is filled with tangentially arranged needle shaped elements. Note the little triangular protrusions on the flange. Culture sample (NAP II), Mediterranean, off Naples, Italy. Fig. 8, 9; Scanning electron micrograph of a H. hyalina coccolith in proximal view (fig. 8) and distal view (fig. 9). Culture sample (NAP 11), Mediterranean, off Naples, Italy. Fig. 10: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. This stage was previously described as “Syracolithus catilliferus” and is referred to as H. carteri HO solid-type. Water sample, N. Atlantic, off Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 11: Scanning electron micrograph of H. carteri coccoliths in the holococcolith stage. Water sample, Antarctic Ocean, cruise JR 48. Fig. 12: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. This stage was previously described as “Syracolithus confusus” and is referred to as H. carteri HO perforate-type. Water sample, western Mediterranean, off Barcelona. Fig. 13: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. Note that presence of coccoliths of both H. carteri HO solid and perforate. This is seen as an example of intraspecific variation in the degree of calcification. Water sample, NW Mediterranean, cruise MESO-96, station F2. Fig. 14: Scanning electron micrograph of a 5. dalmaticus holococcolithophore. Water sam­ ple, western Pacific, Miyake-jima island, Ibo Port, Japan. Fig. 15: Scanning electron micrograph of a detail of S. dalmaticus holococcoliths. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 16: Scanning electron micrograph of a tentative H. wallichii - S. dalmaticus com bina­ tion coccosphere. Water sample. Gulf of Mexico, R/V Gyre cruise 92-G-03, station 9.

Scale bars represent: Figs 1, 4, 7, 10, 12-14: 5/

3. H. carten coccolith 1. H. carten coccosphere 2. H. carten coccolith proximal view distal view

4. H. waw/cn//coccosphere 5. H. wallichii coccolith 6. H. wallichii coccolith proximal view distal view

7. H hyalina coccosphere 8. H. hyalina coccolltti 9. H hyalina coccolith proximal view distal view

10. H. caiten holococcoli- 11. H. carten holococcoliths 12. H. carteri holococcoli- 13. H. carteri holococcolitho thophond solid type solid type thophorid perforate type phorid solid & perforate type

14. S. dalmaticus holococcoli- 15. S. dalmaticus ho ococcoliths 16. H. wallichii - S. dalmaticus thophond possible combination coccosphere 3 Results 62

3.2.5 Remarks on the genusHeUcosphaera

Although coccoliths and coccospheres of the genus HeUcosphaera can easily be distinguished from other species due to their helical flange and their comparatively large size there has been much confusion and instability in the nomenclature. For reviews of the genus HeUcosphaera see Jafar and Martini (1975) and Aubry (1 9 9 0 ). Part of the confusion stems from Wallich's type description of HeUcosphaera carteri. W allich ( 1877) clearly illustrates cells of both C. pelagicus (as Coccosphaera pelagica) and H. carteri (as Coccosphaera carterii) and rightly points out that he did not observe intermediate forms between the oblong cells of H. carteri and the spherical cells of C. pelagicus. However, although Wallich knew about the helical rim o f HeUcosphaera - as his drawings of a specimen from the Indian Ocean in his unpublished notes clearly illustrate - he nevertheless appears to fail to draw (and so to draw attention to) the helical rim of HeUcosphaera by only illustrating the central area structures in the type description (Wallich 1877). Furthermore his cross sections of HeUcosphaera clearly resemble Coccolithus cross sections, probably due to the low power of the microscope used for his studies. As a result C. pelagicus and H. carteri were often confused in the early work on the species. Lohmann (1902) for example combines C. pelagica and C. carterii, probably due to the fact that Lohmann, who worked intensively in the eastern Mediterranean

Fig. 2. Notes from Wallich's field book. The helical shape of the coccoliths on the HeUcosphaera sp. in the drawing can be clearly seen. However if only the central area of a single coccolith is drawn HeUcosphaera can be confused with C. pelagicus. 3 R esults 63 where C. pelagicus is extremely rare, was not too familiar with the appearance of C. pelagicus. In Lohmann (Lohmann 1919) he states that the closed, spherical coccosphere of C. pelagicus and the elongate coccosphere of H. carten are in fact two morphological expressions of a single species. Kamptner (1941) however begs to differ and splits the two species again, pointing out that whilst C. pelagicus cells have an elliptical shape C. carterii coccoliths have a deformed (helical) outline, and consequently - with the help of a polarising light microscope - Kamptner (1954) assigns the new genus name HeUcosphaera (with Coccosphaera carterii becoming the generotype) on the basis of the presence of a “spiral” (helical) flange. Hay et al. (1967) are convinced that as the specimen illustrated by Wallich (Wallich 1877) as C. carterii is in fact C. pelagicus and as the generotype of HeUcosphaera was C. carterii by monotypy (Kamptner 1954) the concept of a helical-rimmed genus HeUcosphaera became invalidated, hence they assigned the new genus and species Helicopontosphaera kamptneri. Re-examination of Wallich "s original drawings and slides archived at The Natural History Museum in London revealed that he indeed illustrated the helical outline of HeUcosphaera on his type figures, albeit very thin (Fig. 2). We have therefore come to the conclusion that Helicopontosphaera kamptneri is a junior synonym of H. carteri, as argued by Jafar and Martini (1975). Extensive work done on culture and natural material did prove the morphology of the H. carteri var. carteri and H. carteri var. hyalina to be stable. This has provided sufficient evidence that the fine scale morphological variation between the, variants as described in Jordan and Young (1990) is not intergradational and can be used to discriminate the taxa. Hence the two variants were risen to species rank, resulting in H. carteri and H. hyalina respectively (Saez et al. 2003). The third variety mentioned by Jordan and Young (1990) H. carteri var. wallichii, has been cultured in the past (Young, unpubl. obs. using material supplied by Inouye) and has a stable morphology both in natural samples and cultures without intermediate morphotypes, so it also seems justified to reinstate this variety to species rank. All the early observations, including the type description, were conducted using light microscopy and documented with drawings or, later in the century, with photo documentation (Kamptner 1954). Some of the type material examined by Wallich is in the collection of the BMNH {Bulldog cruise, 1860, plus additional slides from the S. Atlantic collected on his return from India) and can be examined (e.g. Young and Geisen 2002). In accordance with the ICBN article 8.3 drawings or images are considered to be the type if a specimen cannot be preserved and hence the figures drawn by Wallich (1877) are valid type illustrations. Later taxonomic and cytological work on HeUcosphaera uses SEM (e.g. Gaarder 1962; Nishida 1979; Okada and McIntyre 1977) and TEM (Inouye, unpublished observations). Further data - a large proportion of it gathered during the CODENET project and not yet published - stems from microanatomical cytological sections (Inouye, unpubl. data; Probert unpubl. data), photosynthetic pigment analysis (van Lenning, unpublished data) and molecular biology (Saez et al. 2003). 3 Results 64

3.2.6 Life cycle

Recent evidence from combination cells of H. carteri with S. catilliferus (Cros et al. 2000) and combination cells of the holococcolithophorids S. catilliferus and S. confusus (Geisen et al. 2002) suggest that both holococcolith morpholgies are produced by the haploid phase of the life-cycle of H. carteri. Geisen et al. (2002) explain the close morphological relationship between the two holococcoliths as due to intraspecific variation in the degree of calcification. Although the available examples are few and based on observations on natural populations only they provide convincing evidence to treat both S. catilliferus and S. confusus as junior (heterotypic) synonyms of H. carteri. There is also tentative evidence from a combination coccosphere for a life cycle association of H. wallichii with Syracolithus dalmaticus. This is not a particularly clear example of a combination cell and could indeed be an accidental association. However, in our field samples of the variant wallichii, fi-om Miyake-jima, a small island off Tokyo bay, not far from where Inouye isolated his H. wallichii clone, the holococcolithophore S. dalmaticus is usually occurring as well - so we tentatively conclude that there is evidence for morphological differentiation of both holococcolith and heterococcolith phases between wallichii and carteri. HeUcosphaera wallichii Lohmann (1902) Okada et McIntyre (1977) has priority over Syracolithus dalmaticus Kamptner (1927) Loeblich et Tappan (Loeblich and Tappan 1963), hence if the combination is proven the correct name for the species will be H. wallichii. There is no evidence available on the haploid phase of H. hyalina, although it is interesting to note that the holococcoliths within the genus HeUcosphaera have a highly distinctive ultrastructure, formed predominantly of aligned rhombohedral crystallites. One further holococcolithophore with this ultrastructure occurs, 5.ponticuliferus, and so is a prime candidate as the H. hyalina holococcolithophore.

3.2.7 Cultures

The two H. carteri strains used for the temperature experiments were isolated from water samples taken on an oceanographic cruise in the South Atlantic, off Cape Town. Six further strains maintained at 17° C - five of H. carteri (S. Atlantic) and one of H. hyalina (Mediterranean, off Naples, Italy), were analysed additionally to compare morphologies. Data for all the cultures used is given in figure 3. The cultures were grown under sterile but not axenic conditions on a temperature-light gradient table. Six temperatures covering a range from 14 to 27° C were selected. The mean irradiance level was between 42 (±3) /

For the final experiment the cultures were grown at six temperatures (13.7, 15.5, 16.7, 19.6, 23.1, 26.7° C) starting off with an initial concentration of 1500 cells per ml (c/ml). Subsequently the cultures were counted every 2-5 days using a solution of 0.5% Glutaraldehyde in alcohol to immobilise cells and a Sedgwick-Rafter counting cell. With the initial low cell concentrations 50 fields of view (FOV) were counted, later on a count of more than 100 cells, from a variable number of fields of view proved to be sufficient. The growth rate k of the cultures in divisions per day was then calculated using the following formula k=\og{N^ / N^)*{3.22l t), w ith and being the cell concentrations at the end and the beginning of a period time t in days (Guillard 1973). In addition to the cell counts the coccosphere size of 20 cells per sample was measured using an inverted microscope (Olympus BH-2) with an eyepiece graticule shortly after inoculation (day 2) and on day 18. As the stationary phase was reached cells were harvested, centrifuged and suspended in ethanol before mounting as permanent light microscope mounts with Lakeside. Morphometrical analyses on the

H. carter! culture NS 8-4 (S. Atlantic) coccolith size te m p . In °C n mean length (pm) St. dev. mean width (pm) St. dev. 1 3 .7 5 0 9 .5 0 .9 6 .6 0 .7 1 5 ,5 5 0 9 .4 1 .0 6 .5 0 .7 16.7 50 9.2 1.1 6.4 0.8 1 9 .6 5 0 9 .0 0 .9 6.1 0 .6 2 3 .1 5 0 9 .0 0 .8 6 .0 0 .7 all te m p e r a tu r e s 2 5 0 9 .2 1 .0 6 .3 0 .7

HeUcosphaera spp. coccolith size of cultures at 17'C cu ltu re n mean length (pm) S t. d ev . mean width (pm) St. dev. H. ca rten NS 8-4 (S. Atlantic) 5 0 9 .2 1.1 6 .4 0 .8 H. carteri NS 10-13 (S. Atlantic) 4 0 7 .5 0 .8 5 .2 0 .4 H. carteri NS 10-14 (S. Atlantic) 4 0 7 .6 0 .8 5 .3 0 .5 H. carteri NS 10-6 (S. Atlantic) 4 0 8 .6 0 .9 5 .8 0 .5 H. carteri NS 8-3 (S. Atlantic) 40 8.0 0.8 5.6 0.5 H. hyalina NAP11 (off Naples, Italy) 41 6 .3 0 .6 4 .5 0 .4 all cultures (excluding H. hyalina ) 2 1 0 8 .2 0 .9 5 .7 0 .6

H. carteri culture NS 10-8 (S. Atlantic) coccosphere size after 2 days sa m p le n mean length (pm) St. d ev . mean width (pm) St. dev. 1 3 .7 2 0 1 7 .9 4 .2 1 7 .0 4 .6 1 5 .5 2 0 1 9 .6 2.1 1 8 .4 2 .6 1 6 .7 2 0 1 7 .4 3 .8 1 6 .5 3 .6 1 9 .6 2 0 1 6 .6 2 .9 1 5 .4 1 .9 2 3 .1 2 0 1 6 .9 3 .0 1 6 .8 3 .0 2 6 .7 2 0 1 8 .1 3 .3 1 7 .2 2.1 all temperatures 100 1 7 .8 3 .2 1 6 .9 3 .0

H. carteri culture NS 10-8 (S. Atlantic) coccosphere size after 18 days sample n mean length (pm) St. d ev . mean width (pm) St. d ev . 1 3 .7 2 0 18.6 5.1 17.5 5.1 1 5 .5 2 0 1 8 .6 4 .2 1 7 .9 3 .5 1 6 .7 20 18.4 4.1 16.4 4.4 1 9 .6 20 18.9 3.4 17.2 3.0 23.1 20 17.3 3.2 1 6 .0 2 .4 all temperatures 100 18.4 4 .0 1 7 .0 3 .7

Fig. 3. Sample location and statistics for all the cultures samples of HeUcosphaera studied. 3 Results 66 slides were then performed under phase contrast using a Zeiss Axioplan microscope with a Hamamatsu video camera attached.

Physiology

Both cultures tested (NSlO-8, NS8-4) grew over a temperature range from 13.7 to 23.1° C (Fig. 4-6). Cultures repeatedly inoculated at 26.7° C and at 12° C did not grow so the cut-off temperatures will be somewhere between 23.1° and 26.7° C and between 12° and 13.7° C respectively. Both cultures had their optimum growth temperature between 16.7 and 19.6° C (Fig. 4-6).

H. carten culture NS 8-4 H carteri culture NS 10-8

13 7"C /> /'

20

Fig. 4. Logarithmic plot of the cell densities in both cultures. The yellow area marks the exponential growth phase.

H. carten ■ culture NS8-4

tem perature in “C tem perature in “G

Fig. 5. Cell yield and growth rate as a function of temperature. Both parameters show a nar­ row temperature tolerance. 3 Results 67

NS10-8 growth rates NS8-4 growth rates

T in °C day c/ml A f N2 IN1 k T in “C day c/ml A t N2 INI k 13.7 0 1500 13.7 0 1500 2 1705 2 1.14 0.09 3 2104 3 1.40 0.16 7 2560 5 1 50 0.12 8 3923 5 1 86 0.18 12 3963 5 1 55 0 13 14 5889 6 1 50 0.10 14 4318 2 1 09 0.06 19 6605 5 1.12 0.03 18 4816 4 1 12 0 04 22 9696 3 1.47 0 18 21 5563 3 1 16 0.07 30 10400 8 1 07 001 23 5507 2 0 99 -0 01 for day 0-23 23 3 6 7 0.08 tor day 0-22 22 6.46 0.12 stationary phase not reached tor day 22-30 8 1.07 0.01

T in °C day c/ml A t N2 1 N1 k T in °C day C/ml A f N2 INI k 15.5 0 1500 15.5 0 1500 2 1688 2 1.13 0 08 3 1714 3 1.14 0.06 7 3184 5 1.89 0.18 8 4259 5 2 48 0.26 12 7176 5 2.25 0 23 14 10923 6 2 56 0.23 14 7824 2 1 09 0 06 19 14636 5 1 34 0.08 18 8526 4 1.09 0.03 22 19333 3 1.32 0.13 21 11600 3 1.36 0 15 30 19455 8 1 01 0.00 23 14500 2 1.25 0 16 tor day 0-23 23 9.67 0.14 tor day 0-22 22 12.89 0.17 stationary phase not reached tor day 22-30 8 1.01 0.00

T in “C day C/ml A f N2 ! N1 k T in “C day c/ml A t N2 INI k 16.7 0 1500 16.7 0 1500 2 1286 2 0 8 6 -0 11 3 2680 3 1 79 0 2 8 7 3842 5 2 99 0.32 8 9538 5 3 56 0 3 7 12 5875 5 1 53 0 12 14 26818 6 2.81 0 25 14 10647 2 1 81 0.43 19 28071 5 1 05 0.01 18 19000 4 1 78 0 21 22 27600 3 0 98 -0 01 21 33667 3 1 77 0 28 23 24235 2 0 72 -0 24 tor day 0-21 21 22.44 0.21 tor day 0-14 14 17.88 0.30 tor day 21-23 2 0.72 -0.24 tor day 14-22 8 1.03 0.01

T in “C day c/ml A t N2 / N1 k 19.6 0 1500 T in °C day C/ml A f N2 INI k 2 1694 2 1.13 0.09 19.6 0 1500 7 5520 5 3.26 0.34 3 4140 3 2 7 6 0 4 9 12 18500 5 3.35 0 35 8 138335 3.34 0.35 14 26300 2 1.42 0.25 14 28077 6 2.03 0 17 18 24053 4 0.91 -0.03 19 27875 5 0 99 0 00 21 28300 3 1.18 0.08 22 29000 3 1.04 0 02 23 32750 2 1.16 0.11 tor day 0-14 14 17.53 0.30 tor day 0-14 14 18.72 0.30 tor day 14-23 9 1.25 0.04 tor day 14-22 8 1.03 0.01

T in °C day c/ml A t N2 1 N1 k T in °C day c/ml A t N2 INI k 23.1 0 1500 23.1 0 1500 2 2167 2 1.44 0.27 3 3260 3 2 17 0 3 7 7 7824 5 3.61 0.37 8 9167 5 2.81 0.30 12 35667 5 4.56 0.44 14 24375 6 2 66 0.24 14 22143 2 0.62 -0 34 19 26909 5 1 10 0.03 18 27000 4 1 22 0.07 22 25769 3 0.96 -0.02 21 31727 3 1 18 0.08 23 24500 2 0 77 -0 19 tor day 0-14 14 14.76 0.28 tor day 0-14 14 16.25 0.29 tor day 14-23 9 1.11 0.02 tor day 14-22 8 1.06 0.01

Fig. 6. H. carteri growth rates for culture NS 1Ü-8 and NS 8-4. For the calculation of growth rates see text. 3 R esults 68

For one of the slow growing cultures at low temperatures (NSlO-8 at 13.7° C) the stationary phase was not reached even after 23 days of growth, however the faster growing cultures reached stationary phase after roughly 14 days. The growth rate in the exponential phase in culture NS8-4 ranged from 0.30 divisions per day for the 19.6° C culture down to 0.12 at 13.7° C (Fig. 5,6). Final cell density in the stationary phase ranged from ca. 10000 cells per ml to ca. 29000 cells per ml (Fig. 5,6). Counts in the stationary phase varied slightly as the cultures had the tendency to form clumps of cells with higher densities in the later stages of the experiment. In analogy with the observations on S. pulchra (see 3.1.4), H. carteri exhibited a strong photo- and thermotaxis.

Morphology

Coccosphere morphology (Fig. 3) For culture NSlO-8 coccosphere length after 2 days growth ranges from 16.6 - 19.6 pm with mean of 17.8 pm over the whole temperature range. Coccosphere width ranges from 15.4 - 18.4 p m with a mean of 16.9 pm . For culture NSlO-8 coccosphere length after 18 days growth ranges from 17.3 - 18.9 pm with mean of 18.4 pm over the whole temperature range. Coccosphere width ranges from 16.0 - 17.9 p m with a mean of 17 pm . No statistical correlation between coccosphere size and growth rate could be detected, but there appears to be a slight increase in coccosphere length in the samples taken after 18 days.

Coccolith morphology (Figs 3,7,8) For culture NS8-4 mean coccolith length after harvest on day 22 and day 30 respectively varies from 9 - 9.5 pm with a mean of 9.2 pm over the whole temperature range. Mean coccolith width ranges from 6 - 6.6 pm with an overall mean of 6.3 pm . The distribution of both length and width appears to be unimodal and shows no correlation with temperature/growth rate (Fig. 7). In addition the coccolith size of four additional H. carteri clones and one H. hyalina clone grown at 17° C only were measured. The H. carteri coccoliths revealed a mean coccolith length of 7.5- 9.2 p m (with an overall mean of 8.2 pm ) and a mean coccolith width varying from 5.2 - 6.4 p m (overall mean 5.7 pm ). The distribution of length and width is unimodal, but compared to the variation in the temperature experiment there appears to be a higher variability of morphospace between the clones measured (Fig. 8). The culture of H. hyalina however displayed a mean coccolith length of 6.3 pm with a mean width of 4.5 pm. H. hyalina coccoliths are therefore significantly smaller than H. carteri coccoliths (Fig. 8). 3 R esults 69

HeUcosphaera ca/ter/culture NS8-4

13.7°C 19.6°C

15.5'C 23.r c

16.7X all tem peratures

f/ / 4 /

5 6 7 8 9 10 length in gm

Fig. 7. Density plot for coccolith length and width for H. carteri culture NS 8-4 at various temperatures. 3 R esults 70

HeUcosphaera cultures at 17° C all cultures

H. carten NS 10-13 H. carteri NS 10-14 H. carten NS 10-6

H. carten NS 8-4 H. carten NS 8-3 H. hyalina NAP 11

Fig. 8. Density plot for coccolith length and width for other selected cultures of H. carteri and H. hyalina at 17° C. 3.2.8 Sediment trap samples

To investigate the temporal variation of HeUcosphaera coccolith size sub-samples were taken from a sediment trap mooring off Somalia (MST-9, 10°45 N 53°34 E). Water depth was 4050 m with the specific trap used collecting at a depth of 3050 m. For details of specific trap location and coccolithophore fluxes in a higher trap in the array (MST9-E) refer to Broerse et al. (2000). Due to the low abundance of HeUcosphaera coccoliths in the samples only three samples were analysed. All the measurements have been performed with a light microscope and hence the different HeUcosphaera species could not be differentiated (Fig. 9). 3 R esults 71

HeUcosphaera spp coccolith size in trap MST-9G off Somalia sample starting ctoy dwationliday*) sBquotused SSTfrt*C « 4' mean teigth (pm) dev. mean «nefth ftan) st. dev. G2 07/06/1992 M 1/320 >27 26 9.5 1.5 6.4 1.2 G3 21/06/1992 M 1/640 >22 <27 25 9.0 1.3 5.9 0.9 G4 05/07/1992 14 1/1280 >22 <25 29 ______8V ______U ______L 6 ______0.8 all sam ples 42 80 9.1 1.3 6.0 0.9

Fig. 9. Statistic for all measurements on HeUcosphaera spp. coccoliths from sediment trap material. The sea surface data (SST) is taken from Broerse (2000). M orphology

Due to the low abundance of HeUcosphaera coccoliths in the sediment trap samples the dataset is not suitable for a detailed interpretation and is only used as a comparison with other samples. Mean coccolith length in all samples is 9.1 pm with a mean width of 6 pm. This compares well to the data from the cultured HeUcosphaera clones (Fig. 10).

HeUcosphaera spp. (Sediment trap time series)

Contour lines 0.5 - max @ 0.5 intervals. Contours <1.5 are dashed G2 G3

1=26

G4 all sam ples 7.50

"

i.25 I £

5.00 l l l > 'U II 1 _ / 3.75 / \ - 1=29 n=80 2.50

length in pm

Fig. 10. Density plot for coccolith length and width for all sediment trap samples of HeUcosphaera studied. 3 Results 72

3.2.9 Holocene samples

Eight Holocene coretop samples taken from DSDP/ODP sites covering the North and South Atlantic Ocean, the Indian Ocean and the eastern Pacific Ocean and a boxcore (BC3) from the Meditenanean were analysed (Fig. 11,12). OOP site 704 (S. Atlantic Ocean) was checked, but the abundance of HeUcosphaera coccoliths proved to be too low to allow for a statistical analysis. All the measurements have been performed with a light microscope and hence the different HeUcosphaera species could not be differentiated. Note that the sample from OOP site 872 is of early Pleistocene age and has been included as a comparison with the Holocene sam p les.

BC3

# OOP cores

Fig. 11. Map of the DSD? / OOP sites studied to determine the morphological variation in Holocene HeUcosphaera. Sample BC 3 is from a boxcore in the Mediterranean and was provided by Utrecht University.

Helicosphaera spp. coccolith size in Holocene samples site mean length 0/m) St. dev. mean width (pm) St. dev. ODP 664 108, 664D. IN I. 0-1 cm 0 0 .1 0 N 023.21 W 30 8.3 1.2 5.3 0.9 OOP 806 130. 806B. INI. 2-4cm 00.31 SI 59,35 E 30 9.0 1.5 5.7 1.0 138, 849B. IHI. 2-4 00.18 N 110.53 W 30 9.3 1.6 ODP 872 144, 872A, 1H2, 127cm 10.09 N 162.85 E 25 9.0 1.2 0.8 ODP 704 114, 704A. IHl. 1-3 46.88 S 007.42 E 9 9.5 1.2 0.6 BC3 BC3. EE0019, 6,8-10.4cm 33.37 N 024.76 E 30 9.0 0.9 5.7 0.6 ODP 659 108, 659A. INI. 2-4cm 18.08 N 021.02 W 29 9.0 1.0 5.7 0.6 ODP 709 115. 709C, INI, 3-5cm 0 3 .9 0 S 06 0 .5 5 E 30 8.7 1.7 1.2 ODP 552 81. 552. IH l.g-llcm 56.04 M 023,23 W 30 9.2 1.1 0,5 all samples 243 9.0 1.3 5.7 0.8

Fig. 12. DSDP / OOP site location and statistics for all Holocene samples of HeUcosphaera studied. 3 Results 73

Morphology

Mean coccolith length in the samples varies from 8.4 - 9.5 ptm with mean of 9 /

Helicosphaera spp. : Holocene Conlour ires 0 5 • max @ 05 rlervals Contours < t 5 are dashed

all samples incl ODP 704

A '

:'â "

A .IV \ , Æ

"Oi

Fig. 13. Density plot for coccolith length and width for all Holocene samples of Helicosphaera studied. 3 Results 74

3.2.10 Downcore samples (ODP site 664)

A series of 16 samples (Fig. 14) taken from core ODP site 664 was analysed covering a time interval from 0 to 3.992 Ma. The age model used is based on few magnetochronological datums (Brunhes/Matuyama and Jaramillo-top and bottom) plus nannobiochronological data mostly from LAD of D is c o a s te r forms (Su 1996). As with the Holocene samples and the sediment trap samples the analyses was performed with a light microscope and hence the different Helicosphaera sp ec ies could not be discriminated.

Helicosphaera spp. coccolith size of core ODP 664 samples sample core location sub-bottom depth (m) age in Ma n mean length (pm) S t, dev. mean width (pm) S t. dev. OOP664D0.01 0.1 N 23.21 W 0.01 0.000 30 8.3 1.2 5.3 0.9 ODP664D1.52 1.52 0 .0 4 5 30 9.0 1.1 5.5 0.6 ODP664D3.32 3.32 0.098 30 9.2 1.5 5.8 0.8 ODP664D5.52 5.52 0 .1 6 4 30 8.8 1.1 5.7 0.7 ODP664D8.52 8.52 0.253 30 9.3 1.3 5.7 0.9 ODP664D12.02 12.02 0.357 30 7.8 1.4 5.0 0 .8 ODP664D26.92 26.92 0 .7 5 5 30 8.8 1.2 5.6 0.8 0DP664035.82 35.82 0.991 30 9.3 1.1 5.9 0.7 ODP664D45.92 45.92 1.242 30 9.7 1.5 6.2 0.9 OOP664D60.44 60.44 1.566 30 8.6 1.2 5.4 0.6 0DP664D104.42 104.42 2.436 30 9.1 1.4 5.9 0.8 ODP664D109.34 109.34 2.529 30 8.2 1.3 5.2 0.8 ODP6640125.84 125.84 2.852 30 8.8 0.8 5.8 0.6 ODP664D139.02 139.02 3.207 30 8.5 1.0 5.4 0.5 ODP664D149.23 149.23 3.482 30 8.9 1.1 5.9 0.8 ODP6640166.82 166.82 3.992 30 8.5 1.2 5.4 0.6 all sam ples 420 8.8 1.2 5.6 0.7

Fig. 14. ODP core 664 location and statistics for all samples of Helicosphaera studied. Morphology

Mean coccolith length in the samples varies from 7.8 - 9.7 /

OOP site

Fig. 15. Density plot for coccolith length and width for all samples of Helicosphaera studied from OOP site 664. 3 Results 76

3.2.11 First occurrence

According to Young (1998) the first occunence of H. carteri (as H. carteri var. carteri) in the fossil record is of Late Oligocene age. This age is supported by our own interpretation of data from the ETH Neptune database (Fig. 16-A), pointing to the Early Miocene. The synonym H. kamptneri was also available from the database and was plotted separately (next page, fig. 16-B), pointing to a Late Oligocene origin. H. wallichii (as H. carteri var. wallichii) is of Late Miocene (Tortonian) age (Young 1998), but no data are available from the Neptune database for this species. The timescale used for the Neptune database and for the nannofossil zones in figure 16 is that of Berggren et al. (1995). Molecular data dates the divergence of H. hyalina versus H. carteri b etw een 8.22 and 12.16 Ma (Saez et al. 2003). This would result in a first occurrence of the species in the Middle or Late Miocene. However, data from the Neptune database points to a first occurrence in the middle Pliocene (next page, fig. 16-C). This apparent discrepancy might be due to the species concepts applied by the various OOP nannofossil workers and the difficulty to differentiate the species in the light microscope. Alternative hypotheses would argue for morphological evolution being strongly uncoupled from molecular evolution or for the molecular clock of Saez et al. (2003) being erroneous (see 3.3.8 for a discussion).

Cumulative occurences ofH. carteri in 0.5 Ma time intervals in GDP / DSDP sites (data from ETH Neptune database with 2288 observations from 46 sites)

1 6 0 -

1 4 0 ------

» 120 « I '00 « 3 80 I»

40

20

Stages Piac 1 Zanc Mess 1 Tortonian Serravalian |Lang Burdigalian | Aquitanian Chattian Rupelian Epochs Pieist Pliocene Late Miocene Middle Miocene Early Miocene Late Oligocene E. Oiig (pars)

age in Ma C) 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32

Fig. 16-A. Frequency plot of occurrences of Helicosphaera carteri in DSDP / ODP sites. The data is taken from the ETH Neptune database. 3 Results 77

Cumulative occurences ofH kam ptneh In 0 5 Ma time Intervals in ODP / DSDP sites (data from ETH Neptune database with 811 observations from 28 sites)

I 30

Piac I Z anc~ M ess | Tortonian Serravalian |^ a n ^ Burdigalian | Aquitanian Chattian Rupelian Epochs Pleist Pliocene Middle Miocene Early Miocene Late Oligocene E Olig (pars) age in Ma 0 20 22 26 28 32 34

Cumulative occurences ofH. hyalina in 0.5 Ma time intervals in ODP / DSDP sites (data from ETH Neptune database with 152 observations from 22 sites)

I 35

Z 25 I § 20

15

to

Piac I Zanc" Mass | Tortonian Serravalian [lang Burdigalian | Aquitanian Rupelian Epochs Pleist Pliocene Late Miocene Middle Miocene Early liAocene Late Oligocene E. Olig (pars) age in Ma 0 20 22 26 28 32 34

Fig. 16-B, C. Frequency plot of occurrences of B) H. kamptneri, a synonym of H. carteri and C) H. hyalina in DSDP / ODP sites. The data is taken from the ETH Neptune database. 3 R esults 78

3.2.12 Conclusions

The key observations on Helicosphaera are:

1. Measurements of coccoliths from cultured H. carteri reveal a unimodal, well defined morphospace which is contrasted by a bi- or multimodal distribution in Holocene and time series samples. 2. There is evidence from life cycle observations and from fine scale morphological work that the Helicosphaera variants carteri and wallichii are two separate biological species. 3. There is evidence from molecular data that the Helicosphaera variants carteri and hyalina are separate biological species. 4. In the monoclonal cultures used there was no temperature control on either cell or coccolith size. 5. Optimum growth is at a well defined temperature. 6. As in S. pulchra both photo- and thermotaxis have been observed.

Comparable to the case of 5. pulchra morphometric analyses again produced ambiguous, difficult to interpret patterns. Our key result came from culture isolation of a disputed morphovariant - H. hyalina. Contrary to expectations the coccolith morphology has proven stable in culture indicating again that a subtle morphological variant is under genotypic control. Based on our new results from morphological and molecular studies Saez et al. (2003) have concluded that the two varieties are in fact separate, albeit closely related, species and that their most recent common ancestor would have lived between 8.22 and 12.16 Ma. Although molecular data is still pending evidence from the morphology, life-cycle data and previous culture observations strongly indicate that H. carteri var. wallichii should also be considered a discrete species rather than a variety. These findings have lead to the suggestion that in the Helicosphaera taxa discussed here sympatric evolution of a phenotypically rather plastic population has occurred. This was in turn followed by a change in morphological variation within the single taxa, which exhibit small, but stable morphological differences.

3.2.13 References

Aubry M-P (1990) Handbook of Cenozoic calcareous nannoplankton. Book 4: Heliolithae (Helicoliths, Cribriliths, Lopadoliths and others). Micropaleontology Press, American Museum of Natural History, New York Berggren W A, Kent D V, Swisher C C, Aubry M-P ( 1995) A revised Cenozoic Geochronology and Chronostratigraphy. Geochronology Time Scales and Global Stratigraphie Correlation, SEPM Special Publication 54: 129-212 Borsetti A M, Cati F (1972) II Nannoplancton calcareo vivente nel Tirreno Centro- meridionale. G Geol 38: 395-452 3 R esults 79

Broerse A T C, Brummer G-J A, Hinte J E v (2000) Coccolithophore export production in response to monsoonal upwelling off Somalia (northwestern Indian Ocean). Deep-Sea Res II 47: 2179-2205 Carter H J (1871) On Melobesia unicellularis, better known as the coccolith. Ann Mag nat Hist 7: 184-189 Gros L, Kleijne A, Zeltner A, Billard C, Young J R (2000) New examples of holococcolith- heterococcolith combination coccospheres and their implications for coccolithophorid biology. Mar Micropaleontol 39: 1-34 Deflandre G (1952) Classe des Coccolithophoridés. (Coccolithophoridae. Lohmann, 1902). In: Grasse P P (ed) Traite de Zoologie. Masson, Paris, pp 439-470 Gaarder K R (1962) Electron Microscope Studies on Holococcolithophorids. Nytt Mag Bot 10: 35-51 Gaarder K R (1970) Three new taxa of Coccolithineae. Nytt Mag Bot 17: 113-126 Gaarder K R, Heimdal B R (1977) A revision of the genus Syracosphaera Lohmann (Coccolithineae). Meteor ForschErgebn D 24: 54-71 Geisen M, Billard C, Broerse A T C, Gros L, Probert I, Young J R (2002) Life-cycle associations involving pairs of holococcolithophorid species: intraspecific variation or cryptic spéciation? Eur J Phycol 37: 531-550 Guillard R R L (1973) Division rates. In: Stein J R (ed) Handbook of phycological methods Culture methods & growth measurements. Cambridge University Press, Cambridge, pp 289-311 Hay W W, Mohler H P, Roth P H, Schmidt R R, Boudreaux J E (1967) Calcareous nannoplankton zonation of the Cenozoic of the Gulf Coast and Caribbean-Antillean area, and transoceanic correlation. Trans Gulf Cst Ass geol Socs 17: 428-480 Heimdal B R (1993) Modem Coccolithophorids. A Guide to Naked Flagellates and Coccolithophorids. In: Tomas C R (ed) Marine Phytoplankton. Academic Press, London, pp 147-243 Henriksen K, Young J R, Bown P R, Stipp S L S (2004) Coccolith biomineralisation studied with atomic force microscopy. Palaeontology 57: XXX XXX Huxley T H (1868) On some organisms living at great depths in the North Atlantic Ocean. Q J1 m icrosc Sci 8: 203-212 Jafar S A, Martini E (1975) On the validity of the calcareous nannoplankton genus Helicosphaera. Senckenberg leth 56: 381-397 Jordan R W, Green J C (1994) A check list of the extant haptophyta of the world. J mar biol Ass U K 74: 149-174 Jordan R W, Kleijne A (1994) A classification system for living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 83-105 Jordan R W, Young J R (1990) Proposed changes to the classification system of living Coccolithophorids. Int Nannoplankton Assoc Newsl 1: 15-18 Kamptner E (1927) Beitrag zur Kenntnis adriatischer Coccolithophoriden. Arch Protistenkd 58: 173-184 Kamptner E (1937) Neue und bemerkenswerte Coccolithineen aus dem Mittelmeer. Arch Protistenkd 89: 279-316 3 R esults 80

Kamptner E (1941) Die Coccolithineen der Südwestküste von Istrien. Annin naturh Mus Wien 51: 54-149 Kamptner E (1954) Untersuchungen iiber den Feinbau der Coccolithen. Anz ost Akad Wiss Mathematisch-Naturwissenschaftliche Klasse 87: 152-158 Kleijne A (1991) Holococcolithophorids from the Indian Ocean, Red Sea, Mediterranean Sea and North Atlantic Ocean. Mar Micropaleontol 17: 1-76 Kleijne A (1993) Morphology, taxonomy and distribution of extant coccolithophorids (Calcareous nannoplankton). PhD thesis. Free University Amsterdam (ISBN 90- 9006161-4), p. 321 Loeblich A R, Tappan H (1963) Type fixation and validation of certain calcareous nannoplankton genera. Proc Biol Soc Wash 76: 191-198 Loeblich A R, Tappan H (1966) Annotated index and bibliography of the calcareous nannoplankton. Phycologia 5: 81-216 Lohmann H (1902) Die Coccolithophoridae, eine Monographie der Coccolithen bildenden Flagellaten, zugleich ein Beitrag zur Kenntnis des Mittelmeerauftriebs. Arch Protistenkd 1: 89-165 Lohmann H (1919) Die Bevolkerung des Ozeans mit Plankton nach den Ergebnissen des Zentrifugenfange wahrend der Ausreise der “Deutschland” 1911. Zugleich ein Beitrag zur Biologie des Atlantischen Ozeans. Arch Biontol 4: 1-617 Nishida S (1979) Atlas of Pacific Nannoplanktons. News Osaka Micropaleontol Special Paper: 1-31 Norris R E (1985) Indian Ocean nanoplankton. II. Holococcolithophorids (Calyptrosphaeraceae, Prymnesiophyceae) with a review of extant genera. J Phycol 21: 619-641 Okada H, McIntyre A (1977) Modem coccolithophores of the Pacific emd North Atlantic Oceans. Micropaleontology 23: 1-55 Ostenfeld C H (1899) Über Coccosphaera und einige neue Tintinniden im Plankton des nordlichen Atlantischen Oceans. Zoo! Anz 22: 433-439 Perch-Nielsen K (1985) Cenozoic calcareous nannofossils. In: Bolli H M, Saunders J B, Perch-Nielsen K (eds) Plankton Stratigraphy. Cambridge University Press, Cambridge, pp 427-555 Rice A L, Burstyn H L, Jones A G E (1976) G. C. Wallich M.D.-megalomaniac or mis-used oceanographic genius? J Soc Biblphy nat Hist 7: 423-450 Saez A G, Probert I, Geisen M, Quinn P, Young J R, Medlin L K (2003) Pseudo-cryptic spéciation in coccolithophores. Proc natn Acad Sci USA 100: 7163-7168 Schiller J (1930) Coccolithineae. In: Rabenhorst L (ed) Kryptogamen-Flora von Deutschland, Osterreich und der Schweiz. Akademische Verlagsgesell schaft, Leipzig, pp 89-267 Su X (1996) Development of late Tertiary and Quaternary coccolith assemblages in the Northeast Atlantic. Geomar Report 48: 1-120 Theodoridis S (1984) Calcareous nannofossil biostratigraphy of the Miocene and revision of the helicoliths and discoasters. Utrecht micropaleont Bull 32: 1-271 Wallich G C (1877) Observations on the coccosphere. Ann Mag nat Hist 19: 342-350 3 R esults 81

Winter A, Jordan R W, Roth P H (1994) Biogeography of living coccolithophores in ocean waters. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 161-177 Winter A, Siesser W G (1994) Atlas of living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 107-159 Young J R (1998) Neogene. In: Bown P R (ed) Calcareous Nannofossil Biostratigraphy. Chapman & Hall, London, pp 225-265 Young J R, Geisen M (2002) Xenospheres - associations of coccoliths resembling coccospheres. J nannoplankton Res 24: 27-35 3 R esults 82

3.3 Umbilicosphaera spp. - A case of spéciation?

3.3.1 Introduction

The coccolithophore genus Umbilicosphaera comprises four extant species or va­ rieties (Winter and Siesser 1994) and two extinct fossil species (Young 1998). The most common extant taxa U. sibogae var. sibogae and U. sibogae vzx.foliosa show a broad inter-oceanic occurrence and are common in the fossil record but there are few reliable data on their first occurrence. The type species of U. sibogae var sibogae was described by Weber - van Bosse (1901) whilst on an oceanographic cruise around the Malayan peninsula and named Coccosphaera sibogae after the re­ search vessel RV Siboga. Gaarder (1970) however points out that Coccosphaera is a homonym and hence an invalid genus. As a consequence the genus Umbilicosphaera stems from the junior synonym Umbilicosphaera mirabilis described by Lohmann (1902) and C. sibogae is the type species of the genus Umbilicosphaera. The second Umbilicosphaera taxon studied here in detail is U. sibogae \ar.foliosa, described as Cycloplacolithus foliosus by Kamptner (1963) from Holocene deep-sea sediments from the Pacific. The close morphological affinity between the two taxa was noted by several authors (Inouye and Pienaar 1984; McIntyre and Be 1967; Nishida 1979) and the two species have been considered variants by Okada and McIntyre (1977). Indeed U. sibogae was selected as a CODENET species in part in order to use this varia­ tion as a case study on the morphological variation of two extremely closely related taxa and their morphological evolution in the geological record. Our own research with data from a number of sources has however demonstrated that the varieties are well separated and should be regarded as discrete species, U. sibogae and U. foliosa respectively. Consequently the varieties were risen to species rank in Saez et al. (2003) and the new species names are used in this manuscript. The varieties are only referred to where it is necessary in historical terms (see remarks on the genus Umbilicosphaera). Only a few quantitative studies on the morphology of Umbilicosphaera spp. were performed to date (e.g. Baumann and Sprengel 2000) and we present here our new results from biometrical studies on cultures, sediment samples and Holocene coretop samples.

3.3.2 Taxonomy and description ofUmbilicosphaera sibogae

Umbilicosphaera sibogae (Weber-van Bosse) Gaarder (Plate 1, figs 3, 4, 6, 11-13) Coccosphaera sibogae sp. nov. Weber - van Bosse (1901), p. 140, pi. 17, figs 1, 2. Umbilicosphaera mirabilis gen. nov. sp. nov. Lohmann (1902), p. 139-140, pi. 5, figs 66, 66a; Schiller (1930), p. 249, fig. 126. 3 R esults 83

Coccolithus sibogae Schiller (1930), p. 248, fig. 125. Umbilicosphaera sibogae comb. nov. Gaarder (1970), p. 126, figs 9 (c, d), non 8 e; Borsetti et Cati (1976), p. 223-224, pi. 18, figs 2, 3; Okada et McIntyre (1977), p. 13, pi. 4, fig. 2; Nishida (1979), pi. 5, fig. 2; Heimdal (1993), p. 231-232, pi. 7. Umbilicosphaera sibogae var. sibogae Okada et McIntyre (1977), pi. 4, fig. 2; Kleijne (1993), pp. 197-198, pi. 4, figs 1, 2; Winter et Siesser (1994), fig. 22; Young (1998), pi. 8.5, figs 1,2.

Description of the heterococcoiithophore U. sibogae

U. sibogae forms large spherical to subspherical coccospheres consisting of 40 to more than 100 partly interlocking placoliths (Fig. 1 - A; Plate 1, figs 3,4,6). Placoliths are circular with a large central opening (Plate 1, fig. 11); the proximal shield is flat and typically larger than the convex distal shield (Fig 1 - B, D, Plate 1, fig. 11,12). Both shields monocyclic, with the proximal shield composed of R-units and the distal shield of V-units (Fig. 1 - C; Plate 1, fig. 11,12). Elements on the distal shield are imbricated dextrally, sutures are straight on inner part of the rim, then kinked and incised laevogyre on outer part of the rim. An organic membrane spans the central opening (Plate 1, fig. 12,13). Cells semi-colonial with each cell typically containing 1-2, occasionally 4 protoplasts (Plate 1, fig. 6). The protoplasts do not fill the entire coccosphere (Plate 1, fig. 6). Cells are non motile, but flagellar bases are present (Probert, unpubl.). Uncalcified body scales are absent (Probert, unpubl.).

3.3,3 Taxonomy and description of Umbilicosphaera foliosa

Umbilicosphaera foliosa (Kamptner) Geisen (Plate 1, figs 1, 2, 5, 7-10) Cycloplacolithus foliosus nov. gen. nov. sp. Kamptner (1963), pp. 167-168, pi. 7, fig. 38. Umbilicosphaera sibogae foliosa Nishida (1979), pi 5, figs 1 (a, b). Umbilicosphaera sibogae var. foliosa (Kamptner) comb, nov Okada et McIntyre (1977), pp. 13-14, pi. 4 fig 1; Inouye et Pienaar (1984), pp. 358-361, 366-367, figs 1-14; Kleijne (1993), pp. 198-199, pi. 4, figs 3, 4, pi. 5, fig. 4; Heimdal (1993), p. 232; Winter et Siesser (1994), fig. 21 ; Young (1998), pi. 8.5, figs 3, 4.

Description of the heterococcoiithophore U. foliosa

The compact, spherical coccosphere, consists of up to 25 interlocking placoliths (Fig. 1 - E, Plate 1; figs 1,2,5). The placoliths are circular with a narrow central opening (Fig. 1 - F; Plate 1, figs 7) occasionally with a hook-like spine protruding into the central opening (Fig. 1 - F; Plate 1, fig 9). Both shields are bicyclic and convex, the proximal shield is smaller than the distal shield (Fig 1 - F; Plate 1, figs 7-10). The proximal shield is composed of R-units and the distal shield of V-units (Fig. 1 - G). Inner half of elements on distal shield imbricated dextrally and with straight sutures, outer half kinked, with sinistral imbrication and incised sutures 3 Results 84

(Fig. 1 - F; Plate 1, figs 7,10). An organic membrane spans the central opening (Plate 1, fig. 8). Cells not colonial, the protoplast fills the entire coccosphere (Plate 1, fig. 5). The cells are non motile, but flagellar bases are present (Inouye and Pienaar 1984). Uncalcified body scales are absent (Inouye and Pienaar 1984).

Note: There are two more extant species belonging to the genus U m bilicosphaera, U. anulus and U. hulburtiana. The latter is briefly introduced in the paragraphs first occurrence (see 3.3.8) and conclusions (see 3.3.9). Although U. hulburtiana is rare in natural samples and not the prime focus of this work its close morphological relationship to U. fo lio sa is important for the interpretation of the evolutionary history of the genus. Hence we will present an abridged description here.

Umbilicosphaera hulburtiana Gaarder Gaarder (1970), pp. 121-126, figs 7-9. Coccoliths elliptical and may have nodes around the crest of the distal shield, otherwise similar to U. foliosa. In contrast to U. fo lio sa however Gaarder (1970) has observed body scales.

Remarks on the genus Umbilicosphaera

The type species of Umbilicosphaera, U. sibogae was described by Weber - van Bosse (1901) as Coccosphaera sibogae whilst on an oceanographic cruise around the Malayan peninsula. Although the publication deals mainly with shallow marine macroalgae observations on coccolithophores - conducted with a light microscope - were made from samples obtained with a towed Hensen drag net in pelagic wa­ ter. She states that the cell figured as the type is dividing, but typical U. sibogae cells can be semi-colonial and can easily be mistaken for dividing cells. She also describes the large coccosphere containing only a small protoplast, which is an­ other typical feature of U. sibogae. Equally the presence of a green or yellow-green chromatophore (chloroplast) is noted by her and thus she notes that the algal nature of coccolithophores as proposed by Murray and Blackman (1898) is irrefutably established. As is the case with many early descriptions of coccolithophores the type material could not be preserved and hence under ICBN article 8.3 the drawings are considered the type. Since the type illustration it took almost 100 years until U. sibogae was cultured and made available for cytology, physiological experiments and sequencing. In contrast to the early description of the U. sibogae, U. foliosa was described more than 60 years later by Kamptner (1963) as Cycloplacolithus foliosus using a SEM and sediment samples from the Pacific. The close morphological affinity between the two species was noted by a number of authors (Inouye and Pienaar 1984; McIntyre and Be 1967; Nishida 1979) and although McIntyre and Be (1967) state that the two species can easily be distin­ guished on morphological grounds there has been some confusion in the taxonomy

Fig. 1 (next page). Morphology of U. sibogae and U. fo lio sa as seen in the SEM and LM. 3 Results 85

Umbilicosphaera sibogae

A) scanning electron micrograph of a complete coccosphere ofU. sibogae. Coccosphere ellipsoidal and consisting of 40 to more than 100 partly Interlocking placoliths. The cells are semicolonial and each cell contains typically 1-2, occasionally up to 4 chloroplasts. The protoplasts do not fill the entire coccosphere.

B) scanning electron micrographs ofU. sibogae coccoliths. Placoliths circular, central opening large. Proximal shield (left) larger than distal shield (right). Shields monocyclic. C) light micrographs of aU. sibogae coccolith. Proximal shield composed of R-unlts (# polars - left), distal shield composed of V-unlts (phase contrast - right). D) schematic plan view (top) and cross section (bottom) ofU. asibogae coccolith. ^ { t . Umbiiicosphaera foiiosa

E) scanning electron micrograph of a complete coccosphere ofU. foliosa. Coccosphere spherical and consisting of up to 25 Interlocking placoliths. The cells are not colonial and the protoplast fills the entire coccosphere.

F) scanning electron micrographs ofU. foliosa placoliths Placoliths circular, central opening narror, sometimes with a hook like protrusion. Proximal shield (left) smaller than distal shield (right). Both shields bicyclic. G) light micrographs of aU. sibogae coccolith. Proximal shield composed of R-unlts (# polars - left) with typically curved pseudo-extinction cross. Distal shield composed of V-unlts (phase contrast - right). H) schematic plan view (top) and cross section (bottom) ofU. afoliosa coccolith. 3 Results 86

Plate 1 (next page) Umbilicosphaera spp. Fig. 1: Scanning electron micrograph of two U. fo lio sa coccospheres. U. fo lio sa cells are typically found in clusters of up to four cells. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 69. Fig. 2: Scanning electron micrograph of two U. fo lio sa coccospheres. Note the presence of hook like protrusions in the central opening. Although this appears to be a stable taxonomic character at first cells with both types of coccoliths have been observed frequently. Water sample, N. Atlantic, R/V M eteor cruise 38-1, station 11. Fig. 3,4: Scanning electron micrograph of a U. sibogae coccosphere. Note the organic mem­ brane in Fig. 4 spanning the central opening of some coccoliths. Water sample, S. Atlantic, off Namibia, R/V M eteor cruise M48-4, station 470 (Fig. 3) and station 20 (Fig. 4). Fig. 5: Light micrograph in DlC of a cluster of four U. fo lio sa cells. Culture sample (ESP 6M1), western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 6: Light micrograph in DlC of a dividing U. sibogae cell. U. sibogae cells can be semi­ colonial with typically two cells in a single coccosphere. Note the large extracellular space. Culture material (ASM 39), western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 7: Scanning electron micrograph of the bicyclic distal shield of an U. fo lio sa coccolith. Sediment trap sam ple, Indian Ocean, off Somalia. Fig. 8: Scanning electron micrograph of the bicyclic proximal shield of U. fo lio sa coccoliths. Note the organic membrane spanning the central opening. Culture sample (ESP 6M1), west­ ern Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 9: Scanning electron micrograph of the bicyclic distal shield of U. fo lio sa coccoliths. Note the straight suture lines of the inner cycle in contrast with the ragged suture lines of the outer cycle. Water sample, N. Atlantic, R/V M eteor cruise 38-1, station 13. Fig. 10: Scanning electron micrograph of a U. fo lio sa coccolith in lateral view. The distal shield is larger than the proximal shield. Water sample, western Pacific, Miyake- jima island, Ibo Port, Japan. Fig. 11: Scanning electron micrograph of the monocyclic distal shield of an U. sibogae coccolith. Proximal shield larger than distal shield. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 59. Fig. 12: Scanning electron micrograph of the monocyclic proximal shield of U. sibogae coccoliths. Note the organic membrane spanning the central opening. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 59. Fig. 13: Scanning electron micrograph of U. sibogae coccoliths from a single cocco­ sphere. Note the size variation of the central opening and the rim. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 15.

Scale bars represent: Figs 1-4: 5pm ,figs 5, 6: not to scale, figs 7, 8, 11-13: 2pim, figs 9, 10: l;*m. 3 Results 87

3. U. sibogae coccosphere

2. U. foliosa coccospheres

1. U. foliosa coccospheres

4. U. sibogae coccosphere

5. U. foliosa coccospheres 6. U. sibogae coccosphere

9. U. foliosa coccoliths distal view

12. U. sibogae coccolith 13. U. s/hogae coccoliths 11. U. sibogae coccoliths distal proximal view distal view 3 Results 88

as Okada and McIntyre (1977) recombined C. foliosus into U. sibogae \ai. foliosa on the basis that “extremely rare specimens having both types of coccoliths on a single coccosphere” were observed. Curiously this taxonomic separation was upheld by later researchers, even as research on Umbilicosphaera demonstrated a stable morphology under culture conditions and diagnostic cytological differences between the species (Inouye and Pienaar 1984). Our own research - with data from biometry, molecular genetics, cytology and physiology - has now validated these findings and it seems likely that the coccospheres with both morphotypes observed by Okada and McIntyre (1977) may have been malformed specimens displaying atypical morphologies. Based on our own observations from cultures, morphology and on molecular data (Saez et al. 2003) the two variants were risen to species rank. Although three strains of U. sibogae and one strain of U. foliosa were maintained in culture for several years now there has been no evidence of a life-cycle change. Also despite the amount of SEM work performed on natural samples combination coccospheres have never been observed. So we have no evidence on the alternate stage of the life-cycle of these species although we predict that one will occur and that it probably will bear coccoliths.

3.3.4 Further methods

In general the same methods described in chapter 2 of this thesis were used. The biometrical routines however were adapted to cater for specific requirements. In the early stages of this thesis the two alternative hypotheses about the Umbilicosphaera taxa were that they either represent two discrete species or end morphologies of a single, rather variable species. Hence there was the need to identify characters use­ ful for species discrimination. A number of characters, both on the distal and the proximal shield were tested and it became obvious that due to the specific orienta­ tion of the calcite C-axis only two images per species, one taken in phase contrast and one in cross polarised light, were needed to characterise the taxa and so the macros were changed accordingly (Fig 1 - C). An unsuspected result from here is that the seemingly most obvious character discriminating the two species - the ratio of distal shield / proximal shield diameter (Fig. 1 - D,H) - turns out to be of little use for species discrimination with the LM. This can be explained by relatively large errors in correctly measuring the proximal shield diameter on the images taken in cross polarised light. Another morphological character tested was the distal shield rim width. As the central opening in U. foliosa is narrow and the distal shield diameter is larger com­ pared to U. sibogae the distal shield rim width captures two different aspects of morphological variation previously observed in the taxa. Hence the combination of distal shield rim width with proximal shield diameter proved to be reliable for morphological classification of the two taxa and was used for this study. 3 R esults 89

3.3.5 Cultures

Culture experiments on a strain of U. foliosa were performed by the author of this thesis and Blair Steel whilst studying for his master of science degree at the earth sciences department at UCL. In general the same methods as described in chapter 2, 3.1 and 3.2 were applied, for a detailed methodological description refer to Steel (2002). This thesis can be found on the compact disc in the appendix. Here we will only briefly review the data obtained during this experiment. The strain of U. foliosa (ESP6M1) used for the temperature experiment was isolated from the western Mediterranean, off Masnou (Spain, 100 km north of Barcelona). Experiments with clonal cultures of U. sibogae were repeatedly at­ tempted, but the cultures failed to grow under the experimental conditions.

Physiology

Six temperatures ranging from 13 to 27° C were used in this experiment. The clone tested (ESP6M1) grew over a temperature range from 17 to 23° C (compare figs 2,3). Cultures repeatedly inoculated at 13, 15 and 27° C did not grow and conse­ quently the cut-off temperature will be somewhere below 17 and above 23° C. The optimum growth temperature in the exponential phase was established at 23° C with an average k value of about 0.84 over the first 8 days of growth, after which stationary phase was reached. The culture maintained at 17° C however showed a long lag phase of about 14 day, after which growth occurred at a faster rate, leading to maximal cell densities after 37 days. Final cell density in the stationary phase ranged from ca. 218500 cells per ml for the culture maintained at 17° to ca. 824000 cells per ml for the culture maintained at 19° C. Counts however varied slightly as the cultures had the tendency to form clumps of cells with higher densities in the later stages of the experiment.

Coccolith morphology

According to the cell density coccoliths were harvested from this experiment on three different days. The mean proximal shield diameter was between 4.4 -4.8 /

3.3.6 Holocene samples

Seven Holocene coretop samples taken from ODP sites covering the North and South Atlantic Ocean, the Indian Ocean and the eastern Pacific Ocean and a box- core (BC3) from the Mediterranean were analysed (Figs 5,6). Two more sites were checked, but for both ODP site 704 (S. Atlantic Ocean) and ODP site 552 (N. Atlantic Ocean) the abundance of Umbilicosphaera spp. coccoliths proved to be 3 Results 90 too low to allow for a statistical analysis. Note that the sample from ODP site 872 is of early Pleistocene age and has been included as a comparison with the Holocene samples. During the initial image capturing phase it was unclear whether the two recognised varieties of Umbilicosphaera represented single morphological species with intergradational morphologies or two species with discrete morphologies. Later in the analysis unambiguous separation of the two species was possible with the recognition of additional qualitative characters and as a consequence the data set could be split into the two species, unfortunately resulting in a rather low number of

ESP6M1 growth rates

T in "C day c/m l A/ N2 / N1 k T in “0 day C/ml A f N2 1 N1 k 17 0 3300 19 03300 2 5260 2 1 59 0 3 4 1 8038 1 2 44 1 28 4 4720 2 090 -0 08 4 33429 3 4 16 069 6 5040 2 1 07 0 0 5 6 139000 2 4.16 1 03 8 5700 2 1 13 0.09 8 266000 2 1.91 0.47 10 5500 2 0 96 -0 03 10 339000 2 1.27 0 17 12 7100 2 1 29 0.18 12 687500 2 2 0 3 0.51 14 6970 2 0 98 -001 14 824000 2 1 20 0 13 17 7940 3 1.14 0.06 16 768000 2 0 93 -0 05 20 11240 3 1 42 0.17 18 572000 2 0 7 4 -021 23 13480 3 1.20 0 09 20 353500 2 0 62 -0.35 25 17200 2 1 28 0.18 22 329500 2 0 9 3 -0 05 27 20846 2 1 21 0.14 25 292500 3 089 -0.06 30 37000 3 1 77 0 28 28 118000 3 0.40 -0 44 32 52200 2 1.41 0 25 for day 0-14 14 250 0.57 34 170500 2 3 2 7 0 8 5 for day 14-28 14 0.14 -0 20 37 218500 3 1 28 0.12 for day 14-34 20 24.46 0.23 stationary phase not reached

T in °C day c/m i Û f N2 / N1 k 23 0 3300 1 9038 1 2 74 1 45 4 71667 3 7.93 1 00 6 215000 2 300 0 79 8 341000 2 1.59 0.33 10 307000 2 0 90 -0.08 12 277000 2 0.90 -0 0 7 14 200500 2 0 72 -0 23 16 95750 2 0.48 -0 53 18 201000 2 2.10 0.53 20 173500 2 0.86 -0.11 22 168333 2 0.97 -0.02 for day 0-8 8 103.33 0.84 for day 8-22 14 0.49 -0.07

Fig. 2. U. fo liosa growth rates for culture ESP6M1. For the calculation of growthrates refer to chapter 3.1 and 3.2. 3 Results 91

U. foliosa culture ESP6M1

1000 0 5 10 15 20 25 30 35 40 days

Fig, 3. Logarithmic plot of the cell densities in U. fo lio sa culture ESP6M1 for the tempera­ tures tested. specimens per species and sample.

Morphology

For U. fo lio s a mean proximal shield diameter in the samples varies from 4 - 5.5 //m with mean of 4.5 //m for all Holocene locations. Mean distal shield rim width varies from 1.6-2 jim with an overall mean of 1.8 /tm. The mean ratio of distal

U. foliosa coccolith size in culture ESP6M1 time since mean proximal shield mean distal shield t in ° C inoculation (days) n diameter (^m) S t. dev. rim width (pm) S t. dev 17 34 39 4.7 0.5 1.8 0.3 17 39 48 4.4 0.5 1.7 0.3 19 6 28 4.7 0.7 2.0 0.4 19 11 45 4.7 0.5 1.8 0.3 23 6 55 4.7 0.5 1.9 0.3 23 11 47 4.8 0.5 1.9 0.3 all samples 262 4.7 0.5 1.9 0.3

Fig. 4. Statistics for U. fo liosa clone ESP6M 1 coccolith morphology at different tempera­ tures. 3 Results 92 shield / proximal shield diameter ranges from 1-1.2 with an overall mean of 1.2 (compare fig. 6). For U. sib o g a e mean proximal shield diameter in the samples varies from 3.7 - 4 .7 i4m with mean of 4.1 //m for all Holocene locations. Mean distal shield rim width varies from 1.1 - 1.5 /rm with an overall mean of 1.2 /^m. The mean ratio of distal shield / proximal shield diameter ranges from 1-1.1 with an overall mean of 1 (compare fig. 6). These values agree well with the data taken from the cultured clone of U. fo lio sa (see 3.3.4). In most samples the morphospace occupied by the measurements reveals a bimodal pattern, which reflects the two species, in the sample from ODP site 849 however only U. fo lio s a is present and hence the plot is unimodal (Fig. 7).

B C 3 •- .ODP 659 A #

O D P 709 0

• ODP cores

Fig. 5. Map of the DSDP / OPD sites studied to determine the morphological variation in Holocene Umhilicosphaera spp. Sample BC3 is from a boxcore in the Mediterranean and was provided by Utrecht University. 3 Results 93

U. M iosa coccolith s 1 Holocene samples il shield lean projomal shield core location ipm) diameter (pm) / proximal sh. BC3 BC3,£E0019, 6.8-10.4cm 33.37 N 024.76 E ODP 659 108. 659A. INI, 2-4cm 18.08N021.02 W ODP 664 108, 664D, INI. 0-lcm 00.10 N 023.21 W ODP806 130. 8068, INI. 2-4cm 00.31 S 159.35 £ ODP 849 138, 8498. INI. 2-4 00.18 NI 10.53 W ODP 709 115. 709C, IHl, 3-5cm 03.90 S 0 60.55 E ODP 872 144. 872A. 1H2, 127cm 10.09 N 162.85 E ail samples

in distal shield r U. M iosa coccolith s 1 Holocene samples continued width (pm)

0.3 1,6 0.2 1.5 0.3 1.8 0.5

U. sibogae coccolith size in Holocene samples

mean distal shield mean proximal shield mean distal sh. core location n diam eter (*ot ) St. dev. diameter (pm) St. dev. / proximal sh. BC3.EE0019, 6.8-10.4cm 33.37 N 024.76 E 13 4.6 0.4 4.7 0.5 1.0 108, 659A, IHl. 2-4cm 18.08 N 021.02 W 15 4.0 0.3 3.7 0.5 1.1 ODP 664 108, 6640. IHl. 0-lcm 00.10 N 023.21 W 20 3.8 0.3 3.7 0.4 1.0 ODP806 130. 806B. IHl. 2-4cm 00.31 S 159.35 £ 10 4.0 0.4 0.5 1.0 ODP 849 138. 849B. INI, 2-4 00.18 N 110.53 W 2 115, 709C. IHl, 3-5cm 03.90 S 060.55 E 4.2 0.3 0.3 1.0 144, 872A. 1H2, 127cm 10.09 N 162.85 E all samples 0.4 0-4 1.0

mean co diameter in distal shield r U. sibogae coccolith s 1 Holocene samples continued width (pm)

Fig. 6. DSDP / ODP core location and statistics for all Holocene samples of Umhilicosphaera spp. studied. 3.3.7 Downcore samples (ODP site 664)

A series of 17 samples (Fig. 8) taken from ODP site 664 was analysed covering a time interval from 0 to 3.482 Ma. However some samples had to be excluded from the analysis as the abundance of U. s ib o g a e or U. fo lio sa was either too low, or because only fossil Umbilicosphaera species - U. rotu la and U. ja f a r i - w ere present (see also 3.3.8 for a discussion). The age model used is based on few mag- netochronological datums (Brunhes/Matuyama and Jaramillo-top and bottom) plus nannobiochronological data mostly from LAD of D is c o a s te r forms (Su 1996). As with the Holocene samples the resulting data was divided into the two spe­ cies subsequent to image capturing, which resulted in a low number of specimen per species and sample.

Morphology

For U. f o lio s a mean proximal shield diameter in the samples varies from 3.9 - 4.8

Fig. 7 (next page). Density plot of coccolith proximal shield diameter and distal shield rim width for all Holocene samples of Umbilicosphaera spp. studied. 3 Results 94

Umbilicosphaera spp. ; Holocene

O D P 7 0 9 O D P 8 4 9

O D P 8 0 6 BC 3

O D P 8 7 2 O D P 6 5 9

n-15

O D P 66 4 all samples

20 25 30 35 40 45 50 55 60 6,5 70 3 R esults 95 pim with mean of 4.3 pim for all downcore samples. Mean distal shield rim width varies from 1.6-2 pim with an overall mean of 1.8 pim. The mean ratio of distal shield / proximal shield diameter ranges from 1 -1 .3 with an overall mean of 1.2 (compare fig. 8). For U. sibogae mean proximal shield diameter in the samples varies from 3.6 - 4.2 pim with mean of 3.9 pim for all downcore samples. Mean distal shield rim width varies from 1.0 - 1.2 pim with an overall mean of 1.1 pim. The mean ratio of distal shield / proximal shield diameter ranges from 1 -1 .2 with an overall mean of 1.1 (compare fig. 8). These values agree well with the data taken from the cultured clone of U. foliosa (see 3.3.5) and with the Holocene samples (see 3.3.6). Unlike in the Holocene sam­ ples, where the shield ratio proved useless for the discrimination of the two taxa the

Fig. 8 (next page). ODP 664 core location and statistics for all samples of Umbilicosphaera spp. studied. 3 Results 96

U. foliosa coccolith ize of ODP site 6 6 4 sam ples mean distal shield mean proximal shield mean distal sh. sample core location sub-bottom depth (m) age in Ma diameter (pm) St dev diameter (pm) St. dev. / proximal sh. ODP664D0.01 0.1 N 23.21 W 0.01 0 .0 0 0 9 4.7 0.5 0.5 1.1 0DP664D1.S2 1.52 0.045 14 5.4 0.5 4.4 0.4 1.2 ODP664D 332 3.32 0 .0 9 8 18 5-7 0.9 4.8 0.4 1.2 CDP864D5.52 5.52 0 .164 21 5.3 0.5 4.3 0.3 1.2 ODP664D8.52 8.52 0 .253 22 5.5 0.6 4.5 0.3 1.2 ODP664D12.02 12.02 0.357 19 5.6 0.5 4.5 0.3 1.2 0DP664D15.02 15.02 0 .446 12 5.1 0.6 4.2 0.3 1.2 ODP664D26.92 26.92 0.755 19 5.2 0.7 4.3 0.3 1.2 ODP664D35.82 35.82 0,991 6 5.0 0.5 4.0 0.2 1.3 ODP664D45.92 45.92 1.242 19 4.9 0.4 3.9 0.3 1.3 ODP664D60.44 60.44 1.566 15 5.2 0.4 4.3 0.3 1.2 ODP664D68.90 68.90 1.755 4 5.2 0.2 4.1 0.3 1.3 ODP664D75.93 75.93 1.900 25 5.2 0.5 0.3 1.3 ODP664D104.42 104.42 2,436 ODP664D109.34 1 09.34 2.529 23 5.3 0.4 0.2 1.3 ODP664DÎ 25.84 125.84 2.852 no analysis ODP664D1 66-82 1 66.82 3.992 only U. rotula and U. jafari p rese n t all sam ples 227 5.2 0.5 4.3 0.3 1.2

mean co diameter m ean distal shield rim U. foliosa coccolith ize of ODP site 66 4 sam ples continued (pm) width (pm )

U. Sibogae coccolith size of ODP site 66 4 sam ples

mean distal shield mean proximal shield mean distal sh. sub-bottom depth (m) age in Ma n (Sameter (pm) St. dev. diameter (pm) St. dev. / proximal sh. ODP664D0.01 0.1 N 23.21 W 0.01 0 .0 0 0 20 3.8 0.3 3.7 0.4 1.0 ODP664D1.52 1.52 0-045 16 4.1 0.3 0.3 1.0 0DP664D3 32 3 32 0 .0 9 8 10 3.8 0-4 3.7 0.4 1.0 ODP664D5.52 5.52 0 .164 9 4 2 0 5 4.1 0.5 1.0 ODP664D8.S2 8.52 0 .253 8 4 3 0.3 4.1 0.2 1.0 ODP664D12.02 12.02 0 .357 11 4.2 0-4 4,1 0,3 1,0 0DP664D1 5.02 15.02 0 .446 18 3.7 0.5 0,5 1.0 ODP664D26.92 26.92 0 .755 11 3.9 0.6 0.5 1.0 ODP664D35-82 35.82 0.991 24 3.9 0.3 0.3 1.0 0DP664D45.92 45.92 1.242 11 4.1 0.2 0.2 1.0 ODP664O60.44 60.44 1.566 15 3.9 0.4 0,4 1.0 ODP664O68.90 68.90 1.755 0 no analysis 0DP664D75.93 75.93 1.900 3 4.2 0.4 0.4 1.0 00P664D104.42 104.42 2 .436 ODP664D109.34 109.34 2.529 no U. sibogae present ODP664D1 25.84 125.84 2.852 00P664D166.82 166.82 3.992 all sam ples 156 4.0 0.4 3.9 0.4 1.0

mean co diameter mean distal shield rln U. sibogae coccolith size of ODP site 664 samples continued (pm) St. dev. width (pm ) St. dev. 1,6 0.2 0.2 1.8 0.3 0.1 1.7 0.2 1.1 0.1 1.9 0-3 1.2 0.2 2.0 0.2 1.1 0.2 2.2 0,3 1.0 0.1 1.8 0.3 1.0 0.2 1.8 0.5 1.0 0.1 1.8 0.3 1.1 0.2 1.9 0.1 1.1 0.1 1.9 0.2 1.0 0.2 analysis 1.7 0-2 1.2 0.2

no U. sibogae p resent 3 R esults 97 morphological differences are clearly reflected in both the distal shield rim width and the proximal / distal shield ratio in the downcore samples. Note that the 3.99 Ma timeslice in figure 9 displays the morphospace occupied by the two extinct species U. rotula and U. jafari. The time series reveals no gradualistic morphological evolution of either taxon, hence most of the measurements occupy a discrete bimodal morphospace with the relative abundance of the two taxa in the samples being reflected by varying densi­ ties.

Fig. 9 (next page). Density plot of coccolith proximal shield diameter and distal shield rim width for all Holocene samples of Umbilicosphaera spp. studied. 3 R esults 98

Umbilicosphaera spp. 0.253 Ma 1.242 Ma morphology of coccoliths in O D P s ite 6 6 4

Contour lines 0.5 • ma* @ 05 intervals Contours < 15 are dashed OM a 0.357 Ma 1 566 Ma

f/,-

0.045 Ma 0.446 Ma 1.9 Ma

(lO,

0 098 Ma 0 755 Ma 2.529 Ma

0.164 Ma 0.991 Ma 3.992 Ma

2.0 2.5 3 0 3 5 4 0 4 5 5.0 5.5 6 0 6 5 7.0 proKinal shieta diameter m pm 3 R esults 99

3.3.8 First occurrence

Few precise data on the first occurrence of the two taxa are available. In most of the biostratigraphic work Umbilicosphaera species are lumped together, partly because they have been considered variants, partly because they are relatively rare in fossil samples and partly due to the difficulty of recognising them correctly in the light microscope - which requires careful comparison of cross-polarised light and phase contrast images. Own biostratigraphic work dates the first occurrence of U. sibogae at 2.2 Ma and that of U. foliosa at 2.9 Ma. These ages have been derived from care­ ful re-examination of ODP site 664. Prior to this the two extinct species U. jafari and U. rotula co-occur back to Early Miocene (Young 1998). Figure 10 shows the ranges of the extant and fossil Umbilicosphaera species and their tentative evolu­ tionary history. Two alternative hypotheses can be derived from here:

A) U. foliosa and U. sibogae have separate evolutionary histories with their ancestral species being U. jafari and U. rotula respectively. Evidence here arises from the close morphological and structural relationship between U. foliosa and U. jafari and U. sibogae and U. rotula. In this case the divergence of U. sibogae and U. foliosa would date back to the divergence of U. jafari and U. rotula, palaeontologically dated at ca. 20 Ma. B) Both U. foliosa and U. sibogae have a conomon origin from either U. jafari or U. rotula. This would date their divergence time at ca. 2.2 Ma.

2 1 Ma

] I i : 1 1 I 1 1 1 ! I___ U. sibogae

U. rotula cH U. hulburtiana

5.6 (±1.2) Ma I ' ' P&'______U. foliosa

U. jafari

Fig. 10. Ranges (boxes - own data, grey lines - literature) and hypothetical evolutionary relationships (broken lines) of Umbilicosphaera spp. In this model U. sibogae and U. fo lio sa are not sister taxa, but share a common ancestor at 20 Ma, which is seemingly incompatible with estimates from the molecular clock (yellow box). For further discussion see text. S - {/. sibogae, R - U. rotula, H - [/. hulburtiana, ¥ - U. foliosa, } -U . jafari.

It was however not possible to test these hypotheses from the geological record as no gradualistic morphological change was observed between tentative ancestral and derived species in the studied cores. Recently estimates from a molecular clock for the Calcidiscaceae have become available for the divergence of U. sibogae and U. foliosa. This clock is based on the tufA tree and dates the divergence of the lineage leading to the two species back at 3 Results lÜÜ

5.59 (±1.15) Ma (Sâez et al. 2003). This leads to three alternative hypotheses;

1 ) The two taxa differentiated morphologically much later than they did geneti­ cally (Fig. 11 - A). 2) U .fo lio sa and U. sih o g a e extend much further into the geological record (Fig. 11-B). 3) The molecular clock is inaccurate (Fig. 11 - C).

There is sound evidence that morphological evolution can indeed be strongly un­ coupled from genetic change (cryptic or pseudo-cryptic spéciation, see Know 1 ton 1993). Assuming that this is the case, that the two taxa are sister taxa, and that the biostratigraphical ranges are correct, scenario (A) in figure 11 displays the resulting evolutionary pattern derived from morphological analysis. Equally it is possible that the biostratigraphical ranges for the taxa do not reflect the true ranges of the species but are influenced by the palaeobiogeography of the taxa. In this case scenario (B) in figure 11 would display the evolutionary relation­ ships. However, although only ODP site 664 was studied in detail this is an unlikely explanation as neither of the taxa - U .fo lio sa and U. sih o g a e - was present in any of the 4 Ma samples and according to Young (1998) both taxa have a first occurrence in nannofossil zone 16 (ca. 3.8 - 2.6 Ma). For any molecular clock there are a number of potential sources of error. One problem arises from the calibration, which often carries an error of several Ma. Another problem is the combination of a fast evolving gene with a relatively old calibration node. In this case ages are calculated erroneously due to the gene be­ coming saturated. The divergence date from the molecular clock seems to be clearly incompatible with scenario (C) in figure 11 (compare also fig. 10), which dates the divergence of the lineage leading to the taxa back at 20 Ma. However, the morpho­ logical and structural similarity between U .fo lio sa and U .ja fa r i and U. sihogae and U. rorula makes them possible sister taxa, supporting scenario (C).

M a

mol cular dock estim te f 5.6 (±1.2) Ma

Fig. 11. Possible evolutionary scenarios for Umbilicosphaera spp. The yellow bock high­ lights the divergence time for U .foliosa and U. sihogae (Sâez et al. 2003). Note that in both scenarios A and B the red green clade could alternatively be originating from U .jafari. For further discussion see text. S - U. sihogae, R - U. rotula, W - U. hiilburtiana, F - U .foliosa, J - U .jafari. 3 R esults 101

3.3.9 Conclusions

The key observations on the two Umbilicosphaera species are:

1. Measurements of coccoliths from cultured U .foliosa reveal a unimodal, well defined morphology (see Msc thesis Steel 2002) 2. In the U. foliosa culture used there was no temperature control on coccolith size. 3. There is morphological, structural and molecular evidence that the two previ­ ously described variants of Umbilicosphaera indeed represent distinct species, 4. Compared with other cultured coccolithophores growth rates are relatively high, and optimum growth is at a well defined temperature.

The morphometric data indicates that the two previously recognised Umbilicosphaera variants are separate biological species, with a stable morphology without intermediate morphologies in culture, Holocene samples or the fossil record. Transitional morphotypes have not been observed in culture and quantitative morphology allows for a clear morphological separation of the taxa. Under the assumption that both Umbilicosphaera species are sister taxa molecular data points to a divergence time between 5.59 (±1.15) Ma (Sâez et al. 2003). Biostrati graphie work however allows for a different interpretation of the evolutionary history. Certainly further research is needed here, especially focusing on sequencing U. hulburtiana which is the prime candidate as a sister taxon for U.foliosa.

3.3.10 References

Baumann K-H, Sprengel C (2000) Morphological variations of selected coccolith species in a sediment trap north of the Canary Islands. J nannoplankton Res 22: 185-193 Borsetti A M, Cati F (1976) 11 Nannoplancton calcareo vivente nel Tirreno Centro-meridion- ale. Parte 11. G Geol 40: 209-240 Gaarder K R (1970) Three new taxa of Coccolithineae. Nytt Mag Bot 17: 113-126 Heimdal B R (1993) Modem Coccolithophorids. A Guide to Naked Flagellates and Coccolithophorids. In: Tomas C R (ed) Marine Phytoplankton. Academic Press, London, pp 147-243 Inouye 1, Pienaar R N (1984) New observations on the coccolithophorid Umbilicosphaera sihogae var. fo lio sa (Prymnesiophyceae) with reference to cell covering, cell structure and flagellar apparatus. Br phycol J 19: 357-369 Kamptner E (1963) Coccolithineen-Skelettreste aus Tiefseeablagemngen des Pazifischen Ozeans. Annin naturh Mus Wien 66: 139-204 Kleijne A (1993) Morphology, taxonomy and distribution of extant coccolithophorids (Calcareous nannoplankton). PhD thesis. Free University Amsterdam (ISBN 90- 9006161-4), p. 321 Knowlton N (1993) Sibling species in the sea. A rev ecol syst 24: 189-216 3 R esults 102

Lohmann H (1902) Die Coccolithophoridae, eine Monographie der Coccolithen bildenden Flagellaten, zugleich ein Beitrag zur Kenntnis des Mittelmeerauftriebs. Arch Protistenkd 1: 89-165 McIntyre A, Bé A W H (1967) Modem Coccolithophoridae of the Atlantic Ocean - I. Placoliths and Cyrtholiths. Deep-Sea Res 14: 561-597 Murray G, Blackman V H (1898) On the nature of the Coccospheres and Rhabdospheres. Phil Trans R Soc Ser B 190: 427-441 Nishida S (1979) Atlas of Pacific Nannoplanktons. News Osaka Micropaleontol Special Paper: 1-31 Okada H, McIntyre A (1977) Modem coccolithophores of the Pacific and North Atlantic Oceans. Micropaleontology 23: 1-55 Sâez A G, Probert I, Geisen M, Quinn P, Young J R, Medlin L K (2003) Pseudo-cryptic spé­ ciation in coccolithophores. Proc natn Acad Sci USA 100: 7163-7168 Schiller J (1930) Coccolithineae. In: Rabenhorst L (ed) Kryptogamen-Flora von Deutschland, Osterreich und der Schweiz. Akademische Verlagsgesellschaft, Leipzig, pp 89-267 Steel B A (2002) Physiology, growth and morphometry of selected extant coccolithophorids, with particular reference to Calcidiscus leptoporus (Murray and Blackman, 1889) Loeblich and Tappan 1987. MSc thesis. University College London, p. 83 Su X (1996) Development of late Tertiary and Quatemary coccolith assemblages in the Northeast Atlantic. Geomar Report 48: 1-120 Weber - van Bosse A (1901) Études sur les algues de l’Archipel Malaisien III. Note prélimi­ naire sur les résultats algologiques de l’expédition du Siboga. Annls Jard bot Buitenz 17: 126-141 Winter A, Siesser W G (1994) Atlas of living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 107-159 Young J R (1998) Neogene. In: Bown P R (ed) Calcareous Nannofossil Biostratigraphy. Chapman & Hall, London, pp 225-265 3 R esults 103

3.4 Life-cycle observations involving pairs of holococcoiithophorid species : intraspecific variation or cryptic spéciation?

This is a research paper, which was published in European Journal of Phycology, vol. 37 (2002), pp. 531-550. This manuscript deals with coccolithophore life-cycles and we present here a number of observations, both from wild samples taken on research cruises, but also from cultured material. The key observations here are from samples taken whilst onboard the R/V Hesperides in the Alboran Sea, western Mediterranean in autumn 1999. The chance to participate in this cruise was offered to us by Kees van Lenning (CSIC Barcelona) and so a hastily assembled international team consisting of the author, Alexandra Broerse (then Free Univ. Amsterdam) and Andy Howard (then University College London and NHM) flew out to Malaga after a couple of days of interesting logistical work. By chance a whole team of French scientists had at the last minute withdrawn from this cruise - and as a consequence we had the opportunity to obtain as much water as we wanted from the CTD that the physical oceanographers were running at every station. They simply had no use for the water! Suddenly we had enough water to sample for a whole range of CODENET activities and we also - and to our knowledge for the first time - sampled the deep chlorophyll maximum to isolate living coccolithophores. The cruise went on for 10 days of hard work, but at the end of it we had a set of samples with a very good coverage of the Alboran Sea. Back in the laboratory at the NHM Ian Probert (Univ. Caen) was waiting and we immediately performed isolations on concentrated seawater. We turned out to be extremely successful - a number of species urgently needed in CODENET and a number of additional species, two of them from the deep photic zone were isolated into culture and proved to be important for the interpretation of later observations. Whilst performing SEM analyses of natural samples from the cruise the author however discovered key specimens that led to compelling new insight into processes of spéciation in coccolithophores. Research like this cannot be planned - luck is involved to be at the right place at the right time. However, the Mediterranean has an attraction to coccolithophore workers - Lohmann"s seminal work on living coccolithophores was done just off Syracuse - and I still think it is worth a further trip or a cruise. The author of this thesis performed the sampling for the majority of new observations discussed in this paper, has made the key observations using SEM and has written the manuscript. Alexandra Broerse contributed observations on combination and Jeremy Young and Ian Probert helped with the interpretation of the data. Chantai Billard provided additional insights in haptophyte life cycles. Lluisa Gros (CSIC Barcelona) provided additional examples of combinations and her taxonomical skill has been a great help throughout the project. 3 Results 104

Eur. J. Phycol. (2002), 37: 531-550. © 2002 British Phycological Society 531 DOI: 10.1017/80967026202003852 Printed in the United Kingdom Life-cycle associations involving pairs of holococcoiithophorid species: intraspecific variation or cryptic spéciation?

MARKUS GEISEN’, CHANTAL BILLARD\ ALEXANDRA T.C. BROERSE^, LLUISA CROS\IAN PROBERT^ AND JEREMY R. YOUNG’

' Palaeontology Department, The Natural History Museum, Cromwell Road, London SW7 5BD, UK ^Laboratoire de Biologie et Biotechnologies Marines, Université de Caen, Esplanade de la Paix, 14032 Caen, France ^Faculty of Earth Sciences, Vrije Universiteit, De Boelelaan 1085, 1081 H V Amsterdam, The Netherlands * Institut de Ciències del Mar (CSIC), Passeig Joan de Borbô s/n, 08039 Barcelona, Spain

{Received 15 July 2001 ; accepted 10 M ay 2002)

New holococcolith-heterococcoiith life-cycle associations are documented based on observations of combination coccospheres. Daktylethra pirus is shown to be a life-cycle phase of Syracosphaera pulchra and Syracoliihus quadriperforalus a life-cycle phase of Calcidiscus leptoporus. In addition, new observations from cultures confirm the life-cycle associations of Crystallolithus braarudii with Coccolithus pelagicus and o f Zygosphaera hellenica with Coronosphaera mediterranea. In all four cases previous work has shown that the heterococcolithophorid species is associated with another holococcoiithophorid. Two other examples of a heterococcolithophorid being associated with two holococcolithophorids have previously been identified, so this seems to be a common phenomenon. The six examples are reviewed to determine whether a single underlying mechanism is likely to be responsible for all cases. It is concluded that there is no single mechanism but rather that the six examples fall into three categories: (a) in two cases the holococcolith types are probably simply ecophenotypic morphotypes; (b) in two other cases the holococcolith types are discrete and are paralleled by morphometric differences in the heterococcolith types; (c) in the final two cases the holococcolith types are discrete but are not paralleled by any obvious morphological variation in the heterococcolith morphology. We infer that cryptic spéciation may be widespread in heterococcolithophorid phases and that study of holococcoiithophorid phases can provide key data to elucidate this phenomenon.

Key words: coccolithophorids, cryptic spéciation, haptophytes, holococcolith-heterococcoiith combinations, intraspecific variation, life-cycles

Introduction apparently occurs outside the cell membrane within an organic ‘ skin’ which surrounds the cell (Rowson Coccolithophorids are unicellular marine photo­ et a i, 1986), but the regulatory mechanisms for synthetic algae in the division Haptophyta (syn­ this process remain poorly understood (for reviews onym Prymnesiophyta). They form a major com­ of coccolith morphology and formation see ponent of the oceanic microplankton and are one of Leadbeater, 1994; Pienaar, 1994; Young et al, the main open ocean primary producers. Tradition­ 1999). ally the taxonomy of this group has been based on The dominant reproductive mode of haptophytes morphological characters of the minute calcite is asexual mitotic division. However, in many plates that cover the cell, the coccoliths, of which haptophytes more complex life-cycles have been two major types, heterococcoliths and holo- documented, with two or more morphologically coccoliths, can be distinguished. Heterococcoliths distinct phases (Fig. 1). Evidence for this has come are formed of a radial array of complex crystal units from two main sources; (1) culture observations of of variable shape whereas holococcoliths are formed phase transitions (e.g. Parke & Adams, 1960), of numerous minute identical euhedral crystallites. supported by direct observation of meiosis and Calcification of the heterococcoliths takes place syngamy (e.g. Gayral & Fresnel, 1983), and intracellularly and is consequently under strong chromosome counts (Rayns, 1962; Fresnel, 1994) in cellular control (e.g. Westbroek et al., 1984). In cultured clones; (2) rare observations from natural contrast biomineralization of the holococcoliths populations of combination coccospheres bearing different coccolith types interpreted as representing Correspondence to: M. Geisen. Fax: +44 (0)207 942 5546. the moment of life-cycle phase transition (e.g. e-mail: M.Geisenfi-nhm.ac.uk alternation of haploid and diploid phases). 3 Results 105

M. Geisen et al. 532

Coccolithophorid life-cycles

Meiosis M ito sis Mitosis

ploidy level irom flow cytometry and î nuclear staining

heterococcolith Syngamy holococcolitri stage

Fig. 1. Schematic representation of coccolithophorid life-cycles. The diploid stage is covered with heterococcoliths which are produced intracellularly, whereas the motile haploid stage is covered with holococcoliths which are produced outside the cell membrane. Most examples given in this paper would represent syngamy of two haploid gametes with both holococcoliths and heteroccoccoliths tieing present on a single coccosphere.

Summarizing available data. Billard (1994) inferred haploid holococcolith-bearing phase (Bown, 1998; that there was a common pattern of alternating Y ou n g et ai, 1999, 2000). haploid and diploid phases, both of which were Amongst the reported het-hol combinations, capable of mitotic reproduction, characterized by three cases of one heterococcolithophorid forming consistent differences in their coccolith and scale separate combinations with two holococcolitho- covers. Of particular relevance, she noted that phorids have been reported (see Gros et ai, 2000). available evidence suggested that heterococcoliths For each of these cases of het-hol-hol associations. were characteristic of diploid phases and holo­ G ros et al. (2000) concluded that this phenomenon coccoliths of haploid phases of coccolithophores. is most likely the result of non-genotypic variation This hypothesis has been directly tested, and in the degree of calcification of the holococcoiitho­ supported, by investigation of ploidy level in phorid phase. Three new examples of het-hol cultures of Emiliania hii.xieyi (G reen et ai, 1996) associations are reported here, and in each case the and Coccolithus pelagicus (Y ou n g et ai, 2000). heterococcolithophorid has previously been found Combination coccospheres bear two coccolith in association with a different holococcoiitho­ types which were traditionally regarded as belong­ phorid. There are therefore now a total of six ing to separate species. Numerous new examples of het-hol-hol associations, representing nearly one- holococcoliths and heterococcoliths forming com­ third of all associations discovered to date. This bination coccospheres in field samples have recently increasingly common pattern of one heterococcolith been observed (Cortés, 2000; Gros et ai, 2000; being associated with two holococcoliths is Renaud & Klaas, 2001), indicating that they are intriguing and suggests a common cause; several alternate phases of the life-cycle of single species. possible explanations can be postulated. From such data, associations of ‘species’ are in­ The first possibility is hybridization occurring ferred. Het-hol associations involve one hetero­ between the gametes (i.e. haploid phases) of two coccolith species and one holococcolith species. The closely related species. Hybridization is a wide­ available examples span the biodiversity of cocco­ spread phenomenon within higher plant taxa and lithophorids, strongly supporting the hypothesis some macroalgae (e.g. Gosson et ai, 1984; Gosson, that the primitive state for coccolithophorids is to 1987), though hybridization of marine protists have a diploid heterococcolith-bearing phase and has not been documented. The available evidence 3 R esults 106

Life-cycle associations involving pairs o f holococcoiithophorid species 533 from cultures (e.g. Gayral & Fresnel, 1983) suggests EU-TMR Coccolithophorid Evolutionary Bio­ that all haploid cells can act as gametes. Hence diversity and Ecology Network (CODENET); the hybridization between closely related coccolitho­ disparate data are combined here to allow timely phorids would be predicted to give rise immediately synthesis of this topic. after syngamy to a combination coccosphere bearing holococcoliths of the two species. A second possibility is complex life-cycles; phyto­ Materials and methods plankton life-cycles often include a range of morphologies. Even assuming that the basic het-hol Taxonomic nomenclature division corresponds to a separation between dip­ Traditional coccolithophorid taxonomy was established loid and haploid phases, it is possible that the using the morphological characters of the coccoliths haploid phase may have more than one covering the cell, and crystallographic orientation of the morphotype. Variable cell morphologies have been component crystal units (e.g. Bown, 1998; Young et a l, documented in the haploid phase of the life-cycle of 1999). This taxonomy has been successfully applied to the haptophytes such as Pleurochrysis pseudo- fossil record and compares well with findings from other roscoffensis (Gayral & Fresnel, 1983), other species characterization methods such as cell ultrastructure of Pleurochrysis (Fresnel & Billard, 1991) and and more recently molecular genetics. The discovery that heterococcoiithophorids (HE) and holococcoli tho­ Phaeocystis (Lancelot & Rousseau, 1994), although phorids (HO) can be formed by single species in different at present there is no indication that coccolith phases of the life-cycle does not invalidate this taxonomy morphology varies within a phase. but it does lead to nomenclatural problems. There has A third possibility is sexual dimorphism; if a been much debate as to how nomenclatural taxonomy haplo-diplontic life-cycle also involves sexual should be adjusted to reflect these observations. We differentiation, then discrete male and female hap­ strongly agree with Cros et al. (2000) and Silva (personal loid gametes, potentially with differing holo­ communication) that once such associations are estab­ lished a single scientific name should be adopted for all coccolith morphologies, would be formed. This is a phases following the normal rules of botanical nomen­ well-known phenomenon in other algal classes (e.g. clature, with informal terminology used to indicate the centric diatoms (von Stosch, 1954), cryptomonads phase observed where appropriate. However, since this (Hill & Wetherbee, 1986) and dinoflagellates (von publication is concerned with establishing such associ­ Stosch, 1972)). ations, for clarity we use the original ‘species’ names for Fourth, these observations may represent intra- newly associated phases; the correct names for future speciiic variation, without genotypic control. Holo­ usage, based on nomenclatural priority, are given in the coccolith taxonomy is entirely based on coccolith Results and Discussion section. morphology and since very few holococcolitho- phorids have been maintained in culture there is little information on the degree of variability in Field samples coccolith morphology possible within one species. The key specimens reported here were found in samples Cros et al. (2000) speculated that intraspecific collected from the Alboran Sea (Western Mediterranean) variation in the degree of calcification of the during a cruise of the Institut de Ciènces del Mar (CSIC) holococcolith phase was a possible explanation for on board the R/V Hesperides in 1999 as part of the MATER research project (Mater II, from 26 September the het-hol-hol associations they observed, par­ to 6 October). ticularly since the holococcoliths involved were During this cruise water samples were obtained at morphologically rather similar. selected depths using a rosette sampler with Niskin bottles Finally it is conceptually possible for spéciation attached to a Conductivity, Temperature, Depth (CTD) to occur without obvious morphological change probe. Depending on the concentration of phyto­ (cryptic spéciation) ; indeed molecular genetic results plankton, up to 1000 ml of seawater were filtered using for certain protist groups have suggested that this vacuum filtration. Two types of filters were used: (1) may be a common phenomenon (e.g. Andersen et 25 mm cellulose nitrate filters with 045 ^m retention (Whatman) and (2) 25 mm polycarbonate filters with ai, 1998; Darling et a i, 2000). As a variant we can 0 47 pm pore size (Millipore). Salt was removed by rinsing imagine that morphological change following the filters with mineral water. The filters were oven-dried spéciation may be apparent in the holococcolith at 50 °C for 1 h. For scanning electron microscopy (SEM) phase but cryptic in the heterococcolith phase. a part o f the filter was mounted on a stub and coated with In this paper we present our new evidence of gold-palladium before examination in a Philips XL-30 het-hol-hol associations and, for each case, review PEG or a Cambridge Stereoscan S250 microscope. For which of these potential causes can most reasonably light microscopy (LM) a part of the cellulose nitrate filter was mounted with immersion oil on a slide and covered be invoked to explain this interesting and with a coverslip. LM observations were performed using increasingly commonly detected phenomenon. Our a Zeiss Axioplan with a Hamamatsu CCD video camera. data come from four separate research groups LM was used both for morphological observations and to working on diverse research tasks within the larger determine the crystallographic orientation of the con- Table 1. Sample locations for all combination coccospheres discussed in the text. Where possible the direction of the life-cycle phase transition is given

Transition Combination cells Location Date Station Latitude Longitude Depth (m) Type Fig. mode Notes BMNH image ref. Reference

H . carteri/Sl. catilUferus MESO-96 NW M editerranean Jun.-Jul. 1998 G4 41“ 08-60' N 02“ 45-02' E 70 SLM Not figured Hol het C ros ei al. 2000 Plate 1, Fig. 4 H . carteri/Sl. catilUferus FRONTS-95 NW M editerranean Jun. 1995 24 W 40“ 33 90' N 02“ 38-70' E 70 SLM Not figuredUnknown Cros el al. 2000 - Plate 1, Fig. 3 H . carteri/Sl. confusus Mediterranean 1955 75 LM Not figured Unknown Lecal-Schlauder, 1961 - Photo 4, 5 SI. confusus/SI. catilUferus Meteor 36/2 NE Atlantic 178 33“ 00-20' N 22“ 00-00' W 20 SLM Not figured C ros el al. 2000 - Plate 1, Fig. 6 SI. confusus/SI. catilUferus MESO-96 NW Mediterranean Jun. 1996 F2 41“ 27 02' N 02“ 52-00' E 5 SEM n/a This publ. SI. bannockii/Corisphaera MESO-96 NW M editerranean Jun.-Jul. 1998 G6 40“ 56-30' N 02“ 56-70' E 40 SEM Not figured ?hct4iol C ros ei al. 2000 - Plate 7, Fig. 3 sp. type A Corisphaera sp. type A/ FANS-I NW M editerranean Nov. 1996 127 (141) 39“ 52-80' N 00“ 54-00' E 5 SEM N ot figured C ros ei al. 2000 - Plate 7, Fig. 5 Z. bannockii Corisphaera sp. type A/ Sonne 117 Indian Ocean 20/3 14“ 29-70' N 64“ 44-40' E 20 SEM Not figured Cros el al. 2000 - Plate 7, Fig. 6 Z bannockii Corisphaera sp. type A/ FANS-I NW M editerranean Nov. 1996 123 39“ 59-60' N 00“ 44-40' E 40 SEM 10 n/a This publ. Z. bannockii Corisphaera sp. type A/ FANS-1 NW M editerranean Nov. 1996 127 39“ 52-80' N 00“ 54-00' E 5 SEM N ot figured This publ. Z. bannockii Corisphaera sp. type A / FANS-I NW M editerranean Nov. 1996 127 39“ 52-80' N 00“ 54-00' E 40 SEM 8, 9 n/a This publ. Z. bannockii Cl. pelagiciis/Cr. hyalinus M eteor 7 N. Atlantic Sept. 1985 10 72“ 13-00' N 16“ 05-00' W Surface SEM N ot figured H ol-het Cl. pelagicus Samtleben & Schroder, 1992 - Plate 1, small morphotype Fig. 8 Cl. pelagicus/Cr. hyalinus N. Atlantic, Jun. 1986 2 65“ 30-00' N 00“ 08-00' W 15 SEM 13 H ol-het Cl. pelagicus Samtleben in Norwegian Sea small morphotype W inter & Siesser, 1994 Cl. pelagicus/Cr. hyalinus N. Atlantic, Jun. 1986 2 65“ 30*00' N 00“ 08-00' W 15 SEM Not figured Hol-het Cl. pelagicus Samtleben & Bickert, 1990 - Plate 1, Norwegian Sea small morphotype Fig. 8 Cl. pelagicus/Cr. hyalinus N. Atlantic, Jun. 1986 2 65“ 30-00' N 00“ 08-00' W 15 SEM Not figured I lol het 3 observations, This publ. Norwegian Sea d . pelagicus small morphotype Cl. pelagicus/Cr. hyalinus ARK V II/I N. Atlantic, Jun. 1990 43 70“ 45-00' N 05“ 30-00' W Surface SEM N ot figured Hol -het Cl. pelagicus Baum ann el al. 1990 - Plate 1, Norwegian Sea small morphotype Fig. 1 Cl. pelagicus/Cr. hyalinus N. Atlantic, Sept. 1988 554 72“ 00-00' N 13“ 00-00' W 10 SEM 14 Het-hol 2 observations, This publ. Norwegian Sea Cl. pelagicus small morphotype Cl. pelagicus/Cr. hyalinus OG33A 500 SEM Not figured Het-hol Sediment trap, Andruleit Cl. pelagicus small morphotype Cl. pelagicus/Cr. hyalinus N. Atlantic, Sept. 1988 552 71“ 38-00' N 08“ 25-00' W 32 SEM Not figured Cl. pelagicus This publ. Norwegian Sea small morphotype Cl. pelagicus/Cr. braarudii Arcachon, SW France SEM, LM, 15 Change observed This publ. TEM in culture (LKIA CF4-5, KL2) Cl. pelagicus/ Cr. braarudii English Channel Apr. 1985 50“ 02-00' N 04“ 22-00' W 10 SEM , LM, Not figured Change observed Parke & Adam s, 1960 : TEM in culture Rowson, 1986; Manton & Leedale, 1963 (PLY 128) Cd. leptoporus/Cr. rigidus Snellius II W Mediterranean Jul. 1985 GX-192 36“ 54-00' N 02“ 11-30' E 5 SEM Not figured Hol-het Intermediate morphotype Kleijne, 1991 - Plate 4, Fig. 4 Cd. leptoporus/Cr. rigidus W C Atlantic Ocean M ay 1991 32“ 10-00' N 64“ 30-00' W 25 SEM Not figured Hol-het Intermediate morphotype Cortés, 2000 - Plate 1, Figs 1, 2 Cd. leptoporus/Cr. rigidus WC Atlantic Ocean M ay 1991 32“ 10-00' N 64“ 30 00' W 1 SEM Not figured Hol-het Unknown morphotype Renaud & Klaas (2001) Cd. leptoporus/Cr. rigidus WC Atlantic Ocean May 1991 32“ 1000' N 64“ 30-00' W 25 SEM Not figured Hol-het Intermediate morphotype Renaud & Klaas (2001) Cd. leptoporus/Cr. rigidus WC Atlantic Ocean May 1991 32“ 10-00' N 64“ 30-00' W 25 SEM Not figured H ol-het Intermediate morphotype (Tortés, 2000 - Plate 1, Figs 3, 4 Cd. leptoporus/Cr. rigidus MATER II W. Mediterranean Sept.-O ct. 1999 LM Not figured Het-hol Intermediate morphotype, This publ. phase change observed in culture (AS 31) s PO

Table 1 (contd.) Cd. teploporus/Cr. rigidus MARA S. A tlantic Ocean LM Not figured Intermediate morphotype, 1 his publ. phase change observed in culture (NSlO-2, NS4-2, NS8-2) Cd. leptoporus/ M A T E R II W. M editerranean Oct. 1999 69 39° 25-98' N 02° 25-30' W 5 SEM 21, 22 Hol het Large morphotype MG 124-04 to 13 This publ. St. quadriperforalus Ss. putchra/D. pirus MATER II W. Mediterranean Sept. 1999 15 35° 55-20' N 01° 20-73' W 5 LM 30 Unknow n This publ. Ss. putchra/D. pirus M A T E R II W. M editerranean Sept. 1999 15 35° 55-20' N 01° 20-73' W 5 SEM 33, 34 Unknow n M G 117-67 to 69 This publ. Ss. putchra/D. pirus M A T E R II W. M editerranean Sept. 1999 15 35° 55-20' N 01° 20 73' W 5 SEM Not figured Hol-het This publ. Ss. putchra/D. pirus M A T E R II W. M editerranean Oct. 1999 69 37° 25-98' N 00° 25-30' W 42-5 SEM 31 H ol-het MGI28-4 to 6 This publ. Ss. putchra/D. pirus MATER II W. Mediterranean Oct. 1999 69 37° 25-98' N 00° 25-30' W 42-5 SEM 32 ?hol-het MG 127-26, 27 This publ. Ss. putchra/D. pirus C. Mediterranean Mar. 1961 9 37° 47-00' N 11° 23-00' E 100 LM, TEM Not figured Het hol Saugestad & Heimdal, 2002 - Plate 4, Figs I (a-c) Ss. putchra/D. pirus C. Mediterranean M ar. 1961 9 37° 47-00' N 11° 23-00' E 50 LM Not figured Het-hol Saugestad & Heimdal, 2002 - Plate 4, f-'igs 2 (a-c) Ss. putchra/D. pirus Mediterranean Unknown N ot figured Unknown 1 observation, no photo Lecal-Schlauder, 1961 Ss. putchra/D. pirus C. Mediterranean, 2000 Surface SEM Not figured Het-hol Change observed This publ. o ff Naples in culture (NAP-10) Ss. putchra/Ca. oblonga JG O F S 4 N. A tlantic Jun. 1990 2 53° 30-00' N 20° 30-00' W 30 SEM 26 Unknow n M G 130-6, 7 This publ. Ss. putchra/Ca. oblonga J G O F S 4 N. Atlantic Jun. 1990 2 53° 30-00' N 20° 30-00' W 30 SEM 27 U nknow n M G 130-8 to 10 "Htis publ. Ss. putchra/Ca. oblonga M ED EA -98 N W M editerranean M ar. 1998 41° 28-00' N 02° 19 10 E Surface LM , SEM Not figured Unknown C ros ei at. 2000 - Plate 2, Figs 3, 4 Ss. putchra/Ca. oblonga FR O N TS-96 N W M editerranean Sept. 1996 21 41° 11-70' N 03° 41-60' E 20 SEM Not figured ?hol-het C ros ei al. 2000 - Plate 2, Fig. 2 Ss. putchra/Ca. oblonga C. M editerranean 71902 U nknown LM Not figured U nknow n 2 observations, drawing L ohm ann, 1902 - Plate 6, Fig. 67, Plate 5, Ftg. 54 Ss. putchra/Ca. oblonga A driatic Sea 1926 U nknown LM Not figured Unknown Some observations on K am ptner, 1941 t living cells Ss. putchra/Ca. oblonga M editerranean 1954 25 LM Not figured Unknown Lecal-Sehlauder, 1961 - Photo 2, 3 Cs. m ediierranea/ M ESO-96 NW Mediterranean Ju n .-Ju l. 1998 12 41° 13-90' N 02° 20-70' E 40 SEM N ot figured C ros ei at. 2000 - Plate 4, Fig. 3 Cy. wettsteinii Cs. m ediierranea/ C. Mediterranean, 71939 Unknown LM Not figured Hol-het 2 observations, drawing K am ptner, 1941 - Plate 15, Fig. 152 Cy. wellsleinii near Rovigno Cs. m ediierranea/ H O T S C. N orth Pacific late 1996 22° 45-00' N 158° 00-00' W 5 SEM N ot figured Unknown Cortes & Bollmann, 2002 - Ftgs 1, 2 Cp. hasleana Cs. m ediierranea/ S. Atlantic Ocean 2000 Surface SEM , LM 38, 39 H et-hol Change observed M G 163-49, 50 T his publ. in culture (NS 8-5)

Ca., Calyptosphaera\ Cd., Calcidiscus; Cl., Coccolithus; Cp., Calyptrolithophora; Cr., Crystallolithus; Cy., Calyptrolithina; D., Daktylethra; H., Helicosphaera; SI., Syracolithus; Ss., Syracosphaera; Z., Zygosphaera; LM, light microscopy; SEM, scanning electron microscopy; TEM, transmission electron microscopy. 3 R esults 109

M. Geisen et al. 536 stituent crystallites of the holococcoliths using cross- were added under a laminar flow cabinet. For the polarized illumination (see Moshkovitz & Osmond, 1989, detailed chemical composition of K medium refer to or Young, 1992, for a description of this technique as Keller et al. (1987). applied to coccoliths). Other specimens were obtained from field samples Results and discussion collected during various cruises in the Mediterranean, the North Atlantic Ocean and the Gulf of Mexico (Table 1). We have additional data for each of three het- hol-hol associations identified by Cros et al. (2000), and evidence of three further examples. For clarity Culture samples the previous observations are briefly summarized A large collection of unialgal coccolithophorid strains here together with our new results. For all discussed has been established during the CODENET project, examples of combination coccospheres refer to including many species not previously cultured success­ Table 1 for sample details. fully. Cultures discussed here were isolated from plankton net samples collected off Arcachon (SW France), off South Africa (S. Atlantic), in the Alboran Sea (western Helicosphaera carteri with Syracolithus catilliferus Mediterranean) and off Naples (Italy) (see also Table 1). and S. confusus{Figs 2-6) Cros et al. (2000) illustrated two examples of the Seawater collection and culture isolation heterococcolithophorid Helicosphaera carteri (Fig. 2) forming het-hol combination coccospheres with To obtain concentrated seawater for isolation of Syracolithus catilliferus (Fig. 3). In addition they coccolithophorids small hand-operated plankton nets observed a single holococcolith-holococcolith with a mesh size o f 5 and 10 pm were deployed from ships on station at depths between 5 to 15 m and were left in (hol-hol) combination coccosphere with S. confusus the water for up to 2 h. Additionally the nets were used and S. catilliferus. Finally they reinterpreted a light to concentrate water from the CTD rosette sampler to micrograph of Lecal-Schlauder (1961 ) as illustrating sample species from the deep photic zone. a combination coccosphere of H. carteri with After collection the concentrated seawater samples Syracolithus confusus. were filtered through a 64 pm mesh sieve to remove larger We have found two further examples of S. zooplankton and transferred into translucent storage confusus with S. catilliferus (hol-hol) combination containers. Usually two containers were used per sample and GeOg (to inhibit the reproduction of diatoms) and coccospheres (Figs 4-6). Unlike the examples nutrients were added to one aliquot. The containers were shown in Cros et al. (2000), these specimens clearly stored at the respective ambient water temperature either contain transitional morphotypes between these in an incubator with a 16 h light, 8 h dark (16L/8D) cycle two (morpho-)species. or in a room with continuous light, and were opened daily If hybridization or sexual dimorphism were re­ to allow air exchange. Samples were transported back to sponsible for this het-hol-hol association we would the laboratory in a cool box as soon as possible. expect to observe both hol-hol combination Culture isolation was performed on an inverted micro­ scope using X 80 magnification and a glass micropipette. coccospheres with discrete holococcolith types, and Single cells were captured, transferred into fresh medium, het-hol-hol combination coccospheres, again with picked up again and finally transferred into sterile discrete holococcolith types (the latter being a more polystyrene tissue culture microplates with the wells filled advanced stage in the process of transition from the with a media series ranging from K/2 to K/10 (Keller haploid to the diploid phase). However, no het- et al., 1987). Normally a microplate with 24 wells, each hol-hol combination coccospheres have been dis­ filled with 2 ml medium, was used. After completion covered and the two additional examples which the lid of the microplate was sealed with Parafilm to have been found of hol-hol combination cocco­ prevent evaporation and the microplate was stored in an incubator. The microplates were checked regularly spheres include intermediate stages between the two and growing cultures were then transferred into sterile morphologies (Figs 4-6). Hybridization and sexual 75 ml tissue culture flasks filled with 40 ml of medium. dimorphism therefore seem very unlikely as causes of this association. The similarity of S. catilliferus and S. confusus was noted by Kleijne (1991), and Culture maintenance Cros et al. (2000) predicted that this may be an Cultures were maintained in exponential growth in an example of intraspecific variation in the degree of incubator at 17°C on a 16L/8D cycle. Typically, the calcification, the two holococcolith types differing cultures were checked with an inverted microscope every essentially in the presence/absence of perforations 2 weeks and reinoculated into fresh medium using a in the coccolith structure. The additional obser­ laminar flow cabinet to prevent contamination. Medium vations of hol-hol combination coccospheres with was prepared from seawater collected from the French intermediate morphotypes validate this prediction. coast of the English Channel. The seawater was filtered with an ordinary filter paper circle and autoclaved at Hence it seems very likely that this is an example of 120 °C for 15 min. After cooling, nutrients - nitrate fine-scale, non-genotypic, intraspecific variability in (500//M), phosphate (20//M), trace metals and vitamins- the morphology of the holococcolith-bearing phase 3 Results no

Life-cycle associations involving pairs o f holococcoiithophorid species 537

1

«m

1

. I • . ' *'

Figs 2-13. Scanning electron micrographs (SEMs) of coccospheres of various species. Fig. 2. SEM of a coccosphere of Helicosphaera carteri. Water sample, N. Atlantic, R/V Meteor 42-4B cruise, station USIB. Fig. 3. SEM of a coccosphere of Syracolithus catilliferus. Water sample, N. Atlantic, off Canary Islands, R/V Poseidon cruise 233, station 3. Image courtesy C. Sprengel, AWI Bremen. Figs 4-6. SEMs of two Syracolithus confusus-SyracoUthus catilliferus combination coccospheres. All the figures show the S. catilliferus and .S', cottfusus coccoliths (arrows) as well as transitional forms on one coccosphere. Fig. 5 shows a detail of Fig. 4. Water sample, NW Mediterranean, MESO-96 cruise, station F2. Fig. 7. SEM of a collapsed Syracosphaera delicatus coccosphere. Both endo- and exothecal coccoliths can be seen. Water sample, western Mediterranean, Alboran Sea, R/V IIe.sperides cruise MATER 2, station 69, Figs 8-10. SEMs of two Zygosphaera hannockii Corisphaera sp. type A combination coccospheres. Fig. 9 shows a detail of Fig. 8. Both coccospheres show 3 R esults 111

M. Geisen et al. 538 and all three inorphospecies should be referred to as described, while the genus Syracosphaera has pri­ a single species. ority over Zygosphaera. By analogy to H. carteri an informal classification should be used to distinguish Helicosphaera carteri {Wallich, 1877) Kamptner, the respective intraspecific holococcolith morpho­ 1954. logies: S. bannockii HO-b ridged and S. bannockii H eterotypic sy n o n y m s : Syracolithus catilliferus HO-solid. (Kamptner, 1937) Deflandre, 1952; Syracolithus confusus Kleijne, 1991. As discussed in Cros et al. (2000), Helicosphaera Coccolithus pelagicus with Crystallolithus carteri has priority and so is the appropriate name. hyalinus and Crystallolithus braarudii {Figs 11-15) N.B. Syracolithus Deflandre 1952 would have pri­ Combination coccospheres of Coccolithus pelagicus ority over Helicosphaera Kamptner, 1954, but the and Crystallolithus hyalinus have been illustrated type species of Syracolithus is S. dalmaticus, which is from field-collected samples (Samtleben & Bickert, not known to form associations with Helicosphaera. 1990; Samtleben & Schroder, 1992; Baumann etal., As proposed in Cros et al. (2000) informal terms 1997; C. Samtleben personal communication). Un­ {H. carteri HO-solid for catilliferus type and H. published micrographs of specimens from field- carteri HO-perforate for confusus type holococco­ collected samples from the N. Atlantic Ocean were liths) should be used to distinguish the respective made available to us by Christian Samtleben (Uni­ intraspecific holococcolith morphologies. versity of Kiel) and Karl-Heinz Baumann (Uni­ versity of Bremen) (Figs 13, 14); like the published Syracosphaera bannockii with Zygosphaera micrographs, these reveal unambiguous com­ bannockii and Corisphaera sp. type A. {Figs 7-10) bination coccospheres of C. pelagicus with Cr. The holococcoiithophorid Zygosphaera bannockii hyalinus. was observed by Cros et al. (2000) to form both Several of our monoclonal cultures of C.pelagicus het-hol combination coccospheres with a from Arcachon (SW France) have given rise to the previously undescribed Syracosphaera (Fig. 7) holococcolith-bearing phase, but in each case the species and hol-hol combination coccospheres with holococcoiithophorid associated is Crystallolithus Corisphaera sp. type A. (N.B. Since the braarudii (Fig. 15), rather than Cr. hyalinus. Syracosphaera species was previously undescribed The two holococcolith types involved, Crystallo­ the name S. bannockii is now applied to it, as lithus braarudii and Crystallolithus hyalinus, are recommended by Cros et ai, 2000). structurally very similar, hence Cros et al. (2000) We have found three further examples of the concluded that the Coccolithus pelagicus- hol-hol combination of Z. bannockii with Crystallolithus combination was another example Corisphaera sp. type A (Figs 8-10). In all cases these of variation in the degree of calcification. In this specimens contain transitional morphotypes case, however, the two holococcolith morphotypes between these two (morpho-)species. have not been observed co-occurring on a single This example is directly analogous to the H. coccosphere, and the holococcolith morphology carteri case and so is also interpreted as a result of appears to be consistent within monoclonal cul­ non-genotypic, intraspecific variation. tures maintained under a range of environmental conditions. Syracosphaera bannockii {Borsetti et Cati, 1976) A review of the literature reveals that Cr. hyalinus Cros et al. 2000. and Cr. braarudii have often been confused. Parke H om o typic s y n o n y m : Zygosphaera bannockii & Adams (1960), who first demonstrated the as­ (Borsetti et Cati, 1976) Heimdal, 1982. sociation in cultures, identified the holococcolith- H eterotypic sy n o n y m : Corisphaera sp. type A bearing stage as Cr. hyalinus and Rowson et al. Kleijne, 1991. (1986) maintained this identification. In the original Syracosphaera bannockii was proposed by Cros et description (Gaarder & Markali, 1956) the central al. (2000) as a new combination, since the hetero­ area of Cr. hyalinus is described as being filled with coccolith morphotype had not previously been calcite rhombohedra arranged in parallel rows, with coccoliths o f Z. bannockii and Corisphaera sp. type A (arrows) as well as transitional forms. Water sample, NW Mediterranean, FANS 1 cruise, station 127 (Figs 8, 9) and station 123 (Fig. 10). Figs 11,12. SEMs of coccospheres of Coccolithus pelagicus. The images display the coccolith size variation between the large temperate (Fig. 11) C. pelagicus and the small Arctic morphotype (Fig. 12). Water sample, S. Atlantic, off Namibia, R/V M eteor cruise M48-4 (Fig. 11) and N. Atlantic, off Iceland (Fig. 12). Fig. 13. SEM of Coccolithus pelagicus-Crystallolithus hyalinus combination coccosphere. The C. pelagicus heterococcolith is of the small morphotype. The Cr. hyalinus coccolith shows its typical central area features, with the calcite rhombohedra arranged in parallel rows and covering all the central area. Water sample, N. Atlantic, Greenland Sea. Image courtesy C. Samtleben, University of Kiel. Scale bars represent: Figs 2, 8-10: 1 fim; Figs 3-7, 11-14: 2 //m. 3 R esults 112

Life-cycle associations involving pairs o f holococcoiithophorid species 539 each crystal lying on one face and partly touching Coccolithus pelagicus has priority over the the surrounding crystals at parts of the adjacent heterotypic synonyms. Since the type of Crystallo­ faces (compare Figs 13, 14). Two years after the lithus Gaarder & Markali 1956 is C. hyalinus this publication of the C. pelagicus-Cr. hyalinus life­ genus is a junior synonym of Coccolithus Schiller cycle by Parke & Adams (1960), Gaarder (1962) 1930. This work indicates C. pelagicus consists of described the new holococcoiithophorid Cr. two different biological taxa, so it is suggested that braarudii. Whereas the rim structure in this species C. pelagicus subsp. pelagicus and C. pelagicus subsp. is similar to Cr. hyalinus, the basal layer is in­ braarudii are used for the respective subspecies. Re­ complete with the crystallites being confined to a examination of the type material collected by few radial spokes and sometimes a central ellipse Wallich on the Bulldog cruise indicates that the (compare Fig. 15). The specimens figured in both smaller, Arctic heterococcolith morphotype is the Parke & Adams (1960) and Rowson et al. (1986) type form. Hence if the forms are differentiated as clearly resemble Cr. braarudii rather than Cr. subspecies this form must bear the name C. pelagicus hyalinus. Several of our cultures of C.pelagicus subsp. pelagicus, which is an autonym and so does (strains L K l, 2 & 3, CF4 & 5, all from Arcachon, not need to be formally proposed (ICBN Art 26.3). SW France) have undergone phase change and in each case examination with transmission electron Coccolithus pelagicus subsp. braarudii {Gaarder, microscopy and light microscopy revealed the holo­ 1962) Geisen et al., comb. & stat nov. coccolith Cr. braarudii (Fig. 15). All observations B a sio n y m : Crystallolithus braarudii Gaarder, 1962 made so far from culture material thus seem to {Nytt. Mag. Bott., 10, p. 43, pi. 7). display a C. pelagicus-Cr. braarudii life-cycle. By contrast, the C. pelagicus-Cr. hyalinus com­ Calcidiscus leptoporus with Crystallolithus rigidus binations, figured in Samtleben & Bickert (1989), and Syracolithus quadriperforatus {Figs 16-22) Samtleben & Schroder (1992), Winter & Siesser (1994) and Baumann et al. (1997), have only been Calcidiscus leptoporus (Figs 16-18) has previously observed from plankton samples. There is also a been shown to be associated with the holococco- biogeographic division between these associations, lithophorid Crystallolithus rigidus (Fig. 19) (Kleijne, all plankton observations coming from Arctic 1991), an association that has been confirmed by waters whilst the cultures in which phase trans­ observations from the Bermuda area (Cortés, 2000; formations have been observed have all been Renaud & Klaas, 2001). Recently this observation isolated from temperate waters. It has recently been has been proven by a partial transition in four of shown that the temperate and Arctic C. pelagicus our cultures ofCalcidiscus (AS 31, Alboran Sea, populations show different ecological adaptations, western Mediterranean and NSlO-2, NS4-2, NS8-2, produce different-sized heterococcoliths (Figs 11, S. Atlantic, off South Africa) which have given rise 12), and are genetically differentiated (Baumann et to holococcolithophorids bearing C. rigidus cocco­ a i, 2000; Cachao & Moita, 2000; our unpublished liths. We found a single combination cell of the data). It seems likely that these populations rep­ heterococcolith C. leptoporus with the holococco- resent discrete species, or subspecies. Since in this lithophorid Syracolithus quadriperforatus (Figs 21, case the holococcolith differentiation appears to 22) at MATER cruise station 69. Although we have parallel that of the heterococcoliths, there is no only this single coccosphere as evidence of the new support for inferences such as complex life-cycles, association, it is an exceptionally clear specimen hybridization or sexual dimorphism, and intra­ with a uniquely well-preserved outer cover of specific variation seems unlikely. Instead, it seems holococcoliths. In our view it is highly unlikely that that a recent phylogenetic divergence event has this specimen could be any form of artefact. occurred with slight qualitative separation of the The holococcolithophorids involved, Crystallo­ holococcoliths and quantitative, biometrically lithus rigidus (Fig. 19) and Syracolithus quadri­ measurable, separation of the heterococcoliths (par­ perforatus (Fig. 20), have coccoliths with very allel differentiation). We recommend distinguishing different morphologies and structures. The the morphotypes as subspecies rather than species, coccoliths of Cr. rigidus are essentially plate-like, due to the slight morphological differentiation and consisting of two layers of crystallites in a hexagonal to minimize nomenclatural confusion. array surrounded by a rim three crystallites high. S. quadriperforatus coccoliths, by contrast, have a high tube with internal walls which define four to six Coccolithus pelagicus {Wallich, 1877) Schiller, 1930 openings (compare Figs 19 and 20). On the proximal {type species o f Coccolithus). surface there are two or three concentric rings of H eterotypic sy n o n y m s : Crystallolithus braarudii crystallites and a large central opening usually Gaarder, 1962; Crystallolithus hyalinus Gaarder et covered by an organic membrane. These two Markali, 1956 (type species of Crystallolithus). structures are very different and close affinity 3 Results 113

M . Geisen et al. 540

%

: s w « «1- • ' I

r*

S' Figs 14-25. Electron micrographs of coccospheres and coccoliths of various species. Fig. 14. SEM of Coccolithus 3 R esults 114

Life-cycle associations involving pairs o f holococcoiithophorid species 541 between them has never been predicted. Crystallo- morphologies have been observed between the two graphically there is more affinity between these holococcoiithophorid species, intraspecific vari­ structures; the internal walls of S. quadriperforatus ation can be ruled out in this case. As with C. pel­ and the hexagonal meshwork plate of Cr. rigidus are agicus, there seems to be strong evidence of phylo­ both formed of calcite crystallites with vertical c- genetic differentiation of biological (sub-)species axes whilst the tube and rim are formed of that show slightly different morphologies in the crystallites with radial c-axes. heterococcolithophorid phase, each associated with C. leptoporus heterococcoliths show considerable a different holococcoiithophorid stage. variation in size and in certain elements of their morphology. Kleijne (1991) and Knappertsbusch et al. (1997) distinguished three morphotypes: (1) Calcidiscus leptoporus (Murray et Blackman, 1898) small morphotype (Fig. 16) - liths 3 -5 //m, 10-20 Loeblich et Tappan, 1978. elements, distal shield sutures often angular and H eterotypic sy n o n y m s : Crystallolithus rigidus serrated (Kleijne, 1991), sometimes the inner part of Gaarder in Heimdal et Gaarder, 1980; Syracolithus the distal shield elements shows a dextral inclination quadriperforatus (Kamptner, 1937) Gaarder in (our observations); (2) intermediate morphotype Heimdal et Gaarder, 1980. (Fig. 17) - liths 5-8 yum, 15-30 elements, sutures Calcidiscus leptoporus has priority, but as C. variable ; (3) large morphotype (Fig. 18)-liths leptoporus seems to incorporate three biological 7-11 yum, 20-35 elements, sutures smoothly curved, subspecies it is appropriate to introduce subspecies. usually with a zone of obscured sutures around the The size range of coccoliths pictured in the type crest of the tube (Baumann, personal communi­ description of C. leptoporus is that of the inter­ cation; our observations). These morphotypes, par­ mediate morphotype, consequently the name C. ticularly the large and intermediate forms, seem to leptoporus subsp. leptoporus should be applied to intergrade in morphology and there is no simple this form. (G. Murray worked at the Natural pattern to their biogeography (Renaud & Klaas, History Museum London, but we have been unable 2001), but morphometric studies have consistently to locate any coccolith preparations of his, and it supported their discrimination (Kleijne, 1991; seems likely that he used water mounts. Fixed Knappertsbusch et a i, 1997; Baumann & Sprengel, samples of his do exist but these have decalcified. 2000; Renaud & Klaas, 2001). All het-hol Hence the type illustrations are the only available associations involving Cr. rigidus are with the evidence.) C. leptoporus subsp. leptoporus is an intermediate-size C. leptoporus morphotype autonym and so does not need to be formally (Cortés, 2000; our culture observations). The proposed (ICBN Art 26.3). heterococcoliths of our new combination specimen The combination with the holococcolith S. quadri­ (C. leptoporus with S. quadriperforatus) measure perforatus bears heterococcoliths of the large 6 7-8 3 yum, on the borderline between intermediate morphotype so the name C. leptoporus subsp. and large morphotypes, but the central area quadriperforatus should be used. As there are no characters indicate that it is the large morphotype. observations of holococcoliths being associated On the basis of the different morphologies of the with the small morphotype of C. leptoporus it is holococcoliths involved in the mentioned cases and suggested here that an informal classification C. on the fact that no intermediate holococcolith leptoporus subsp. SMALL be used, pending identi- pelagicus-Cryslallolithus hyalinus combination coccosphere. The C. pelagicus heterococcolith is of the small morphotype. The Cr. hyalinus coccolith shows its typical central area features, with the calcite rhombohedra arranged in parallel rows and covering all the central area. Water sample, N. Atlantic, Greenland Sea. Image courtesy of C. Samtleben, University of Kiel. Fig. 15. Transmission electron micrograph of Crystallolithus braarudii coccoliths. The Cr. braarudii coccoliths show the typical central ellipse with radial spokes connecting to the rim. Culture material, SW France, off Arcachon. Figs 16-18. SEM of coccospheres of Calcidiscus leptoporus. The images display the coccolith and coccosphere size variation between the small (Fig. 16), intermediate (Fig. 17) and large (Fig. 18) morphotype. Water samples, S. Atlantic, off Namibia, R/V M eteor cruise M48-4, station 20 (Figs 16, 17) and Western Pacific Ocean, Miyake-jima island, Japan (Fig. 18). Fig. 19. SEM of a coccosphere of Crystallolithus rigidus. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 59. Fig. 20. SEM of a coccosphere of Syracolithus quadriperforatus. Water sample, N. Atlantic, off Canary Islands, R/V Poseidon cruise 233, station 2. Image courtesy of C. Sprengel, University of Bremen. Figs 21, 22. SEM of a Calcidiscus leptoporus-Syracolithus quadriperforatus combination. Fig. 22 shows a detail of Fig. 21. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 23. SEM of a coccosphere of Syracosphaera pulchra. Both endo- and exothecal coccoliths can be seen. Water sample, N. Atlantic, off Canary Islands, R /V Poseidon cruise 233, station 3. Image courtesy of C. Sprengel, AWI Bremen. Figs 24, 25. SEMs of Calyptrosphaera oblonga. Fig. 24 shows a collapsed coccosphere and Fig. 25 shows a detail of the circumflagellar coccoliths. Note the typical hexagonal structure of the calcite rhombohedra and the absence of an offset between base and hood. The circiunflagellar coccoliths often have a pointed hood. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 15 (Fig. 24) and station 69 (Fig. 25). Scale bars represent: Fig. 22: \ pm \ Figs 14-21, 23-25: 2 jum. 3 Results 115

M. Geisen et al. 542

■ ' ' '

^ ' - r " ' "

m .

%

(tùkb.

t a - ' -ÎOT Figs. 26-41. SEMs of various coccolithophorids. Figs 26, 27. SEM of .S. pulchra-C. ohlonga combination coccospheres. Water sample, N. Atlantic, JG OFS leg 4 (1990) cruise. Figs 28, 29. Scanning electron micrograph of Daktylethra pirus. 3 R esults 116

Life-cycle associations involving pairs o f holococcoiithophorid species 543 fication of the holococcolith phase which will types associated with Syracosphaera pulchra - determine the correct subspecies to be used. Calyptrosphaera oblonga and Daktylethra pirus (Figs 23-25, 28, 29) - have been placed in different Calcidiscus leptoporus subsp. quadriperforatus genera. They are, however, rather close in mor­ {Kamptner, 1937) Geisen et al., comb. & stat nov. phology, so the similarities and the differences in their morphology require some discussion: B asionym : Syracosphaera quadriperforata Kamptner, 1937 {Arch. Protistenk., 89, p. 302, pi. Similarities : (1) Both are cavate holococcoliths 15, figs. 15, 16). consisting of a single-layered tube and convex distal cover; (2) LM observations indicate that all crystallites are arranged with their e-axes perpen­ dicular to the surface of the coccolith ; (3) in both Syracosphaera pulchra with Calyptrosphaera cases the proximal surface consists of three or four oblonga and Daktylethra pirus {Figs 23, 34) concentric rings of crystallites, with a distinct central Cros et al. (2000) showed one unambiguous and one opening usually covered by an organic membrane, questionable combination coccosphere involving and the outermost ring protrudes beyond the tube the heteroeoeeolithophorid Syracosphaera pulchra to form a basal flange (Figs 25, 29). None of these (Fig. 24) and the holococcoiithophorid three features is uncommon for holococcoliths, but Calyptrosphaera oblonga (Figs 24, 25), confirming the co-occurrence in these two holococeolith types the previous observations of Lohmann (1902) and does suggest close affinity. Moreover, (4) in both Kamptner (1941). We have subsequently found two coecolith types eircumflagellar coccoliths have dis­ further examples of this association in a sample tinctive pyramidal bosses on the distal surfaee, a from the North Atlantic (Figs 26, 27). However, we feature not shown by any other holococcolith types have also observed several specimens from the (Figs 25, 28, 29). Alboran Sea where S. pulchra coccoliths are Differences; (1) In C. oblonga coccoliths the tube associated on combination eoeeospheres with the wall is initially vertieal and curves into the distal holococcoiithophorid Daktylethra pirus (Figs 28, cover with no obvious break, whereas in D. pirus 29). Four SEM and one LM specimen from two coccoliths the tube wall flares outward and there is stations have been observed (Figs 30-34). a major inflection between the tube and distal cover; One example of a S. pulchra-D. pirus con\h\m.\.\on (2) in D. pirus large pores are present around the (described as a S. pulchra-C. oblonga combination) distal cover of the coccolith (Fig. 29) ; (3) C. oblonga was recorded without illustration by Lecal- coccoliths have a perforated hexagonal crystallite Schlauder (1961), and several further examples have arrangement (Fig. 25), whereas D. pirus coccoliths been observed in a study of samples from the have a non-perforate crystallite arrangement with­ Tyrrhenian Sea (Saugestad, 1967; Saugestad & out obvious hexagonal pattern (Fig. 29). The aflSnity Heimdal, 2002). in eoccolith structure of these two species is clear Recently a phase change has occurred in one of and their placing within different genera is little our cultures ofS. pulchra (NAP-10 from offshore more than a historical accident. Nonetheless, the Naples, Italy). The resulting motile phase bears two morphologies are entirely discrete, being holococcoliths. These are often malformed, but the separated by multiple independent characters. better-formed specimens are unambiguously identi­ Moreover, although both species are very common, fiable as D. pirus with both LM and SEM. As in neither intermediate morphotypes nor co-occur­ the case of C. leptoporus, the two holococcolith rence of the two coccolith morphologies on a single

Fig. 28 shows a collapsed coccosphere. Note the pointed hood of the circumflagellar coccoliths. Fig. 29 shows a detail of D. pirus coccoliths. Note the clear offset between the hood and the base as well as the perforations in the hood. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Figs 30-34. Light micrographs and SEMs of Syracosphaera pulchra-Daktylethra pirus combination coccospheres. Fig. 30 displays a com bination cell observed with a light microscope. Left, cross-polarized light; right, phase contrast. Figs 31-34 display SEMs of three further combination coccospheres. Fig. 34 shows a detail of Fig. 33. Water samples, western Mediterranean, Alboran Sea, R /V Hesperides cruise M ATER 2, station 15 (Figs 30, 33, 34) and station 69 (Figs 31, 32). Fig. 35. SEM o f a coccosphere of Coronosphaera mediterranea. Water sample. South Atlantic, off Namibia, R/V Meteor cruise M48-4, station 476. Figs 36, 37. SEMs of Calytrolithophora hasleana. Fig. 36 shows a collapsed coccosphere and Fig. 37 shows a detailed view of the coccoliths. Note the hexagonal crystal arrangement of the distal cover. Water samples, western Pacific Ocean, Miyake-jima island, Japan. Figs 38-41. SEMs of Zygosphaera hellenica. Fig. 38 shows coccoliths of both Coronosphaera mediterranea and Z. hellenica in a culture in partial transition. Fig. 39 displays a detail of Fig. 38. Fig. 40 shows a coccosphere of Z. hellenica from a field sample. Note the difference in comparison with the culture material in Figs 38, 39. In Fig. 41 the arrangement of crystals in concentric rings and the ring of pores around the base of the coccoliths can be clearly seen (arrows). Culture material, S. Atlantic, off South Africa (Figs 38, 39) and water sample, western Mediterranean, Alboran Sea, R /V Hesperides cruise MATER 2, station 69. Scale bars represent: Figs 29, 37: 1 pm \ Figs 26-28, 31-34, 39-41 : 2/zm; Figs 30, 36: 5//m . 3 R esults 117

M. Geisen et al. 544

Coccolith length & width for S. pulchra combinations

*■ pirus m oblonga

6.5 -

6 -

I I 5.5

5 -

4.5 -

n = 44 (S. pulchra ' 0. pirus) n = 13 (S. pulchra I C. oblonga)

2.5 3.5 4 4.5 width in pm X------

Fig. 42. Length and width of endothecal heterococcoliths of Syracosphaera pulchra measured on scanning electron micrographs of seven combination cells. A total of 44 coccoliths on combinations with Daktylethra pirus and 13 on combinations with Calyptrosphaera oblonga were measured. All measurements occupy the same morphospace, with those taken on combination cells with C. oblonga showing slightly higher mean coccolith lengths and widths. The means and the standard deviation for the measurements on the two types of combination cells are displayed near the axis.

cell have ever been reported. The S. pulchra hetero­ pulchra in the sedimentary record will prove coccoliths on the cells involved in the combination difficult. coccospheres show normal morphologies, including In this case intraspecific variation can be ruled both endothecal and exothecal coccoliths. Although out as the morphology of the two holococcoliths S. pulchra coccoliths have unusually complex mor­ involved is significantly different and no cells have phologies, we have not been able to detect any been observed bearing both holococcolith types or consistent differences in morphology between the intermediate stages. By analogy to C. pelagicus and coccoliths occurring on combination coccospheres C. leptoporus, we believe the most likely cause of with C. oblonga and those occurring with D. pirus. this het-hol-hol association is genotypic differ­ Measurements of the length and width of all suitably entiation, but with only the holococcolith phase oriented endothecal coccoliths on the combination having changed morphologically, i.e. a case of coccospheres (Fig. 42) showed intriguingly that the cryptic spéciation in the heterococcolith phase. An C. oWo«ga-associated heterococcoliths had a alternative hypothesis of a complex life-cycle with narrower range of sizes and higher mean size than two holococcolith phases cannot be ruled out, how­ the D. pfrM5-associated heterococcoliths. However, ever. Obviously it will be interesting to study this there is complete overlap between the datasets, and case further, particularly with molecular genetics. we have too few observations to be able to conclude that these size variations will prove consistent. Syracosphaera pulchra Lohmann, 1902. Equally, size measurements ofS. pulchra coccoliths H eterotypic sy n o n y m s : Calyptrosphaera oblonga from Holocene sediments show a clear monomodal Lohmann, 1902 (type species of Calyptrosphaera distribution pattern (Fig. 43), indicating that a Lohmann, 1902); Daktylethra pirus (Kamptner, morphology-based species discrimination of S. 1937) Norris, 1985. 3 Results il8

Life-cycle associations involving pairs of holococcoiithophorid species 545

Contour lines 1 - max @ 2 intervals. produced by this culture are generally not well Contours < 6 are dashed. formed and tend to collapse into unidentifiable mounds of crystallites in SFM preparations. A 6 limited number of intact holococcoliths have, how­ pulchra ever, now been observed (Figs 38, 39) and these are unambiguous specimens of a third holococcolith 5 species, Zygosphaera hellenica (Figs 40, 41). Identi­ fication criteria include the arrangement of crystallites in numerous concentric rings, and the

4 presence of a ring of pores around the base of the coccolith . The available data in this case are limited but each combination appears very convincing and the fact that three holococcolith species are apparently Ê involved, rather than two, gives it particular interest. All three holococcolithophorids are di­ 2 morphic, and have similar-shaped coccoliths. In each case the body coccoliths are flat-topped tubes n = 1 7 4 with an irregular distal boss whilst the circum­ flagellar coccoliths have an elevated transverse 3. 3.4. 4.5 5. 5 6 . 6 . 7. 7. 8 . m ajor axis in jim bridge. They differ, however, in numerous other characteristics. C. wettsteinii coccoliths are cavate, Fig. 43. Density plot of Syracosphaera pulchra endothecal i.e. they have large central openings, the tube is non- coccoliths measured on 174 coccoliths from a number of perforate and the distal cover is broken by several Holocene samples. The plot shows the monomodal distribution of lengths and widths in the samples. large openings. C. hasleana coccoliths are also probably cavate but the tube and distal cover both have perforate hexagonal crystal arrangements and Neither 5". pulchra, the type species of Syracosphaera there are no large openings in the cover (Figs 36, Lohmann, 1902, nor C. ohlonga has clear priority 37). Z. hellenica coccoliths by contrast are non- since they were described in the same publication. cavate with usually the entire coccolith being filled On the grounds of nomenclatural stability Cros et by concentric layers of crystallites; the tube wall is al. (2000) have recommended use of S. pulchra. W e predominantly non-perforate but there is always a recommend here the use of 5. pulchra HE for the row of perforations around the base and variable heterococcolith phase (until future research allows numbers of perforations above this (Figs 38-41). separation of the heterococcolith subspecies within T he Z. hellenica coccoliths appear very different to S. pulchra) and the informal S. pulchra HO oblonga- the others; however, it is noticeable that some C. type and S. pulchra HO pirus-iyxsc for the respective hasleana coccoliths show partial development of the holococcolith phases. concentric layered structure and that some Z. hellenica coccoliths show perforate hexagonal wall structure. So the morphologies are perhaps less Coronosphaera mediterranea with different than they appear initially. Nonetheless the Calyptrolithophora hasleana, Calyptrolithina differences between these three holococcolith types wettsteinii ty/rc/ Zygosphaera hellenica {Figs 35-41) are sufficiently large and consistent to make it Coronosphaera mediterranea (Fig. 35) has pre­ unlikely that these morphotypes result from non- viously been shown to be associated with the genotypic variation. holococcoiithophorid Calyptrolithina wettsteinii By contrast there is no obvious differentiation (K am ptner, 1941 ; C ros et a i , 2000). Subsequently of the heterococcoliths of Coronosphaera mediter­ a single cell of C. mediterranea has been observed ranea. A s w ith S. pulchra these are morphologic­ with the holococcoiithophorid Calyptrolithophora ally complex coccoliths which appeared to define a hasleana (Figs 36, 37) in a field-collected sample very clear morphospecies. However, despite the from the Pacific (Cortés & Bollmann, 2002). large number of available morphological characters We have not found further examples of these we cannot find any distinctive features which sep­ associations in plankton samples, but we have arate the heterococcoliths associated with C. w e tt­ recently been successful in isolating a culture of C. stein ii from those associated with C. hasleana or mediterranea from a water sample collected in the Z. hellenica. We therefore conclude that this case South Atlantic (NS 8-5). This culture has sub­ is analogous to that of 5". pulchra, i.e. spéciation has sequently undergone a partial transition to the occurred but that this is only obviously reflected in holococcoiithophorid phase. The holococcoliths the morphology of the holococcolith phase, even 3 R esults 119

M. Geisen et al. 546 though in this case genotypic differentiation has Het-hol-hol associations thus seem to fall into occurred twice. three groups (Fig. 44) : (1) Helicosphaera carteri with Syracolithus Coronosphaera mediterranea {Lohmann, 1902) catilliferus and S. confusus', and Syracosphaera Gaarder in Gaarder et Heimdal, 1977. bannockii with Zygosphaera bannockii and Coris­ H eterotypic sy n o n y m s : Calyptrolithina wettsteinii phaera sp. A. In these cases the holococcolith (Kamptner, 1937) Kleijne, 1991 (type species of ‘ species ’ appear to be intraspecihe morphotypes, as Calyptrolithina Heimdal, 1982); Calyptrolithophora demonstrated by the oecurrence of intergradational hasleana (Gaarder, 1962) Heimdal, in Heimdal et morphotypes and co-occurrence of the two mor­ Gaarder, 1980 (type species of Calyptrolithophora photypes on single eoeeospheres. Holocoecolith Heimdal in Heimdal et Gaarder, 1980; Zygosphaera morphology thus appears to be more plastic hellenica Kamptner, 1937 (type speeies of than heterococcolith morphology, perhaps un­ Zygosphaera Kamptner, 1936 by subsequent des­ surprisingly given the relative large number and ignation of Loeblich & Tappan, 1963). simple arrangement of crystals and the observation The species Coronosphaera mediterranea, the type that holococcoliths are formed outside the cell species of Coronosphaera Gaarder in Gaarder et membrane. Heimdal, 1977, has priority over the three (2) Coccolithus pelagicus with Crystallolithus associated holococcolith species. Strictly, the genus hyalinus and Cr. braarudii', and Calcidiscus lepto­ Zygosphaera Kamptner 1937 has priority over porus with Crystallolithus rigidus and Syracolithus Coronosphaera Gaarder 1977. However, the genus quadriperforatus. In these cases qualitative differen­ Coronosphaera is much more widely used and better tiation in holocoecolith morphology is paralleled by established than the genus Zygosphaera and use of morphometric differentiation in the hetero- Coronosphaera would involve fewer new com­ coecoliths. The holococcolith differentiation thus binations. We are preparing a submission to the provides strong support for previous inferences of ICBN to conserve the name Coronosphaera and genotypic diversification. In the case of C. pelagicus, suppress the name Zygosphaera. Pending this ap­ there is clear evidence that the two subspecies peal we recommend use of the genus Coronosphaera. occupy different geographic ranges, suggesting that As this work suggests that the heterococcolith allopatric spéciation has occurred. For C. lepto­ phase of Coronosphaera mediterranea eonsists of porus, however, there is no evidence of present or three morphologically indistinguishable biological past spatial isolation of populations, suggesting that species or subspecies, we recommend the use of C. the inferred spéciation was sympatric, i.e. a result of mediterranea HE for the heterococeolith phase ecologieal niche separation within the same geo­ (until future research allows separation of the graphical zone. At present there are not enough heteroeoccolith speeies in C. mediterranea spp.) and data to determine whether the different C. lepto­ the informal names C. mediterranea HO wettsteinii- porus subspecies occupy distinct niches in con­ type, C. mediterranea HO hasleana-typc and C. temporary oceans. A detailed study of the seasonal mediterranea HO hellenica-type for the respective and depth distribution of these taxa in relation to holococcolith phases. variation in physico-chemical parameters would elearly be of interest. (3) Syracosphaera pulchra with Calyptrosphaera Conclusions oblonga and Daktylethra pirus', and Coronosphaera In total only about 20 het-hol associations have mediterranea with Calyptrolithina wettsteinii, been discovered. However, since these span the Calyptrolithophora hasleana and Zygosphaera evolutionary biodiversity of coccolithophorids we hellenica. In these cases the holococcolith differen­ predict that this will prove to be a common pattern, tiation provides compelling evidence for previously and that the infrequency of such observations may unsuspected cryptic spéciation within the hetero­ be a result of the temporally and spatially sporadic coceolith species. An interesting theoretical expla­ nature of most sampling and the fact that syngamy nation for this phenomenon can be postulated. and meiosis are likely to be rapid processes that Protection against the expression of deleterious occur infrequently in the natural environment. mutations is often cited as a potential advantage of Rather surprisingly, in six of these cases, i.e. nearly diploidy over haploidy (for a review see Valero et ai, a third of the total, the heterococcolithophorid 1992). It can be hypothesized that a corollary of this involved has been shown to form associations with is that in diploid cells potentially advantageous not one, but two or three holococcolithophorids. mutations are not necessarily expressed, depending Despite the limited number of het-hol-hol com­ on the relative dominance of alleles. Any gene binations observed, a close inspection of each case mutation will necessarily be expressed in a haploid allows certain conclusions on the possible causative eell, and even though the rate of evolution of genes faetors to be drawn. may not differ between the phases, one might expect 3 Results 120

Life-cycle associations involving pairs of holococcoiithophorid species 547

Fine scale spéciation

unknown

(Cr. hyalinus) C. pelagicus

(Cr. rigidus) (Cr. braarudii)

(S. quadrip erfo ratu s)

Intraspecific variation C, leptoporus (Corisphaera sp.) (S. catilliferus)

hi carteri S. bannockii (S. confusus) Cryptic speciation (Z. bannockii)

^ A (C. hasleana) (D. pirus)

S. pu lch ra C. mediterranea

(C. oblonga) (Z. hellenica)

Fig. 44. Summary of life-cycle associations of sets of holococcolithophorids with a single heterococcolithophorid. The lines represent observed combination coccospheres. intraspecific variation, with transitional morphotypes in the holococcoiithophorid phase, is a likely cause for both Ifelicosphaera carteri and Syracosphaera hannockii and fine-scale speciation is seen as the likely cause for both Coccolithus pelagicus and Calcidiscus leptoporus. Two or more discrete holococcolithophorids in combination with one heterococcolithophorid species as observed in Coronosphaera mediterranea and Syracosphaera pulchra makes cryptic speciation the likely cause. (See Table 1 for abbreviations.)

to preferentially observe the result of gene appears clear that holococcolith morphology is mutations which cause neutral or advantageous more readily variable than heterococolith mor­ changes in haploid cells. Following a speciation phology and is thus a more sensitive indicator of event, therefore, the rate of morphological evol­ fine-scale variation, but less useful for identifying ution in haploid cells would be predicted to be phylogenetic relationships. greater than that of diploid cells. In any case it Het-hol associations are known in five hetero- 3 Results 121

M. Geisen et al. 548

Table 2. Comparison of numbers of holococcoiithophorid species and of heterococcolithophorid species in families known to form associations

In N ot in Total combinations combinations

Holococcolithophorids 59 18 41 Heterococcoiithophorids Rhabdosphaeraceae 20 1 19 Syracosphaeraceae (including Coronosphaera and Calciosolenia, but not Alisphaera) 50 8 42 Coccolithaceae 10 4 6 Zygodiscaceae 8 1 7 Total heterococcoiithophorids 88 14 74

Numbers of taxa are based on the taxon list of Jordan et al. (1994) with the addition of undescribed taxa known to the authors. NB; The Papposphaeraceae and likely associated holococcoliths are not included here as they appear to form discrete consistently identifiable groups (14 heterococcolith species and 10 holococcolith species have been described and 5 combinations recognized). The Noelaerhabdaceae, Pleurochrysidaceae and Hymonomonadaceae are excluded since they are known to be non-calcifying in the haploid phase. coccolithophorid families (Cros et a i, 2000). Since Acknowledgements this association must be a primitive feature derived We are grateful to numerous colleagues for sharing from a common ancestor (it is highly unlikely that their observations and insights with us, in particular the complex calcification mode of holococcoliths Karl-Heinz Baumann, Jorg Bollmann, Mara evolved independently on more than one occasion), Cortés, Jacqueline Fresnel, Berit Heimdal, Michael it might be predicted that all members of these Knappertsbusch, Annelies Kleijne, Vita Pariente, families will ultimately be shown to have an Jeremy Rowson, Alberto Garcia-Saez, Katharina holococcoiithophorid phase. An obvious impli­ von Salis, Christian Samtleben and Claudia cation is that there is a shortage of holococcoiitho­ Sprengel. We are also grateful for assistance with phorid species; as shown in Table 2 approximately sampling on several cruises, and especially to the 59 holococcoiithophorid species are known but master and crew of the R/V Hesperides during the there are some 88 heterococcolithophorid species MATER II cruise. Additionally we would like to in the relevant families. When species in known thank Chris Maggs, Paul Kugrens and two anony­ associations are removed then the discrepancy be­ mous reviewers for their valuable comments on the comes stronger-4 1 holococcoiithophorid species manuscript. This publication is a contribution of versus 74 heterococcolithophorid species. We con­ the EU TMR network project CODENET which clude that het-hol-hol associations, which are an funded our collaboration and supported M.G. and increasingly commonly discovered phenomenon, IP. are the result of intraspecific variation in the degree of calcification of the holococcoiithophorid phase or result from non-cryptic or cryptic speciation. This suggests that this discrepancy will widen References further as more het-hol-hol associations are dis­ A n d e r s e n , R.A., B r e t t , R.W., P o t t e r , D. & Sexton, J.P. (1998). covered. From this shortfall it might be inferred Phylogeny of the Eustigmatophyceae based upon 188 rDNA, that at least 30-50 % of the heterococcolith-forming with emphasis on Nannochloropsis. Prolist, 149; 61-74. B a u m a n n , K.-H ., A ndruleit, H., Schroder-Ritzrau, A. & species in these families have secondarily lost the Sam tleben, C. (1997). Spatial and temporal dynamics of ability to calcify in the haploid phase, or have be­ coccolithophore communities during non-production phases in come asexual. However, Cros (2001) has illustrated the Norwegian-Greenland Sea. In Contributions to the Micro- a large number of undescribed, rare holococcolith paleontology and Paleoceanology o f the Northern North Atlantic, vol. 5 (Hass, H.C. & Kaminski, M.A., editors), 227-243. morphotypes in field samples and we suspect that Grzybowski Foundation. the most likely explanation is that many species B a u m a n n , K.-H ., Y o u n g , J R., C a c h a o , M. & Z i v e r i, P. (2000). have only short-lived holococcolith phases. It Biometric study of Coccolithus pelagicus and its palaeoen- should be noted also that the reverse case, one vironmental utility. Abstracts of the 8th INA Conference, holococcocolithophorid being associated with two Bremen. Newsl. Int. Nannoplankton Assoc., 22: 82. B a u m a n n , K.-H. & S p r e n g e l , C. (2000). Morphological variations heterococcoiithophorids, cannot be ruled out. of selected coccolith species in a sediment trap north of the Finally we note that the production of different Canary Islands. /. Nannoplankton Res., 22: 185-193. biomineralized periplasts rich in phylogenetic data B i l l a r d , C. (1994). Life cycles. In The Haptophyte Algae (Green, within two phases of the life-cycle of cocco­ J.C. & Leadbeater, B.S.C., editors), 167-186. Systematics As­ sociation Special Volume 51. lithophorids gives the group special potential for B o w n , P R. (1998). Calcareous Nannofossil Biostratigraphy. Chap­ studies of m icroevoluti onary pattern and process. man and Hall, London. 3 R esults 122

Life-cycle associations involving pairs o f holococcoiithophorid species 549

Cachao, M. & M oita, M.T. (2000). Coccolithus pelagicus, a K e l l e r , M .D ., S e l v i n , R.C., C l a u s , W. & G uillard, R.R.L. productivity proxy related to moderate fronts off Western Iberia. (1987). Media for the culture of oceanic ultraphytoplankton. Mar. Micropaleontol., 39: 131-155. J. Phycol., 23: 633-637. C o r t é s , M.Y. (2000). Further evidence for the K l e i j n e , A. (1991). Holococcolithophorids from the Indian Ocean, heterococcolith-holococcolith combination Calcidiscus Red Sea, Mediterranean Sea and North Atlantic Ocean. Mar. leptoporus-Crystallolithus rigidus. Mar. Micropaleontol., 39: Micropaleontol., 17: 1-76. 35-37. K l e i j n e , A. (1993). Morphology, taxonomy and distribution o f extant C o r t é s , M.Y. & B o l l m a n n , J. (2002). A new combination coccolithophorids {calcareous nannoplankton). Ph.D. thesis. Free coccosphere of the heterococcolith species Coronosphaera University Amsterdam, Drukkerij FEBO, Enschede, The mediterranea and the holococcolith species Calyptrolithophora Netherlands. hasleana. Eur. J. Phycol., 37: 145-146. Knappertsbusch, M., Cortés, M .Y. & Thœrstein, H R. (1997). CossoN, J. (1987). Croissance des sporophytes résultant Morphologic variability of the coccolithophorid Calcidiscus leptoporus in the plankton, surface sediments and from the Early d’hybridations interspécifiques et intergénériques chez les Pleistocene. Mar. Micropaleontol., 30: 293-317. Laminariales. Cryptogamie Algol., 8: 61-72. Lancelot, C. & Rousseau,V. (1994). Ecology of Phaeocystis: the CossoN, J., G a y r a l , P. & O l i v a r i , R. (1984). Sur des échantillons key role o f colony forms. In The Haptophyte Algae (Green, J.C. & de Laminaria digitata à Stipe bifide : une hypothèse relative à leur Leadbeater, B.S.C., editors), 229-245. Systematics Association origine. Crytogamie Algol., 5: 15-20. Special Volume 51. C r o s , L. (2001). Planktic coccolithophores o f the N W Mediterranean. Leadbeater, B.S.C. (1994). Cell coverings. In The Haptophyte Ph.D. thesis. University of Barcelona. Algae (Green, J.C. & Leadbeater, B.S.C., editors), 23-46. C r o s , L., K l e i j n e , A., Z e l t n e r , A., B i l l a r d , C . & Y o u n g , J R. Systematics Association Special Volume, 51. (2000). New examples of holococcolith-heterococcoiith com­ Lecal-Schlauder, J. (1961). Anomalies dans la composition des bination coccospheres and their implications for coccolithophorid coques de flagellés calcaires. Bull. Soc. Hist. Nat. Afr. N., 52: biology. Mar. Micropaleontol., 39: 1-34. 63-66. D a r l i n g , K .F., W a d e , C M., S t e w a r t , J.A., K roon, D., Dingle, L o e b l i c h , A.R. & T a p p a n , H. (1963). Type fixation and validation R. & B r o w n , A.J.L. (2000). Molecular evidence for genetic of certain calcareous nannoplankton genera. Proc. Biol. Soc. mixing of Arctic and Antarctic subpolar populations of Washington, 16: 191-198. planktonic foraminifers. Nature, 405: 43-47. L o e b l i c h , A.R. & T a p p a n , H. (1978). The coccolithophorid genus D eflandre, G. (1952). Classe des Coccolithophoridés (Cocco­ Calcidiscus Kamptner and its synonyms. J. Paleontol., 52: lithophoridae Lohmann, 1902). In Traité de zoologie. Anatomie, 1390-1392. systématique, biologie I, part 1, Phylogénie. Protozoaires: L o h m a n n , H. (1902). Die Coccolithophoridae, eine Monographie généralités. Flagellés (Grassé, P.P., editor), 439-470. Masson, der Coccolithen bildenden Flagellaten, zugleich ein Beitrag zur Paris. Kenntnis des Mittelmeerauftriebes. Arch. Protistenk., 1: 89-165. F r e s n e l , J. (1994). A heteromorphic life cycle in two coastal M oshkovitz, S. & O s m o n d , K. (1989). The optical properties and coccolithophorids, Hymenomonas lacuna and Hymenomonas microcry stall ography of Arkhangellskielaceae and some other coronata (Prymnesiophyceae). Can. J. Bot., 72: 1455-1462. calcareous nannofossils in the Late Cretataceous. In Nannofosils F r e s n e l , J. & B i l l a r d , C. (1991). Pleurochrysis placolithoides sp. and their Applications (Cnui, J.A. & van Heck, S.E., editors), nov. (Prymnesiophyceae), a new marine coccolithophorid with 76-97. British Micropalaeontology Society Series. remarks on the status of cricolith-bearing species. Br. Phycol. J., M urray, G. & Blackman, B.A. (1898). On the nature of the 26: 67-80. coccospheres and rhabdospheres. Phil. Trans. R. Soc. Lond., Ser. G a a r d e r , K.R. (1962). Electron microscope studies on holo­ B, 190: 427-441. coccolithophorids. Nytt. Mag. Bot., 10: 35-50. N o r r i s , R E. (1985). Indian Ocean nannoplankton. II. Holo­ G a a r d e r , K.R. & H e im d a l, B.R. (1977). A revision of the genus coccolithophorids (Calyptrosphaeraceae, Prymnesiophyceae) Syracosphaera Lohmann (Coccolithineae). Meteor. Forsch. with a review of extant taxa. J. Phycol., 21: 619-641. Ergebn., D24: 54-71. P a r k e , M . & A d a m s , I. (1960). The motile {Crystallolithus hyalinus G a a r d e r , K.R. & M a r k a l i , J. (1956). On the coccolithophorid Gaarder and Markali) and non-motile phases in the life-history o f Crystallolithus hyalinus n. gen., n. sp. Nytt. Mag. Bot., 5: 1-5. Coccolithus pelagicus (Wallich) Schiller. J. Mar. Biol. Assoc. UK, 39: 263-274. G a y r a l , P. & F r e s n e l , J. (1983). Description, sexualité et cycle de P i e n a a r , (1994). Ultrastructure and calcification of développement d'une nouvelle Coccolithophoracée R.N. coccolithophores. In Coccolithophores (Winter, A. & Siesser, (Prymnesiophyceae): Pleurochrysis pseudoroscoffensis sp. nov. W.G., editors), 13-37. Cambridge University Press, Cambridge. Protistologica, 19: 245-261. R a y n s , D.G. (1962). Alternation of generations in a cocco­ GREEN, J.C., C o u r s e , P.A. & T a r r a n , G. A. (1996). The life-cycle of lithophorid, Cricosphaera carterae (Braarud & Fragerl.) Braarud. Emiliania hu.xleyi : a brief review and a study of relative ploidy J. Mar. Biol. Assoc. UK, 42: 481-484. levels analysed by flow cytometry. J. Mar. Syst., 9: 33-44. R e n a u d , S. & K l a a s , C . (2001). Seasonal variations in the H e im d a l, B.R. (1982). Validation of the names of some species of morphology of the coccolithophore Calcidiscus leptoporus off Zygosphaera Kamptner. Int Nannoplankton Assoc. Newsl. 4: Bermuda (N. Atlantic). J. Plankton Res., 23: 779-795. 52-56. R o w s o n , J.D., Leadbeater, B.S.C. & G r e e n , J.C. (1986). Calcium H e im d a l, B.R. & G a a r d e r , K.R. (1980). Coccolithophorids from carbonate deposition in the motile {Crystallolithus) phase o f the northern part of the eastern central Atlantic. I. Holo­ Coccolithus pelagicus (Prymnesiophyceae). Br. Phycol. J., 21: coccolithophorids. Meteor. Forsch. Ergebn., D32: 1-14. 359-370. K a m p t n e r , E. (1936). Uber die Coccolithineen der Sudwestkuste Sam tleben, C. & B i c k e r t , T. (1990). Coccoliths in sediment traps von Istrien. Anz. Akad. Wiss. Wien Math-Naturw. Kl., 75: from the Norwegian Sea. Mar. Micropaleontol, 16: 39-64. 243-247. Samtleben, C. & S c h r o d e r , A. (1992). Living coccolithophore K a m p t n e r , E. (1937). Neue und bemerkenswerte Coccolithineen communities in the Norwegian-Greenland Sea and their record in aus dem Mittelmeer. Arch. Protistenk., 89: 279-316. sediments. Mar. Micropaleontol, 19: 333-354. K a m p t n e r , E. (1941). Die Coccolithineen der Sudwestkuste von Saugestad, A H. (1967). Planteplankton i vestlige Middlehav mars- Istrien. Ann. Naturh. Mus. Wien., 54: 54-149. april 1961. Thesis, University of Oslo. K a m p t n e r , E. (1950). Über den submikroskopischen Aufbau der Saugestad, A H. & H e im d a l, B.R. (2002). Light microscope studies Coccolithen. Anz. Akad. Wiss. Wien., 87: 152-158. on coccolithophorids from the western Mediterranean Sea, with K a m p t n e r , E. (1954). Untersuchungen fiber den Feinbau der notes on combination cells of Daktylethra pirus and Coccolithen. Arch. Protistenk., 100: 1-90. Syracosphaera pulchra. Plant Biosyst., 136: 3-28. 3 R esults 123

M . Geisen et al. 550

S c h i l l e r , J. (1930). Coccolithineae. In Krypiogamen-Flora von V rind, J.P.M ., Kok, D., de B ruun,W.C. & P arker, S B. (1984). Deutschland, Osterreich und der Schweiz, vol. 10, Flagellatae, M echanism of calcification in the m arine alga Emiliania huxleyi. 89-226. Akademische Verlagsgesellschft, Leipzig. Phil. Trans. R. Soc. Lond., Ser. B, 304: 435-444. T h o m s e n , H.A., B u c k , K.R. & C h a v e z , F.P. (1991). Choano- W estbroek, P., Y o u n g , J R. & Linschooten, K. (1989). Coccolith flagellates of the central California waters: taxonomy, mor­ production (biomineralization) in the marine alga Emiliania phology and species assemblages. Ophelia, 33: 131-164. huxleyi. J. Protozoal., 36: 368-373. T h o m s e n , H.A., 0 s t e r g a a r d , J.B. & H a n s e n , L.E. (1991). Het­ W i n t e r , A. & S ie s s e r , W.G. (1994). Atlas of living Cocco­ eromorphic life-histories in arctic coccolithophorids lithophores. In Coccolithophores (Winter, A. & Siesser, W.G., (Prymnesiophyceae). J. Phycol., 2 7 : 634-642. editors), 107-159. Cambridge University Press, Cambridge. V a l e r o , M., R i c h e r d , S., P e r r o t , V. & D e s t o m b e , C. (1992). Y o u n g , J R. (1992). The description and analysis of coccolith Evolution of alternation of haploid and diploid phases in life structure. In Proceedings of the 4th INA Conference, Prague 1991 cycles. Trends Ecol. Evoi, 7 : 25-29. (Hamrsmid, B. & Young, J R., editors). Knihovnicka ZPN, 14a: VON S t o s c h , H.A. (1954). Die Oogamie von Biddulphia mobiUensis 35-71. und die bisher bekannten Auxosporenbildungen bei den Cen­ Y o u n g , J R., D a v is , S.A., B o w n , P R. & M a n n , S. (1999). trales. Eighth Congr. Int. Bot., Rap. et Com. Sect., 1 7 : 58-68. Coccolith ultrastructure and biomineralisation. J. Struct. Biol., VON S t o s c h , H.A. (1972). La signification cytologique de la cyclose nucléaire dans le cycle de vie des Dinoflagellés. Mem. Soc. Bot. 126: 195-215. Fr., 201-212. Y o u n g , J R., B illard, C., Bown, P R., Cros, L., Davis, S.A., W a l l i c h , G.C. (1877). Observations on the coccosphere. Ann. Mag. Geisen, M., K aw achi, M., M ann, S., M edlin, L.K. & Probert, Nat. Hist. (Ser. 4), 19: 342-350. I. A.S. (2000). Phylogeny of coccolithophores and the evolution of W e s t b r o e k , P., d e J o n g , E.W., v a n d e r W a l , P., B o r m a n , A.H ., d e calcification. J. Nannoplankton Res., 22: 153-155. 4 Discussion 124

4 Discussion: Species level variation in coccolithophores

This discussion chapter has been written for a book to be published by Springer in 2003. This book follows from the conference Coccolithophores - from molecular processes to global impact, chaired by Hans Thierstein and Jeremy Young and held in early 2002 at the Monte Verita - a lovely conference centre near Ascona and run by the Swiss Federal Institute of Technology Zurich. The book aims to be a sequel to the well known book Coccolithophores, edited by Amos Winter and Bill Siesser in 1994. In addition to reviewing the state of art in research the various chapters will summarise the advances made in coccolithophore research over the last 10 years and especially in key CODNET working areas such as ecology and evolution. For further information about the book please visit http://www.coccoco.ethz.ch/. In the chapter presented here - Species level variation in coccolithophores - we tie together data from various sources, namely morphology, molecular biology, life cycle work and biogeography, to elucidate speciation and evolutionary processes in the CODENET key species. Data were contributed by Alberto Garcia-Saez (molecular biology), Ian Probert (cultures), Karl-Heinz Baumann (life-cycles and biometry), Lluisa Cros (taxonomy and life-cycles) and Jeremy Young (taxonomy, life-cycles and interpretation). The first author has contributed data from the following fields: biometry, life­ cycles, cultures, physiology and taxonomy. The interpretation of the various data was performed in collaboration with Jeremy Young and the author of this thesis wrote the text. The manuscript is accepted for publication with minor revision, the version in this thesis corresponds to the submitted manuscripts and includes changes suggested by the panel during the PhD examination as well as those of two independent reviewers. 4 Discussion 125

Species level variation in coccolithophores

Markus Geisen'’^, Jeremy R. Young^, Ian Probert\ Alberto G. Saez", Karl-Heinz Baumann^, Claudia Sprengel', Jorg Bollmann®, Lluïsa Gres’, Colomban de Vargas®, and Linda K. Medlin’

* * Pelagic Ecosystems, Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, D-27570 Bremerhaven, Germany, [email protected]. [email protected]. [email protected] ’ Palaeontology Department, The Natural History Museum, Cromwell Road, London SW7 5BD, England, [email protected] ^ Laboratory of Marine Biology and Biotechnology, University of Caen, F-14032 Caen, France. [email protected] Department of Biological Sciences, Imperial College London, Silwood Park Campus, Ascot, Berkshire SL5 7PY, England, [email protected] ^ Department of Geosciences, University Bremen, Postfach 330440, D-28334 Bremen, Germany, [email protected] ^ Geological Institute, Swiss Federal Institute of Technology, Sonneggstrasse 5, CH-8092 Zurich, Switzerland. [email protected] ’ Institut de Ciències del Mar, CMIMA-CSIC, Passeig Maritim de la Barceloneta, 37-49, 08003 Barcelona, Spain, [email protected] ® Institute for Marine and Coastal Sciences, Rutgers University, New Brunswick, New Jersey, USA. [email protected]

Abstract

Oceans cover roughly seventy percent of the earth’s surface. With few exceptions, minute photosynthetic primary producers like diatoms, dinoflagellates, silicoflagellates and coccolithophores inhabit the euphotic zone of this vast expanse. This phytoplankton forms the basis of the marine food chain and plays an important role in geochemical cycles. Knowledge of species level biodiversity and spéciation is thus important for understanding marine ecology and biogeochemistry. Amongst the groups mentioned, coccolithophores appear to be the ideal test group since their biomineralised periplasts - the coccoliths - provide a rich suite of qualitative and quantitative morphological characters and a uniquely extensive fossil record. In addition to this, extant coccolithophore species can be grown in culture and hence are available for molecular genetic studies and cytological research. For the CODENET project a number of extant species with seemingly global occurrence and spanning the biodiversity of coccolithophores were selected and different case studies were carried out to elucidate their species level biodiversity. The key merit of our approach was our using of multiple, independent lines of evidence 4 Discussion 126 so as to remove ambiguity implicit in any one type of study. Observations on living material can thus further the understanding of species concepts applied to coccolithophores and - in turn - can refine the interpretation of the evolutionary record of coccolithophores. Here we present the results and discuss tentative models of spéciation for the coccolithophores in general.

Introduction

Coccolithophores are a group of marine calcifying algae of the division Haptophyta (Jordan and Chamberlain 1997). The conventional interpretation of coccolithophore systematics is that there are about 165 well-described heterococcolithophore species (own data), with, in almost all cases, inter-oceanic distributions within broad ecological boundaries (Jordan and Chamberlain 1997; Jordan and Kleijne 1994; Young and Bown 1994). A perception of very widely distributed, rather homogeneous species is well supported by geological evidence of synchronous apparently sympatric evolution across the world oceans and for near synchronous (on scales of less than a few thousand years) extinction events (Chepstow-Lusty et al. 1992; Wei 1993; Wei and Shilan 1996). Research over the past decade via the integration of data from various sources (morphology, life-cycles, and molecular studies) has greatly refined our knowledge of fine scale diversity in this group. This has led to the development of hypotheses of causation in terms of ecophenotypic or genotypic variation. New data have come from different sources which will subsequently be discussed briefly.

Morphology

The primary source of data has been morphometric investigation of selected taxa using plankton, sediment trap, culture material, surface sediment samples, and often time series studies of geological samples. On the basis of quantifiable morphometric parameters - mostly coccolith size, but also the number of elements (in Calcidiscus spp.) and the bridge angle (in Gephywcapsa spp.) - this has led to the identification of sub-morphotypes and morphological gradients within species (Bollmann 1997; Knappertsbusch 1997, 2000). In addition to quantitative methods, qualitative morphological data including shield structure, crystallographic axis orientation (Young et al. 1999; Young et al. 1992), suture lines, and ornamentation of coccoliths (Kleijne 1993) have provided an important additional source of information about intra-specific variation. Examples include Calcidiscus spp., where the expression of suture lines and the appearance of the zone surrounding the central pore have proved to be crucial for the understanding of the diversity within the genus (Kleijne 1993). In Helicosphaera spp., the stability of fine morphological characters in the central area was validated as a species discriminating feature (Geisen 2003). In Umbilicosphaera spp., mono- 4 Discussion 127 versus bicyclic shield structure and central area features were used to differentiate species (Geisen 2003).

Ecological / biogeographical separation

As the concept of biological species requires reproductive isolation, a sound knowledge of the geographical and ecological ranges of the putative morphotypes is important. Evidence here stems both from the biogeographical mapping of distributions of differing morphologies and from studies of species associations from sediment trap time series. Different factors have been identified: In Coccolithus spp., there appears to be a spatial variation, separating a subarctic species from a temperate species with partially overlapping ranges (Cachao and Moita 2000); in Calcidiscus spp., Renaud and Klaas (2001) have demonstrated a temporal succession of morphologies; and finally the two “pseudo-cryptic” species (for a review on cryptic species refer to Knowlton 1993) detected in Symcosphaera apparently share the same habitat (Geisen et al. 2002).

Culture studies

A critical test for the interpretation of fine morphological differences is whether these features are stable in culture. Testing comes from studies of the degree of morphological variation within monoclonal strains grown under varying ecological conditions. Doing this enables one to assess to what extent there is genetic control of the morphological expression of a species or if the observed variation is ecophenotypic in nature. This method was first applied to coccolithophores by Young and Westbroek (1991) to assess the degree of genotypic variation in E. huxleyi. However, isolating and maintaining species in clonal cultures is very labour intensive and since then further tests have only been performed on Calcidiscus leptoporus (Quinn et al. 2003). The CODENET keystone species and a number of additional species isolated during the project have now provided the opportunity for a more detailed study of the morphological variability of coccolithophores in culture. Data here include quantitative and qualitative morphological characters as described above (Geisen 2003).

Genetic separation

A set of molecular markers was used to study the extent of genotypic variation between and within recognised morphological species. Different genes from the nucleus and the chloroplast - each with varying rates of nucleotide substitution - were sequenced. The slowly evolving and nuclear gene 18S rDNA was used to quantify inter-specific differences, reconstruct the molecular phylogeny of 4 Discussion 128 coccolithophores and Haptophyta, and infer times of divergence in these groups (e.g. Edvardsen et al. 2000; Fujiwara et al. 2001; Medlin et al. 1997). Within the CODENET project, we used it for the same purposes, extending our analyses to a much larger set of species (Saez et al, this volume, Saez et al. 2003). Faster evolving plastid genes have also been used for resolving coccolithophorid and Haptophyta phytogenies: rbcL (Fujiwara et al. 2001) and tu f A (Saez et al. 2003). In addition to the nuclear 18S rDNA, including the very fast-evolving ribosomal internal transcribed spacer region (ITS rDNA), the molecular marker tufA was also sequenced in isolates of the same species, to assess the biological significance of phenotypic variation observed within recognised morphological species (de Vargas et al. and Quinn et al., this volume, Saez et al. 2003). Although DNA has been amplified from single cells using the Polymerase Chain Reaction (PCR) (e.g. de Vargas et al. 1999), this has not yet proved successful with coccolithophores, so our molecular work still depends on cultures to correlate the fine-scale morphological differences with molecular results. As the number of strains in culture was the limiting factor, our approach was to sample genes from different loci - the chloroplast and the nucleus - to obtain, if possible, concordant results. Molecular genetic data can be used not only to test whether putative fine-scale morphological variation is genotypic in origin but also to provide estimates of the likely time of divergence. Our data indicate that both 18S rDNA and tufA of coccolithophores evolved in a clock-like manner and basic calibration of this record is possible (Saez et al. 2003). There are substantial uncertainties in any molecular clock estimate. Nonetheless it is certainly possible to use genetic distance data to infer order of magnitude estimates of divergence times and hence to discriminate between possibilities such as that pairs of genotypes diverged within the last 1 Ma or several Ma ago.

Life-cycles and holococcolith phase differentiation

A fifth source of data has come from recognition of alternate life-cycle phases. It is now clear that the typical life-cycle of coccolithophores consists of independent haploid and diploid phases, both of which are capable of indefinite asexual reproduction (Billard et al., this volume). Both phases usually produce coccoliths but via distinctly different biomineralisation processes resulting in consistent structural differences. In the diploid phase biomineralisation occurs intra-cellularly and produces heterococcoliths formed of radial arrays of complex crystal-units. In the haploid phase biomineralisation occurs extracellularly and produces holococcoliths formed of numerous minute euhedral crystallites. Evidence of this stems from culture observations of phase transitions (e.g. Parke and Adams 1960), observations of meiosis and syngamy (Gayral and Fresnel 1983), chromosome counts (Fresnel 1994; Rayns 1962), flow cytometry on cultured clones (Probert unpubl.), and observations of combination coccospheres from natural populations (for a review refer to Cros and Fortuno 2002; Cros et al. 2000; Geisen et al. 2002). Thus the two phases potentially provide independent morphological evidence 4 Discussion 129 of differentiation. This potential has been realised in several cases, particularly through the recognition of rare combination coccospheres produced during phase transitions but also by the direct observation of phase changes in culture.

Summary - integrated methodology

Combining data from the sources mentioned above has allowed us to investigate fine-scale genotypic variation within selected coccolithophores. Although we have not been able to gather evidence from all the various methods for all the species studied, there are strong patterns emerging, which allows for some generalization even if some information is missing. The different strands of evidence have produced significant evidence of different levels of genotypic variation within conventional species, which we will discuss in this review.

Results

In this section we present the data from each of the six CODENET species. Each case study was different but some strong patterns emerged. Although this is mainly a review, unpublished data from our own observations are included as well. The sequence in which we present the various sources of data roughly represents the sequence in which the evidence became available and in some cases - as in Calcidiscus and Coccolithus - demonstrates how the introduction of further available data can radically change the interpretation.

Umbilicosphaera sibogae and U. foliosa (Plate 1)

Two varieties ((/. sibogae var. sibogae and U. sibogae \ 2ct. foliosa) have traditionally been recognised in this species based largely on coccosphere characters. There are correlative differences in coccolith morphology but it had been speculated that these were a consequence of the different coccosphere morphology. Hence alternative hypotheses were that these varieties were different life-cycle stages of a single species or discrete taxa. Umbilicosphaera was selected as a CODENET species to conduct a case study on infra-specific variation of two closely related taxa with an opportunity to study the evolutionary history as well. Based on molecular differences Saez et al. (2003) have raised the two varieties to species rank.

Quantitative and qualitative morphology

The coccolithophorid genus Umbilicosphaera comprises four extant species (Winter and Siesser 1994) and two extinct fossil species (Young 1998). The most common extant species U. sibogae and U. foliosa show a broad inter-oceanic occurrence 4 Discussion 130 and are common in the fossil record but there are few reliable data on their first occurrence. The genus Umbilicosphaera has a continuous and well-documented fossil record back to the Early Miocene (23 Ma) (Young 1998). Although McIntyre and Be (1967) state that the two species can easily be distinguished on morphological grounds there has been some confusion in the taxonomy as Okada and McIntyre (1977) recombined Cycloplacolithus foliosus into U. sibogae vdcc. foliosa on the basis that “extremely rare specimens having both types of coccoliths on a single coccosphere were observed”. Curiously this taxonomic separation has been upheld by later researchers, even though research on Umbilicosphaera demonstrated a stable morphology under culture conditions and diagnostic cytological differences between the species (Inouye and Pienaar 1984). Our own research - with data from cultures, natural samples, and sediments - has now validated these findings and it seems likely that Okada and McIntyre (1977) distinguished the coccoliths of the two varieties using the expression of the suture lines, which is a far less reliable criterion for species determination (see below). The two species exhibit significant diagnostic differences in both coccolith and coccosphere morphology. U. foliosa forms a compact spherical coccosphere consisting of up to 25 interlocking placoliths (Plate 1, figs 1, 2) which are circular with a narrow central opening typically with a few hook-like spines protruding into the central opening (Plate 1, figs 2, 7). Both shields are bicyclic and convex, and the proximal shield is smaller than the distal shield (Plate 1, figs 8, 10). The proximal shield is composed of R-units and the distal shield of V-units. The inner half of the elements on the distal shield are imbricated dextrally and have straight sutures, the outer half kinked, with sinistral imbrication and serrated sutures (Plate 1, fig. 7, 9). The central opening is spanned by an organic membrane (Plate 1, figs 1, 8). The cells are not colonial and the protoplast fills the entire coccosphere (Plate 1, fig. 5). Cells are non-motile, but flagellar bases are present, uncalcified body scales are absent (Inouye and Pienaar 1984). Although intensive research has been conducted on cultures both within CODENET, but also by other researchers (Inouye and Pienaar 1984), the life-cycle associations of the Umbilicosphaera spp. remain yet unknown. In contrast, U. sibogae forms a large spherical to subspherical coccosphere consisting of 40 to more than 100 partly interlocking placoliths (Plate 1, figs 3, 4, 6). The placoliths are circular with a large central opening (Plate 1, figs 11, 13), the proximal shield is flat and typically larger than the convex distal shield (Plate 1, fig. 12). Both shields are monocyclic, with the proximal shield composed of R-units and the distal shield of V-units. Elements on the distal shield are imbricated dextrally, sutures straight on the inner part of the rim, then kinked and incised laevogyre on the outer part of the rim (Plate 1, fig. 11). An organic membrane spans the central opening (Plate 1, fig. 12). Cells are semicolonial, with each cell typically containing 1-2, occasionally four protoplasts (Plate 1, fig. 6). The protoplasts do not fill the entire coccosphere (Plate 1, fig. 6). Cells are non-motile, but flagellar bases are present (Probert, unpubl.). As in U. foliosa, uncalcified body scales are absent (Probert, unpubl.). 4 Discussion 131

Working with sediment trap material, Baumann and Sprengel (2000) successfully used pore size and distal shield diameter to distinguish the two species and their data show little inter-annual size variation for U. sibogae coccoliths. In addition to this, own measurements of coccolith rim width and shield diameter reveal a distinct morphospace for the two taxa and no gradualistic morphological shift between them since their first occurrence in the fossil record.

Culture studies and life-cycles

Although intensive research has been conducted on cultures within CODENET, and also by other researchers (Inouye and Pienaar 1984) the life-cycle associations of the Umbilicosphaera spp. remain unknown. Culture studies in CODENET have, however, demonstrated the morphological stability of the two species in culture. Strains of both species have been maintained in culture for several years. During this period, the cultures were repeatedly checked by both light microscope (LM) and scanning electron microscope (SEM) but no evidence whatever of transitions between the two species was observed. A range of ecological conditions - mainly light and temperature - was tested and no significant variation in coccolith size and characters was detected.

Molecular studies

Three genes were sequenced: rbcL (Fujiwara et al. 2001), ISS rDNA and tufA (Saez et al. 2003). All of them show a high number of substitutions between the varieties. Based on the differences in morphology and the molecular differences Saez et al. (2003) concluded that the two variants are indeed distinct biological species. Accordingly they suggest referring to them as separate species. Estimates from a molecular clock based on the tufA tree date the divergence of the lineage leading to the two species back at 5.59 (±1.15) Ma (Saez et al. 2003). Our own biostratigraphic work however dates the first occurrence of U. sibogae at 2.2 Ma and that of U. foliosa at 2.9 Ma, so it can be hypothesised here that U. sibogae and U. foliosa are not direct sister taxa or differentiated morphologically much later than they did genetically.

Summary

All the data indicate that they are discrete species. This includes morphometric evidence that the coccolith morphotypes do not intergrade and the recognition of additional species distinguishing qualitative characters. Transitional morphotypes have not been observed in culture. Molecular data have revealed a significant number of substitutions between the species and points to a divergence time between 4.44 and 6.74 Ma. Nonetheless, the two taxa cluster together on all molecular trees so they are clearly closely related, as suggested by the coccolith morphology. 4 Discussion 132

Plate 1 Umbilicosphaera spp. Fig. 1: Scanning electron micrograph of two U. fo lio sa coccospheres. U. fo lio sa cells are typically found in clusters of up to four cells. Water sample, western Mediterranean, Alboran Sea, R /V H esperides cruise MATER 2, station 69. Fig. 2: Scanning electron micrograph of two U. fo lio sa coccospheres. Note the presence of hook like protrusions in the central opening. Although this appears to be a stable taxonomic character at first cells with both types of coccoliths have been observed frequently. Water sample, N. Atlantic, R/V M eteor cruise 38-1, station 11. Fig. 3,4: Scanning electron micrograph of a U. sibogae coccosphere. Note the organic mem­ brane in Fig. 4 spanning the central opening of some coccoliths. Water sample, S. Atlantic, off Namibia, R/V M eteor cruise M48-4, station 470 (Fig. 3) and station 20 (Fig. 4). Fig. 5: Light micrograph in DIC of a cluster of four U. fo lio sa cells. Culture sample (ESP 6M1), western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 6: Light micrograph in DIC of a dividing U. sibogae cell. U. sibogae cells can be semi­ colonial with typically two cells in a single coccosphere. Note the large extracellular space. Culture material (ASM 39), western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 7: Scanning electron micrograph of the bicyclic distal shield of an U. fo lio sa coccolith. Sediment trap sample, Indian Ocean, off Somalia. Fig. 8: Scanning electron micrograph of the bicyclic proximal shield of U. fo lio sa coccoliths. Note the organic membrane spanning the central opening. Culture sample (ESP 6M1), west­ ern Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2. Fig. 9: Scanning electron micrograph of the bicyclic distal shield of U. fo lio sa coccoliths. Note the straight suture lines of the inner cycle in contrast with the ragged suture lines of the outer cycle. Water sample, N. Atlantic, R/V M eteor cruise 38-1, station 13. Fig. 10: Scanning electron micrograph of a U. fo lio sa coccolith in lateral view. The distal shield is larger than the proximal shield. Water sample, western Pacific, Miyake- jima island, Ibo Port, Japan. Fig. 11: Scanning electron micrograph of the monocyclic distal shield of an U. sibogae coccolith. Proximal shield larger than distal shield. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 59. Fig. 12: Scanning electron micrograph of the monocyclic proximal shield of U. sibogae coccoliths. Note the organic membrane spanning the central opening. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 59. Fig. 13: Scanning electron micrograph of U. sibogae coccoliths from a single cocco­ sphere. Note the size variation of the central opening and the rim. Water sample, western Mediterranean, Alboran Sea, R/V H esperides cruise MATER 2, station 15.

Scale bars represent: Figs 1-4: 5pm ,figs 5, 6: not to scale, figs 7, 8, 11-13: 2pm , figs 9, 10: \p m . 4 Discussion 133

3. U. sibogae coccosphere

4. U. Sibogae coccosphere

e

12. U. sibogae coccolith 13. U. sibogae coccoliths 11. u. sibogae coccoliths distal proximal view distal view 4 Discussion 134

Coccolithus spp. (Plate 2)

Coccolithus pelagicus (Wallich 1877) is one of the most robust and longest ranging of extant coccolithophores and its life-cycle is exceptionally well documented (Geisen et al. 2002; Parke and Adams 1960; Rowson et al. 1986). It appeared in the early Palaeocene and it is common throughout the Cenozoic (Perch-Nielsen 1985), although this very long fossil record is based on a very broad species concept. Today this species is commonly found in high latitudes, where it is known from the North Atlantic and the Subarctic area (Geitzenauer et al. 1977; McIntyre and Be 1967; Okada and McIntyre 1977; Samtleben and Schroder 1992). In addition, this genus is observed in low abundance in upwelling areas (e.g. Cachao and Moita 2000). Two morphotypes have been identified based on coccolith characters - a large, temperate form and a smaller, subarctic form. The morphotypes have generally been considered to represent ecophenotypic end-members of a single rather variable species. Recently however, based on life-cycle observations (Geisen et al. 2002) and genetic studies (Saez et al. 2003), the species character of the putative ecosphenotypes has been conclusively demonstrated and consequently named C. pelagicus (the subarctic species) and C. braarudii (the temperate species).

Quantitative and qualitative morphology

A range of studies with material from sediment trap, cultures and seawater samples have been carried out on the qualitative morphology of Coccolithus pelagicus sensu lato using both SEM and LM. Baumann et al. (2000) have used material from sediment traps from a north-south transect in the North Atlantic Ocean and demonstrated a clear change from a unimodal size distribution of the placoliths for the northern locations to a bi-modality in the more temperate locations. However, these data, based on coccolith rather than coccosphere measurements proved difficult to interpret due to the broad overlap in coccolith size between the two morphotypes. If measurements are performed on coccospheres, the two populations become clearly separable, with the temperate populations (C. braarudii) being dominated by large coccospheres (Plate 2, fig. 1) and the subarctic populations (C. pelagicus) by small coccospheres (Plate 2, fig. 5). Both morphotypes share most of their qualitative characters. However, in the large form the central area is open and spanned by a cross-bar (Plate 2, figs 2, 3) whereas the central area is usually closed in the small form (Plate 2, figs 6,7). There is however evidence that these characters are secondary, size dependent characters as small specimens of the temperate form can have closed central areas and large specimens of the arctic form can have bars and open central areas. Due to the nature of the morphological evidence there was a strong need to test whether the observed variations were ecophenotypic or genotypic in origin. Culture studies of both morphotypes under a range of different temperatures and light levels demonstrated morphological stability within each type, indicating that the 4 Discussion 135 differences in morphology (coccolith and coccosphere size and shape) are under genetic control.

Life-cycles and holococcolith phase differentiation

On the basis of life-cycle observations and morphology, Geisen et al. (2002) have demonstrated that temperate and subarctic populations of Coccolithus produce different holococcoliths in the alternate life-cycle phase (Plate 2, figs 4, 8 respectively). Evidence here arises from both culture studies (Probert, unpubl. and Parke and Adams 1960; Rowson et al. 1986) and life-cycle associations from natural associations (Baumann et al. 1997; Samtleben and Bickert 1990; Samtleben and Schroder 1992; Winter and Siesser 1994). Both holococcolithophores involved share the same coccolith rim structure. The morphology of the central area however is different. In the holococcolithophore stage of the temperate Coccolithus, the central area consists of a central ellipsis of crystallites with spokes radiating towards the rim (Plate 2, fig. 4). The holococcolithophore stage of the subarctic Coccolithus however features coccoliths with the calcite rhombohedra arranged in parallel rows with each crystal lying on one face and partly touching adjacent faces (Plate 2, fig. 8) On the basis of life-cycle studies and morphological observations Geisen et al. (2002) have amended the taxonomy and recombined Coccolithus to include two subspecies, C. pelagicus ssp. pelagicus and C. pelagicus ssp. braarudii.

Ecological / biogeographical separation

The two extant Coccolithus species show different, but partly overlapping biogeographies with C. pelagicus preferring colder, sub-arctic water masses with temperatures ranging from -1°C-14°C (Okada and McIntyre 1979; Winter et al. 1994) and the C. braarudii preferring temperate waters and upwelling regimes (Baumann et al. 2000; Cachao and Moita 2000; Geisen et al. 2002) with optimal growth conditions in a temperature range between 13-18°C (Cachao and Moita 2000). The discovery that extant Coccolithus consists of two subspecies with significant differences in ecological preference and geographic distribution can shed a new light on the interpretation of results from the palaeobiogeography of Coccolithus spp. (Ziveri et al., this volume).

Molecular studies

Only one sub-arctic strain was available in culture so a multi-gene approach was used to compensate for the lack of strains. The two subspecies of Geisen et al. (2002) turned out to be identical on the conservative 18S rDNA gene. However, the faster evolving genes tufA and ITS rDNA showed variation between the two subspecies, whereas no net variation was observed among 8 strains of C. pelagicus ssp. braarudii (Saez et al. 2003). Estimates from a molecular clock date the divergence of the two taxa at 2.15 (±0.57) Ma (Saez et al. 2003). On the basis 4 Discussion 136 of these important molecular divergences between the two subspecies, Saez et al. (2003) have concluded that Geisen et al. (2002) were too conservative in assigning the intra-specific subspecies rank and have raised the subspecies to species rank with the large, temperate species being C. braarudii and the small, subarctic species being C. pelagicus.

Summary

It has been demonstrated that Coccolithus pelagicus includes discrete, arctic and temperate species (Baumann et al. 2000; Saez et al. 2003). These can be separated according to heterococcolith size, holococcolith morphology and temperature tolerance - all characters which remain constant in culture (Cachao and Moita 2000; Geisen et al. 2002). Molecular genetic data from a range of genes support this differentiation into two discrete, but closely related species. Results from a molecular clock estimate the divergence time of the sister taxa between 1.58 and 2.72 Ma (Saez et al. 2003). 4 Discussion 137

Plate 2 Coccolithus spp., Gephyrocapsa spp. and Erniliania spp. Fig. 1: Scanning electron micrograph of a C. braarudii coccosphere. This species was previ­ ously known as the large, temperate morphotype of C. pelagicus. Water sample, S. Atlantic, off Namibia, R/V Meteor cruise M48-4, station 476. Fig. 2: Scanning electron micrograph of a C. braarudii coccolith in distal view. Culture sam­ ple (AS55T), western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2. Fig. 3: Scanning electron micrograph of a C. braarudii coccolith in proximal view. Sediment trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 4: Transmission electron micrograph of the holococcolithophore stage of C. braarudii. This stage was previously described as “Crystallolithus braarudii”. Note the central ellipse of crystallites with radial spokes connected to the rim. Culture sample (LKl), SW France, off Arcachon. Fig. 5: Scanning electron micrograph of a C. pelagicus coccosphere. This species was previ­ ously known as the small, arctic morphotype of C. pelagicus. Water sample, N. Atlantic, off Iceland. Fig. 6: Scanning electron micrograph of a C. pelagicus coccolith in distal view. Culture sam­ ple (IBV 74), N. Atlantic, off Iceland. Fig. 7: Scanning electron micrograph of a C. pelagicus coccolith in proximal view. Water sample, N. Atlantic, off Iceland. Fig. 8: Scanning electron micrograph of the holococcolithophore stage of C. pelagicus. This stage was previously described as ”Crystallolithus hyalinus”. Note the central area with the crystallites arranged in parallel rows. Water sample, N. Atlantic, JGOFS cruise. Fig. 9: Scanning electron micrograph of a G. oceanica coccosphere. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 59. Fig. 10: Scanning electron micrograph of a G. oceanica coccolith in proximal view. Sediment trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 11: Scanning electron micrograph of a G. ornata coccosphere. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 44. Fig. 12: Scanning electron micrograph of a G. ornata coccolith in lateral view. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 13: Scanning electron micrograph of a G. muellerae coccosphere. Water sample, west­ ern Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 14: Scanning electron micrograph of a G. muellerae coccoliths. Sediment trap sample, N. Atlantic, JGOFS cruise. Fig. 15, 16: Scanning electron micrograph of a G. ericsonii coccosphere. Fig. 16 displays an enlarged view of the same specimen. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 15. Fig. 17: Scanning electron micrograph of a E. huxleyi coccosphere. The coccoliths exhibit the A-type morphology. Water sample, N. Atlantic, R/V Meteor cruise 42-4B, station US IB. Fig. 18: Scanning electron micrograph of a E. huxleyi var. corona coccosphere. Note the col­ lar surrounding the central area. Water sample, N. Atlantic, off Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 19: Scanning electron micrograph of a E. huxleyi coccosphere. The coccoliths exhibit the C-type morphology. Water sample, N. Atlantic, R/V Meteor cruise 38-1, station 12. Fig. 20: Scanning electron micrograph of an E. huxleyi coccosphere. The coccoliths exhibit the A-type morphology with the central area being overcalcified. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 64.

Scale bars represent; Figs 1, 5: 5/

1. c. braarudii coccosphere 2. C. braarudii coccolith 3. C braarudii cocco- 4. C braarudii distal view lith proximal view holococcoliths

5. C, pelagicus coccosphere 6. C. pelagicus 7. C. pelagicus 8. C. pelagicus coccolith distal view coccolith proximal view holococcoliths

1

15. G. ericsonii c occ osp h er e13. G. muellerae 14 G muellerae coccosphere13. 16. G. ericsonii coccosphere coccolith distal view coccosphere. detail

19. fc. huxleyi 20. E huxleyi 17. E. huxleyi coccosphere coccosphere coccosphere 18 E. hux/eyi var. corona coccosphere 4 Discussion 139

Helicosphaera spp, (Plate 3)

According to Jordan and Young (1990) the extant cosmopolitan (Okada and McIntyre 1977; Winter et al. 1994) genus Helicosphaera consists of two species {H. carteri and H. pavimentum) with three varieties in H. carteri (Jordan and Green 1994; Jordan and Kleijne 1994). Helicosphaera is common in the sedimentary record and palaeontologists have successfully used fine morphovariants for biostratigraphy. More than 40 species are consequently recognised in the fossil record (Perch-Nielsen 1985). The most common extant form H. carteri var. carteri has a first occurrence of Late Oligocene age and H. carteri var. wallichii of Late Miocene (Tortonian) age (Young 1998). Three varieties have been described based on central area structures. These varieties - N. carteri var. carteri, H. carteri var. wallichii and H. carteri var. hyalina - have either been considered as separate species (Jordan and Young 1990) or as morphological extremes of intergradational morphotypes (Nishida 1979). Recent evidence from molecular studies (Saez et al. 2003) has led to the conclusion that the variants H. carteri var. carteri and H. carteri var. hyalina are separate biological species, resulting in the species H. carteri and H. hyalina. Based on life-cycle observations and morphology we argue here that H. carteri var. wallichii should equally be raised to full species rank.

Quantitative and qualitative morphology

Coccoliths of the genus Helicosphaera can be easily distinguished from other species due to their helical flange and their comparatively large size. For a review of the genus Helicosphaera see (Jafar and Martini 1975; Theodoridis 1984). The three extant varieties assigned to Helicosphaera share a number of morphological characters. All form ellipsoidal coccospheres with spirally arranged asymmetrical helicoliths. The cells are flagellate and the flagellar pole is surrounded by modified coccoliths, which usually exhibit a larger wing and tooth-like protrusions on the flange (Plate 3, figs 1, 4, 7). The central area characters however are distinctly different. In Helicosphaera carteri the central area of the helicoliths show a bar, which separates 1-2 dextrally angled, aligned openings (pores or slits) (Plate 3, figs 1-3). In contrast the central area of H. hyalina is characterised by tangentially arranged needle-shaped elements and the absence of pores or slits (Plate 3, figs 7-9). Finally, H. wallichii has a central area with a bar separating two dextrally aligned, oblique slits with kinked ends (Plate 3, figs 4-6). Simple size measurements of coccolith size within Helicosphaera sensu lato, based on water column, sediment, and sediment-trap samples resulted in hi- or multimodal morphospace that were difficult to interpret and could either represent genotypic or ecophenotypic variation. During CODENET multiple H. carteri (Plate 3, fig. 1) cultures became available and the morphological stability of coccolith size and shape if tested under a range of environmental conditions was demonstrated. However slit/pore size and shape 4 Discussion 140 were remarkably variable even within the same clonal culture. A single clone of H. hyalina (Plate 3, fig. 7) was cultured and - in comparison with H. cart. var. carteri - did exhibit a consistently smaller coccolith size in addition to the central area features described above. These features remained stable in culture and proved important for species recognition. Observations on material from cultured H. wallichii (Plate 3, fig. 4) provided by Isao Inouye confirmed the stability of the fine morphological characters, with all coccoliths showing oblique slits in the central area. A review of our large collection of scanning electron micrographs of coccoliths and coccospheres of Helicosphaera from all oceans exhibited a rather high morphological variability of the central area characters in H. carteri, even on a single coccosphere. However the central area characters of H. hyalina and H. wallichii proved to be consistently different. The findings on the natural samples hence support our morphological observations on the cultured clones. These results from morphological work indicate that there is a strong genotypic control on the morphological variation in Helicosphaera.

Life-cycles and holococcolith phase differentiation

Recent evidence from combination cells of H. carteri with a holococcolithophore “Syracolithuscatilliferus” (Plate 3, figs 10-12) (Cros et al. 2000) and combination cells of the holococcolithophores “Syracolithus catilliferus” and “Syracolithus confusus” (Plate 3, fig. 13) (Geisen et al. 2002) suggests that both holococcolithophores are the haploid phase of the life-cycle of H. carteri. Cros et al. (2000) and Geisen et al. (2002) note the close morphological relationship between the two holococcoliths involved (Plate 3, figs 10, 12) and explain this similarity by (intra-specific) variations in the degree of calcification. Although the available examples are few and based on observations on natural populations only, they provide convincing evidence to treat both holococcolithophores as junior (heterotypic) synonyms of H. carteri. We have further tentative evidence from a single combination coccosphere for a life-cycle association of H. wallichii with a holococcolithophore Syracolithus dalmaticus (Plate 3, figs 14-16). This is not a particularly clear example of a combination cell and could indeed be an accidental association. However, in our field samples of the variant wallichii the holococcolithophore S. dalmaticus usually occurs as well. So we tentatively conclude that there is evidence for differentiation of both holococcolith and heterococcolith between H. wallichii and H. carteri.

NB Helicosphaera wallichii (Lohmann 1902) Boudreaux and Hay (1969) has priority over Syracolithus dalmaticus (Kamptner 1927) Loeblich and Tappan (1966), hence - if the combination is proved - the correct name for the species will be H. wallichii. There is no evidence available on the haploid phase of H. hyalina, although it is interesting to note that the holococcoliths within the genus Helicosphaera have a highly distinctive ultrastructure, formed predominantly of aligned rhombohedral crystallites (Plate 3, fig. 11). One further holococcolithophore with this ultrastructure occurs, S. ponticuliferus, and so is a prime candidate as the H. hyalina holococcolithophore. 4 D iscussion 141

Molecular studies

Two genes - the conservative 18S rDNA and the faster evolving tufA - have been targeted for the two varieties carteri and hyalina within Helicosphaera available in culture. The tufA gene exhibited a large number of substitutions between both varieties. However, and despite numerous attempts with H. hyalina, 18S rDNA it was only possible to sequence for H. carteri. Based on the differences in morphology and the molecular data Saez et al. (2003) have raised these variants to species rank, consequently named H. carteri and H. hyalina. They also inferred a most recent common ancestor for these two taxa at 10.19 (±1.97) Ma. Due to the lack of new isolates ofH. wallichii comparative molecular data are not available.

Summary

Morphometric analyses again produced ambiguous patterns, which were difficult to interpret. Our key result came from culture isolation of a disputed morphovariant - H. hyalina. Contrary to expectations the coccolith morphology has proved stable in culture indicating again that a subtle morphological variant is under genotypic control. On the basis of our new results from morphological and molecular studies Saez et al. (2003) have concluded that the two varieties are in fact separate, albeit closely related, species and that their most recent common ancestor would have lived between 8.22 and 12.16 Ma. Although molecular data are still pending, evidence from the morphology, life-cycle data and previous culture observations strongly support that the previously described variant H. cart. var. wallichii should be eonsidered a discrete species as well. These findings have led to the suggestion that in this species sympatric evolution of a phenotypically plastic population might lead to gradualistic change in the range of morphological variation within a single species. 4 Discussion 142

Plate 3 Helicosphaera spp. Fig. 1; Scanning electron micrograph of a H. carteri coccosphere. The helicoliths show the typical spiral arrangement and possess enlarged flanges in the circumflagellar coccoliths. The central area of this specimen shows the typical morphology with two aligned slits, which are separated by a bar. Morphotypes with 1 -2 pores and intermediate central area morphologies have also been observed. Note the little triangular protrusions on the flange. Water sample, N. Atlantic, Portuguese shelf, R/V Andromeda cruise CODENET 2, station 6. Fig. 2: Scanning electron micrograph of a H. carteri coccolith in proximal view. Sediment trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 3: Scanning electron micrograph of a H. carteri coccolith in distal view. Sediment trap sample, S. Atlantic. Image courtesy Babette Bockel, Univ. Bremen. Fig. 4: Scanning electron micrograph of a coccosphere of H. wallichii. The central area shows the typical morphology with two angled slits with kinked ends which are separated by a bar. This fine morphological feature is stable in culture. Note the little triangular protrusions on the flange. Water sample, western Pacific, Miyake-jima island, Miike Port, Japan. Fig. 5: Scanning electron micrograph of a H. wallichii coccolith in proximal view. Note the kinked ends of the aligned slits. Water sample, western Pacific, Miyake-jima island, Miike Port, Japan. Fig. 6: Scanning electron micrograph of a//, wallichii coccolith in distal view. Sediment trap sample, Indian Ocean, off Somalia. Fig. 7: Scanning electron micrograph of a H. hyalina coccosphere. The central area is filled with tangentially arranged needle shaped elements. Note the little triangular protrusions on the flange. Culture sample (NAP 11), Mediterranean, off Naples, Italy. Fig. 8, 9: Scanning electron micrograph of a H. hyalina coccolith in proximal view (fig. 8) and distal view (fig. 9). Culture sample (NAP 11), Mediterranean, off Naples, Italy. Fig. 10: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. This stage was previously described as “Syracolithus catilliferus” and is referred to as H. carteri HO solid-type. Water sample, N. Atlantic, off Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 11: Scanning electron micrograph of H. carteri coccoliths in the holococcolith stage. Water sample, Antarctic Ocean, cruise JR 48. Fig. 12: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. This stage was previously described as “Syracolithus confusus” and is referred to as H. carteri HO perforate-type. Water sample, western Mediterranean, off Barcelona. Fig. 13: Scanning electron micrograph of a H. carteri coccosphere in the holococcolith bear­ ing stage. Note that presence of coccoliths of both H. carteri HO solid and perforate. This is seen as an example of intraspecific variation in the degree of calcification. Water sample, NW Mediterranean, cruise MESO-96, station F2. Fig. 14: Scanning electron micrograph of a S. dalmaticus holococcolithophore. Water sam­ ple, western Pacific, Miyake-jima island, Ibo Port, Japan. Fig. 15: Scanning electron micrograph of a detail of S. dalmaticus holococcoliths. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 16: Scanning electron micrograph of a tentative H. wallichii - S. dalmaticus combina­ tion coccosphere. Water sample. Gulf of Mexico, R/V Gyre cruise 92-G-03, station 9.

Scale bars represent: Figs 1, 4, 7, 10, 12-14: 5pm, figs 2, 3, 5, 6, 8, 9, 11, 15: 2pm. 4 Discussion 143

3. H. carten coccolith 1. H. carten coccosphere 2. H. carten coccolith proximal view distal view

4M

4, H. wallichii coccosphere 5. H. wamch;/. coccolith 6 . H. wallichii coccolith proximal view distal view

B H. hyalina coccolith 7. H. hyalina coccosphere 9. H. hyatina coccolith proximal view distal view

10. H. carten holococcoli- 11 H carter? holococcoliths 12. H carten holococcoli- 13. H. cartert holococcolitho thophorid solid type solid type thophorid perforate type phorid solid & perforate type

14. S. da/nraf/cüs holococcoli- 15. S. dalmaticus holococcoliths 16. H. wallichii - S dalmaticus thophorid possible combination coccosphere 4 Discussion 144

Calcidiscus spp. (Plate 4)

Together with the Gephyrocapsa / Erniliania plexus, Calcidiscus is probably the best documented coccolithophorid genus (Baumann and Sprengel 2000; Brand 1981; Knappertsbusch 1997,2000; Renaud and Klaas 2001 ; Renaud et al. 2002). Like the aforementioned genera it has a broad, inter-oceanic occurrence spanning a range of ecological variation and a very good, continuous fossil record, spanning the last 23 Ma. Three extant morphotypes have been tentatively identified, based largely on size of the coccoliths and coccospheres and there has been much speculation as to whether this variation represents distinct species or ecophenotypes. Recently however it has been conclusively demonstrated that these morphotypes indeed represent distinct biological species (Geisen et al. 2002; Saez et al. 2003). A detailed review of this species complex is presented by Quinn et al. (this volume) and for reasons of completeness we will summarise the available data here as well.

Quantitative and qualitative morphology

Three morphotypes (large, intermediate and small) were identified within extant populations of Calcidiscus, based on the size of coccoliths and coccospheres (for references see Quinn et al. this volume) (Plate 4, figs 1-3). The morphospace based on coccolith size measurements and element counts reveals a trimodal distribution, albeit with broadly overlapping margins and the modes compare well with the size range reported for Holocene material. It has, however, been pointed out that size measurements as a sole character are not sufficient to distinguish the morphotypes. If other, qualitative characters are added, the morphotypes become easily separable (Baumann and Sprengel 2000; Geisen et al. 2002; Kleijne 1993) (Plate 4, figs 4- 9). Research during CODENET has additionally shown that at least two of the morphotypes exhibit a stable morphology in culture under varying environmental conditions (Quinn et al., this volume Quinn et al. 2003).

Life-cycles and holococcolith phase differentiation

Observations on plankton samples and on cultured clones of Calcidiscus have demonstrated that the large and intermediate morphotype independently form life-cycle associations with distinctly different holococcolithophores (Plate 4, figs 10-12). In combination with the morphological observations this led Geisen et al. (2002) to amend the taxonomy, assigning the morphotypes to subspecies rank.

Molecular studies

The two subspecies identified by Geisen et al. (2002) can be distinguished using both the conservative IBS rDNA and the faster evolving tufA genes. With this information Saez et al. (2003) concluded that Geisen et al. (2002) were too conservative in using the intra-specific rank subspecies and have therefore raised the subspecies to species level. Furthermore they present evidence for two distinct 4 Discussion 145 genotypes within C. quadriperforatus and assume that this represents a case of recent cryptic spéciation. In the same study they calculate the divergence time for C. leptoporus (the intermediate “morphotype”) and C. quadriperforatus (the large “morphotype”) at 11.57 Ma (±1.61), which correlates well with the results of Knappertsbusch (2000) from the morphological classification of Calcidiscus in the fossil record.

Summary

Initial results from oceanographic and culture studies produced conflicting data. However, subsequent data from holococcolith morphology (Geisen et al. 2002) and a combined morphological and molecular genetic (Saez et al. 2003) study of a large collection of culture isolates proved that the variation was predominantly genotypic. This has provided the key for a reinterpretation of previous data with respect to the occurrence and ecological preferences of the species (Baumann and Sprengel 2000; Henderiks and Renaud 2001; Renaud and Klaas 2001). The molecular clock and findings from the fossil record point to a relatively deep divergence of the Calcidiscus species. An additional surprising result from the molecular studies was the discovery of recent cryptic spéciation within C. quadriperforatus. 4 Discussion 146

Plate 4 Calcidiscus spp. Fig. 1: Scanning electron micrograph of a C. quadriperforatus coccosphere. This species was previously known as the large morphotype of C. leptoporus. Water sample, western Pacific, Miyake-jima island, Miike Port, Japan. Fig. 2: Scanning electron micrograph of a C. leptoporus coccosphere. This species was pre­ viously known as the intermediate morphotype of C. leptoporus. Culture sample (NS 10-2), S. Atlantic, off South Africa, R/V Agulhas cruise MARE 2. Fig. 3: Scanning electron micrograph of a Calcidiscus sp. SMALL coccosphere. Note the kinked suture lines that can be traced into the central pore. This species was previously known as the small morphotype of C. leptoporus. As no holococcolithophore stage has been identified an informal classification is used. Water sample, S. Atlantic, off Namibia, R/V Meteor cruise M48-4, station 472. Fig. 4: Scanning electron micrograph of a C. quadriperforatus coccolith in distal view. Note the curved suture lines and the obscured zone around the central pore. Culture sample (ASM 27), western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2. Fig. 5: Scanning electron micrograph of C. leptoporus coccoliths in distal view. Note the curved suture lines that can be traced into the central pore. Water sample, western Pacific, Miyake-jima island, Chyotarou Port, Japan. Fig. 6: Scanning electron micrograph of Calcidiscus sp. SMALL coccolith in distal view. Note the kinked and ragged appearanee of the suture lines that can be traced into the central pore. Water sample, western Pacific, Miyake-jima island, Chyotarou Port, Japan. Fig. 7: Scanning electron micrograph of a C. quadriperforatus coccolith in proximal view. Water sample, western Pacific, Miyake-jima island, Ibo Port, Japan. Fig. 8: Scanning electron micrograph of C. leptoporus coccoliths in proximal view. Water sample, western Pacific, Miyake-jima island, Chyotarou Port, Japan. Fig. 9: Scanning electron micrograph of Calcidiscus sp. SMALL coccoliths in proximal view. Water sample, N. Atlantic, R/V Meteor cruise 38-1, station 12. Fig. 10: Scanning electron micrograph of the holococcolithophore stage of C. quadriperfo­ ratus. This stage was previously described as “Syracolithus quadriperforatus”. Water sample, N. Atlantic, off the Canary Islands, R/V Poseidon cruise P233B, station 2. Fig. 11: Scanning electron micrograph of the holococcolithophorid stage of C. leptoporus. This stage was previously described as '‘Crystallolithus rigidus". Culture sample (NS 10-2), S. Atlantic, off South Africa, R/V Agulhas cruise MARE 2. Fig. 12: Detail of a scanning electron micrograph of a combination coccosphere bearing eoc- coliths of both the holococcolithophore stage of C. quadriperforatus and the associated holo­ coccolithophore stage. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69.

Scale bars represent: Figs 1-3, 10: 5;

m

4 C quadiiperforatus coccoliths 5. C. leptoporus 6. C. sp small distal view coccoliths distal view coccolith distal view

7. C. quadriperforatus coccolith 8, C, leptoporus 9. C. sp, small coccoliths proximal proximal view coccoliths proximal view and lateral view

10. C. quadriperforatus U .C . leptoporus holococcoliths 12. C. quadriperforatus holococcolithophorid distal view detail of combination coccosphere 4 Discussion 148

Gephyrocapsa oceanica and related species (Plate 2)

The genus Gephyrocapsa constitutes a late Neogene group of heterococcolithophores that first developed in the late Pliocene, about 3.5 Ma ago {Reticulofenestra pseudoumbilicus Zone, NN 15), became dominant in the Early Pleistocene, and has yielded ecological dominance during the last 85 kyr to the closely allied descendant species Emiliania huxleyi. Because of the close environmental relationship of E. huxleyi with genus Gephyrocapsa, E. huxleyi has been regarded as a modem ecological surrogate for the small gephyrocapsids of the Pleistocene, which it has replaced in the modern phytoplankton (e.g. Gartner 1988). Phylogenetically E. huxleyi emerges from the Gephyrocapsa clade and hence needs to be considered within the group. Gephyrocapsa is a complex genus, which shows high morphological variability and has been intensively studied by palaeontologists in order to produce a high- resolution biostratigraphic subdivision of the Quaternary. Thus, there are many extant and extinct species classified under the genus Gephyrocapsa, and most of these species are defined solely by morphologic characters (e.g. Pujos 1985). According to this taxonomy the modem Gephyrocapsa assemblage consists of four species with first appearances between times of 1-3.5 Ma (Matsuoka and Okada 1990; Samtleben 1980; Young 1991). However, detailed morphometry, qualitative morphological, and biogeographical work indicates that each of these species consists of different morphotypes (Bollmann 1997; Bollmann et al. 1998; Cros 2002; Hagino and Okada 2001). Emiliania huxleyi is the most abundant living coccolithophore and appears to have extremely broad ecological affinities, occurring in all of the main oceanic conditions. Large-scale blooms which mainly consist of E. huxleyi, are regularly observed during early summer in the northem North Atlantic (e.g. Brown and Yoder 1993; Holligan et al. 1993). This species is easy to maintain in culture and has been subject to intensive studies, combining observations from culture and oceanic populations (e.g. Westbroek et al. 1993; Westbroek et al. 1994; Westbroek et al. 1989; Young 1994; Young and Westbroek 1991). In addition, Emiliania huxleyi has a very well constrained first occurrence at only 270 ka (Thierstein et al. 1977).

Quantitative and qualitative morphology

The genus Gephyrocapsa was first described by Kamptner (1943) and included only one species, G. oceanica, which he later divided into two varieties, G. oceanica var. typica and G. oceanica var. californiensis (Kamptner 1956). Much later, the first small Gephyrocapsa species, G. aperta, was described (Kamptner 1963). Since then more than 20 species have been described within this genus on the basis of various criteria. Many of these species have been distinguished using overall morphologic criteria detectable with LM such as size, proportion of the central area and bridge angle (e.g. Boudreaux and Hay 1969; Hay and Beaudry 1973). Some have been cited only once, whereas others, such as Gephyrocapsa 4 Discussion 149 reticulata (Nishida 1971), lack a crossbar and, therefore, cannot be assigned to the genus Gephyrocapsa. Whereas large species like G. oceanica. (Plate 2, figs 9, 10) and G. muellerae (Plate 2, figs 13, 14) are relatively easy to distinguish even with light microscopy it was only after the SEM replaced LM and trans electron microscopes that some characteristic small species such as G. ornata (Plate 2, figs 11, 12), G. ericsonii (Plate 2, figs 15, 16) or G. aperta were accurately defined. However, a major step forward in the classification of Gephyrocapsa species came when Bréhéret (1978) and Samtleben (1980) independently demonstrated the feasibility of distinguishing different morphological species by using simple quantifiable morphometrical characters (mainly coccolith size and bridge angle). Comparable size criteria have been applied to the entire Gephyrocapsa complex when Matsuoka and Okada (Matsuoka and Okada 1989, 1990) investigated time- progressive variations in the morphology of the genus. All extant species of the genus Gephyrocapsa build spherical to subspherical coccospheres of oval placoliths with a diagonal bridge crossing the central area (Plate 2, figs 9-16). This distinctive feature makes even the smallest specimens identifiable at the generic level. Many variations of the coccolith bridge exist. Some have a high-arched bridge, whereas others have a low-profile bridge formed by fine rod-like elements. Recently, morphological analysis of Holocene Gephyrocapsa assemblages revealed six dominant morphological associations (Bollmann 1997), which the author described informally, since he reserved judgment on whether these corresponded to discrete species or, in part, to ecophenotypes. At least, they appear to have distinct environmental preferences with respect to temperature and productivity. During CODENET only G. oceanica was isolated and studied in detail in culture. Data emerging from there demonstrate a strong stability of morphological characters within the species. Emiliania huxleyi (Lohmann 1902), Hay and Mohler, 1967 in Hay et al., 1967 forms spherical coccospheres consisting of fewer than 10 to up to more than 50 partially interlocking placoliths (Plate 2, figs 17-20). These oval placoliths typically are formed of T-shaped elements and have an elliptical central area. Young and Westbroek (1991) distinguished four varieties of E. huxleyi - types A (Plate 2, figs 17, 20), B, C (Plate 2, fig. 19), and var. corona (Plate , fig. 18) - although they stated that biometrical analyses do not separate these types easily. Culture studies were limited because we failed to isolate most of the morphotypes, though those that were cultured (types A & C) showed strong morphological stability, with no transitions into other morphotypes.

Molecular studies

To date E. huxleyi has been genetically characterised from a number of studies. Medlin et al. (1996) have sequenced a number of clonal strains from all oceans which were identical with regard to the chlorplasticl6S and the nuclear 18S rDNA and the spacer region between the Rubisco rbcL and rbcS. However, genetic 4 Discussion 150 fingerprinting techniques such as RAPDs and AFLPs (for a review refer to Mueller and LaReesa Wolfenbarger 1999) revealed significant genetic differences between strains. Using RAPD Medlin et al. (1996) found that all strains except for one were genetically distinct and that this genetic diversity is reflected in the morphology and the ecological distribution of the strains. On the basis of the RAPD study by Medlin et al. (1996) the proposed variants by Young and Westbroek (1991) were formally emended. Recently Iglesias-Rodrfguez et al. (2002a; 2002b) further demonstrated a high degree of polymorphism in isolates of E. huxleyi of different geographical origin using AFLP and microsatellite loci. Gene flow and calculations of genetic diversity using population statistics are presently under way. All of these studies predict that E. huxleyi must be undergoing frequent sexual recombination to maintain such high diversity among populations reproducing vegetatively to maintain high biomass. In CODENET a number of E. huxleyi and G. oceanica and an additional Emiliania morphotype Emiliania huxleyi type R have been sequenced (Saez et al. unpublished). With the slowly evolving 18S rDNA all the sequences obtained are identical, which is consistent with a recent divergence of these genotypes. But isolates, either from the same morphological species or not, appear genetically distinct at the fast evolving tufA gene. Unlike the case of C. pelagicus or C. leptoporus, where each morphotype was monophyletic, E. huxleyi and G. oceanica species and morphotypes were mixed in relation to the tufA genotypes. These unexpected findings have been interpreted to be the result of ancient (or shared) polymorphisms at the tufA gene, which have persisted through the spéciation events in these different lineages. Thus future work regarding the genetic variability of the species in correlation to their ecological preferences remains an interesting topic and is likely to be solved using DNA fingerprinting methods, such as microsatellites or SNPs analyses. Due to the failure to isolate other Gephyrocapsa and Emiliania species, information about their genetic variability is still lacking.

Life-cycles and holococcolith phase differentiation

The life-cycle of E. huxleyi consists of coccolith bearing non-motile, diploid C cells and motile, haploid, scale-bearing S cells (see Billard, this volume). Unlike in other coccolithophores, where life-cycles have been demonstrated, the S cells do not produce holococcoliths. A third type of cells - the naked N cells - is diploid and does not appear to be part of the haplo-diplontic life-cycle, but a mutant. All cell types are capable of indefinite asexual reproduction by binary fission. Few life­ cycle observations are available for Gephyrocapsa spp. but as in Emiliania spp. the haploid stage is motile and covered with unmineralized scales (Probert unpubl. data). It has been hypothesised that sexual recombination can, however, lead to consolidating reproductive barriers between closely related species and hence can explain the diversity of the group (Paasche 2002). 4 D iscussion 151

Ecological / biogeographical separation

Measuring the reproduction rates of a large number of E. huxleyi and G. oceanica clones under the same environmental conditions, Brand (1981; 1982) has demonstrated both the stability of this parameter in single clones, but also considerable variation among clones. He therefore concluded that a natural population is not clonal but consists of a mixture of genotypes with different reproductive potentials (Brand 1982). However he did not claim that this differentiation could only be due to the break of reproductive barriers: “these species (...) either undergo extensive genetic recombination with the resulting genotypes having different reproductive potentials or exist as complexes of coexisting clonal lines” (Brand 1982). Paasche (2002) points out the relative tolerance of coastal clones of E. huxleyi to salinity variations, contrasted by less tolerant oceanic E. huxleyi clones. Young and Westbroek (1991) describe E. huxleyi type A, B and C as well as the variant corona but they reserve judgement whether these are typical of distinct environments. Emiliania appears to have diverged into at least five discrete sub-species with partially overlapping biogeographies (Findlay and Giraudeau 2001; Medlin et al. 1996; Young and Westbroek 1991). In the genus Gephyrocapsa culture studies were limited because most of the recognized species were not isolated into culture. However, the hypothesis of differential ecologies of the species has been tested in detail through derivation of a temperature transfer function based on the distribution of Gephyrocapsa species (Bollmann et al. 2002).

Summary

In the case of E. huxleyi there is evidence of different levels of genotypic variation. Firstly, there is the well-documented divergence of genus Emiliania at about 270 ka, which has already diverged into a number of well-defined species, detectable by molecular and morphological methods. Secondly, there is evidence of a high genotypic variability on the population scale. It has been demonstrated that a population is not a single clone or genotype, but a mixture of genetically distinct clones, which are attuned to yield maximum growth rates in a range of environmental conditions. Recent molecular studies using microsatellite loci have confirmed that populations of E. huxleyi - separated by major oceanic boundaries - show a distinct genetical fingerprint and can thus be spatially separated. The overall pattern of variation in Gephyrocapsa is comparable with Emiliania. A relatively young genus has diverged into a number of well-defined morphological species (G. muellerae, G. ericsonii, G. oceanica, G. ornata). However, different clones of G. oceanica have been tested for their environmental preferences and they reveal a genotypic variability of the same nature as in Emiliania (Probert, unpubl. data). Furthermore Emiliania and Gephyrocapsa species play an important role in geochemical cycles and knowledge of both intra- and inter-specific genotypic variation is crucial to determine which species are the most important actors in. 4 Discussion 152 for example, the carbon cycle. Therefore, the development of specific markers to map the spatial and temporal variability of both E. huxleyi and Gephyrocapsa spp. remains an important target for future research.

Syracosphaera pulchra (Plate 5)

Syracosphaera pulchra is the most common member of the very diverse extant genus Syracosphaera and the only one to have been successfully isolated into culture (Inouye and Pienaar 1988). It occurs globally in all oceans (Okada and Honjo 1973; Okada and McIntyre 1977; Winter et al. 1994) and is common in sediments. Its first occurrence in the fossil record is in the early Pliocene at 4.8 Ma (our data based on analysis of the NEPTUNE database, GDP site 664 and the GDP reference slide collection). As S. pulchra coccospheres and coccoliths are very rich in morphological characters the species was regarded as very well defined and was selected as a key species for the CGDENET project, in part to act as a control species to quantify the degree of intra-specific variation. The recent discovery of "pseudo-" cryptic species in S. pulchra (Geisen et al. 2002) however has shed a new light on this interpretation.

Quantitative and qualitative morphology

Coccospheres of the heterococcolith stage of S. pulchra are spherical to ellipsoidal and typically pear-shaped, a feature which is shared with the two holococcolithophores. The coccospheres are normally dithecate with dimorphic endothecal caneoliths (muroliths) (Plate 5, figs 1, 3). The cell bears two flagellae and a haptonema and the flagellar opening is surrounded by modified coccoliths possessing a spine which is forked at the end (Plate 5, figs 1, 2, 4). The elliptical body caneoliths have a corrugated wall with three external flanges (Plate 5, fig. 3). The central area is formed of numerous radial laths that extend towards the centre of the coccolith and are partly joining (Plate 5, figs 3, 4). The monomorphic dome­ shaped exothecal coccoliths are elliptical with a narrow central depression and slitted walls (Plate 5, fig. 5). The heterococcolithophore species Syracosphaera pulchra contains two chloroplasts and exhibits two flagella. For quantitative morphological analysis of the heterococcolith stage simple size measurements on endothecal coccoliths have been performed. Data shows a strong bi- to multimodality in samples from the fossil record and sediment trap material (Geisen et al. 2002). It has been speculated whether this apparent variation is due to genotypic or ecophenotypic variability. As cultures became available, this hypothesis was tested and cultures were grown under a range of environmental parameters. Morphometrical analyses performed on clonal cultures, however, revealed a strong unimodality, and no temperature related size variation was observed, indicating that the morphology is under strong genotypic control. Following the discovery of discrete "pseudo”-cryptic species within S. pulchra 4 Discussion 153 based on holococcolith morphology a re-examination of the culture material revealed a slight variation in heterococcolith size and it has been hypothesised that this variation is species diagnostic (Geisen et al. 2(X)2). Molecular techniques are being used to test this hypothesis (see paragraph on molecular biology). Unlike in other species discussed in this paper the rich set of qualitative characters has been of limited use and characters allowing for the discrimination of the two cryptic species have not yet been identified.

Life-cycles and holococcolith phase differentiation

It has recently been demonstrated that the heterococcolithophore S. pulchra independently forms life-cycle associations with two holococcolithophores previously assigned to different genera (for a review refer to Cros et al. 2000; Geisen et al. 2002). Data here stem from observations both from a phase change observed in culture and from combination cells from water samples. Geisen et al. (2002) have concluded that this is an example of “pseudo”-cryptic spéciation where morphological separation between the species is only visible in the haploid, holococcolith bearing stage. Although they conclude that S. pulchra comprises two biological species it is currently impossible to separate them in the heterococcolith phase and so they introduced the informal terms S. pulchra pirus-type (Plate 5, figs 6-8) and S. pulchra oblonga-type (Plate 5, figs 9-11) until the heterococcolith phase can be discriminated. Re-examination of the morphometrical data from water samples and culture samples suggests that there may be a slight differentiation in the mean size of the endothecal heterococcoliths. Testing of the predictions from this, using molecular genetics, is in progress.

Ecological / biogeographical and molecular separation

The co-occurrence of the two holococcolithophores associated with S. pulchra in the same natural water samples has led to speculation about the nature of spéciation and - comparable with the case of Emiliania - an ecological spéciation has been hypothesised. Until now, however, the two Syracosphaera species can only be discriminated in their holococcolithophore stage and there is few information on their biogeographical and ecological ranges. 13 clonal strains of 5. pulchra are currently under molecular investigation to test the hypothesis of a slight differentiation in the mean size of the heterococcoliths between the two potential species. The IBS rDNA genes are strictly identical between the strains, but the first tufA DNA sequences show the presence of at least two different types. As for the other species described in this chapter, a comparison between genetic and morphological data will provide a powerful tool to discriminate which subtle morphological character(s) may allow distinction at the species level. 4 Discussion 154

Summary

In this case, heterococcolith morphology is remarkably complex and stable and S. pulchra had been regarded as a particularly well-defined species. Detailed study of geological and oceanographic samples however yielded more complex and variable morphological patterns than predicted, but these were initially interpreted as essentially noise, i.e. random population-level variation. However, our observations of holococcolith-heterococcolith combination coccospheres and phase changes in cultured strains have indicated strong morphological differentiation in the haploid phase holococcoliths (Geisen et al. 2002). This discovery of cryptic species in S. pulchra has now severely challenged the interpretation of S. pulchra as a single species with a global occurrence. Until now any testing of the predictions arising from this has been critically dependent on the presence of the relatively rare holococcolithophores and hence biogeographical mapping of the two holococcolithophores involved remains an important target for future research. A further opportunity to solve this challenging problem will be the use of genetic markers to discriminate the two species. 4 Discussion 155

Plate 5 Syracosphaera spp. Fig. 1: Scanning electron micrograph of a S. pulchra coccosphere. This typical specimen displays endothecal and exothecal coccoliths. Coccoliths surrounding the flagellar pole are spine bearing. Water sample, N. Atlantic, off the Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 2: Scanning electron micrograph of a S. pulchra coccosphere. Note the lack of exothecal coccoliths. Water sample, N. Atlantic, R/V Meteor cruise 42-4B, station US IB. Fig. 3: Scanning electron micrograph of S. pulchra endothecal coccoliths. The specimen in lateral view shows the typical wall structure with three flanges. Note the lack of exothecal coccoliths. Water sample, N. Atlantic, off Canary Islands, R/V M eteor cruise 42-4B, station US IB. Fig. 4: Scanning electron micrograph of a S. pulchra circumflagellar endothecal coccolith. Note the typical central spine with forked end. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 5: Scanning electron micrograph of a S. pulchra exothecal coccolith. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 6: Scanning electron micrograph of the holococcolithophorid stage of 5. pulchra. This stage was previously described as “Daktylethra pirus” and is referred to as S. pulchra HO pirus-type. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 7: Scanning electron micrograph of a S. pulchra HO pirus-type circumflagellar holococ­ colith in lateral view. The circumflagellar coccoliths typically have a pointed hood. Note the clear offset between the hood and the base and the presence of perforations in the hood. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 8: Scanning electron micrograph of a S. pulchra HO pirus-type holococcoliths. The cir­ cumflagellar coccoliths typically have a pointed hood. Water sample, N. Atlantic, off Canary Islands, R/V Meteor cruise 42-4B, station US IB. Fig. 9: Scanning electron micrograph of the holococcolithophorid stage of S. pulchra. This stage was previously described as “Calyptrosphaera oblonga” and is referred to as S. pulchra HO oblonga-type. Water sample, N. Atlantic, off the Canary Islands, R/V Poseidon cruise P233B, station 3. Fig. 10: Scanning electron micrograph of a S. pulchra HO oblonga-type holococcolith. Note the hexagonal arrangement of the calcite rhombohedra and the absence of an offset between the hood and the base. Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69. Fig. 11: Scanning electron micrograph of a S. pulchra HO oblonga-type circumflagellar holococcoliths. The circumflagellar coccoliths typically have a pointed hood Water sample, western Mediterranean, Alboran Sea, R/V Hesperides cruise MATER 2, station 69.

Scale bars represent: Figs 1, 2,9: 5pm ,fig. 6: 10/

7

3. S. pulchra endothecal coccoliths 4. S. pulchra arcumflagellar 5. S. pulchra exothecal coccolith In distal and sideview endothecal coccolith

6. S. pulchra holococcolithophorid 7. S. pulchra circumflagellar 8 S pulchra pirus-type holococcolith pirus-type holococcoliths pirus-type

10. S. pulchra holococcolith oblonga-type

U .S . pulchra circumflagellar holococcoliths oblonga-iypa 9. S. pulchra holococcolithophorid oblonga-type 4 Discussion 157

Synthesis

Spéciation and divergences (Fig. 1)

In a few cases our recent research has shown that fine-scale morphovariants of classic species in fact represent fully isolated species, which have diverged in the Pliocene (2-5 Ma) or earlier. For instance Umbilicosphaera sibogae has conventionally been subdivided into two varieties with alternative hypotheses that they represented ecotypes, life-cycle stages or closely related genotypes (Inouye and Pienaar 1984). Molecular evidence has clearly shown that they have been genotypically independent for more than 5 Ma according to the molecular clock of Saez et al. (2003). Similarly, in modem Calcidiscus three morphotypes have been tentatively distinguished based on size variation and subtle morphological differences (Kleijne 1991, 1993; Knappertsbusch 1997), but there has been uncertainty as to whether they intergrade or are discrete and hence whether they are ecophenotypes (Knappertsbusch 1997; Renaud and Klaas 2001). Our new data from geological studies, life-cycle stages, culture studies and molecular genetic work all indicate that they in fact represent discrete species that probably diverged in the Middle Miocene (Geisen et al. 2002; Renaud and Klaas 2001; Renaud et al. 2002; Saez et al. 2003). Those results suggest that morphological differentiation can be strongly uncoupled from genetic divergences (see de Vargas et al., this volume, for more discussion on this aspect). In contrast to the successful - on the range of several Ma - species discussed above a second set of studied cases reveals another level of genotypical variability equally coupled with slight morphological variation. Amongst the species studied here Emiliania huxleyi, Gephyrocapsa spp., Coccolithus spp., Calcidiscus spp. (see to Quinn et al., this volume for a discussion on cryptic spéciation in the large Calcidiscus quadriperforatus) and Syracosphaera pulchra reveal a similar pattern of variation, but there is tentative evidence for much more recent divergences. These results indicate that, unless coccolithophores are undergoing a remarkable radiative evolution, evolution is a dynamic process continuously producing and eliminating species. This process can be studied by a combination of morphological and molecular genetic research (see fig. 4 of de Vargas et al., this volume).

Local adaptation - ecological spéciation?

Paasche (2002) synthesised a range of evidence to suggest that globally distributed species, such as E. huxleyi should be regarded as mosaics of locally adapted populations. Evidence includes: (a) genetic fingerprinting data for high levels of genetic recombination, within a haplo-diplontic life-cycle; (b) physiological experiments indicating genetic differentiation within populations and significant variation between environments (Brand 1981, 1982; Fisher and Honjo 1989; Paasche 2002; Paasche et al. 1996; Young and Westbroek 1991); (c) the very 4 Discussion 158

ï 5 2 81 5Ï5 Coccolithus '" 1 1 1 II ll î pelagicus 1.5-2.5 Ma

m m

Calcidiscus leptoporus 10-13 Ma (iarge/fntm)

NB Holococcolith Umbilicosphaera phases have not been sibogae ‘^ 1 Kjentffied but are likely 4.5-7 Ma

Helicosphaera carteri 12-8 Ma (cart/hyal)

Syracosphaera NB In this separation is cryptic pulchra the helerococc phase but clear holococcolith phase 0-4.5 Ma

simplification Gephyrocapsa there are at least 5 sub-types in these oceanica-muellerae two sp e cie s plus 3 group the sister species £. huxleyi "I il il li il

Fig. 1. Outline of intra-specific variation in the CODENET taxa. Heterococcolith phases are illustrated above the holococcolith phases. Boxes indicate the type and the quality of avail­ able evidence to support the interpretation. Black boxes - strong data, grey boxes - weak data, white boxes - no data. 4 Discussion 159 broad distribution of such species, occupying improbably wide ranges of habitats, contrasting with narrower ecological tolerances for individual culture isolates. These data suggest local adaptation as a possible origin of the sub-species level variation. We have demonstrated in this review that the same level of genotypic variation applies not only to E. huxleyi, but also to a range of species studied in detail, including Coccolithus spp., Calcidiscus spp. and Syracosphaera pulchra. We suggest that this apparent local specialization may be a key factor for spéciation in the coccolithophores, and possibly more generally for the evolutionary success of marine planktonic organisms.

Final remarks

These patterns arguably intergrade - indeed it is often difficult to determine which pattern applies in a particular case. Hence a possible model is that local ecological adaptation leads to continuous evolution of new geographically restricted genetic varieties, which in certain cases differ sufficiently to form discrete sub-species that disperse globally into similar ecological environments. A constant turnover of such sub-species may occur, possibly because of environmental change causing shifts in the extent of the ecological conditions to which they are adapted. If particular sub­ species differentiate sufficiently both ecologically and genotypically, then they may diverge into discrete biological species. A key factor to elucidate the underlying mechanisms of this hypothetical pattern of evolution is the understanding of the coccolithophore life-cycle and reproductive strategy. The presence of chloroplasts provides the algae with a seemingly unlimited source of energy. Unlike in other species however the energy produced is not stored in carbohydrates and fatty acids, but seems to maintain - in the presence of sufficient nutrients - a high biomass level by asexual reproduction (Smetacek 2001). This strategy might well be described as protection by the sheer outnumbering of possible predators, an extreme example being a phytoplankton bloom. The apparent decoupling of sexual recombination of genes and reproduction by mitotic fission allows for both rapid (local) spéciation and maintenance of high cell densities. Hence the ability to exchange genes within a population is an important - and long overlooked - tool for adaptation and spéciation especially for planktonic organisms competing in a rapidly changing environment. As a logical consequence of this strategy it can be hypothesised that coccolithophores, which have lost or not evolved calcification in the haploid stage, might have an evolutionary advantage as this part of the life-cycle only occurs occasionally. Unlike in land plants, where the skeleton serves as a supporting structure in the competition of the chloroplasts for light (Smetacek 2001), this function seems unlikely for marine planktonic algae living in the absence of gravity. Hence mechanical protection against viruses and bacteria trying to enter the cell and against grazers with their mainly acidic stomachs (e.g. copepods) can be hypothesized as a function of coccoliths. However, protection for the relatively short time spent in the haploid stage is not needed and more energy is thus available for reproduction via mitosis. Important future 4 Discussion ] 60

research tasks will therefore focus on life-cycles and reproductive strategies of coccolithophores and will include the identification of possibly ecological triggers to induce phase changes. Coccolithophores appear to show evidence for both classic models of evolution - phyletic gradualism and punctuated equilibria (allopatric spéciation) (Benton and Pearson 2001 ; Eldredge 1971 ; Eldredge and Gould 1972; Gould and Eldredge 1977; Pearson 1993; Young and Bown 1994). We have demonstrated here that combining different strands of research can enable us to acquire detailed information on coccolithophore diversity and evolution and to gain further under-standing of the underlying processes.

Acknowledgements

This work is a contribution to the EU-funded TMR network CODENET (ERBFMRX CT97 0113) which funded IP, AGS and MG. Further funding was provided by the EU project Ironages (EVK2-CT-1999-00031), which supported MG. We would like to thank J. Giraudeau and an anonymous reviewer for their valuable comments which helped improve the quality of this manuscript.

References

Baumann K-H, Andruleit H, Schroder-Ritzrau A, Samtleben C (1997) Spatial and temporal dynamics of coccolithophore communities during low production phases in the Norwegian-Greenland Sea. In: Hass H C, Kaminski M A (eds) Contributions to the Micropaleontology and Paleoceanography of the Northern North Atlantic. Grzybowski Foundation, pp 227-243 Baumann K-H, Sprengel C (2000) Morphological variations of selected coccolith species in a sediment trap north of the Canary Islands. J nannoplankton Res 22: 185-193 Baumann K-H, Young J R, Cachao M, Ziveri P (2000) Biometric study of Coccolithus pelagicu s and its palaeoenvironmental utility. J nannoplankton Res 22: 82 Benton M J, Pearson P N (2001) Spéciation in the fossil record. Trends Ecol Evol 16: 405- 411 Bollmann J (1997) Morphology and biogeography of Gephyrocapsa coccoliths in Holocene sediments. Mar Micropaleontol 29: 319-350 Bollmann J, Baumann K-H, Thierstein H R (1998) Global dominance of Gephyrocapsa coccoliths in the late Pleistocene: Selective dissolution, evolution, or global environmental change? Paleoceanography 13: 517-529 Bollmann J, Henderiks J, Brabec B (2002) Global calibration of Gephyrocapsa coccolith abundance in Holocene sediments for paleoteraperature assessment. Paleoceanography 17: 7-1-7-9 Boudreaux J E, Hay W W (1969) Calcareous nannoplankton and biostratigraphy of the late Pliocene-Pleistocene-Recent sediments in the Submarex cores. Revta Esp Micropaleont 1: 249-292 4 D iscussion 161

Brand L E (1981) Genetic variability in reproduction rates in marine phytoplankton populations. Evolution 38: 1117-1127 Brand L E (1982) Genetic variability and spatial patterns of genetic differentiation in the reproductive rates of the marine coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica. Limnol Oceanogr 27: 236-245 Bréhéret J G (1978) Formes nouvelles quaternaires et actuelles de la famille des Gephyrocapsaceae (Cocolithophorides). C R Acad Sci Série D 287: 447-449 Brown C W, Yoder J A (1993) Blooms of Emiliania huxleyi (Prymnesiophyceae) in surface waters of the Nova Scotian Shelf and the Grand Bank. J Plankton Res 15: 1429-1438 Cachao M, Moita M T (2000) Coccolithus pelagicus, a productivity proxy related to moderate fronts off Western Iberia. Mar Micropaleontol 39: 131-155 Chepstow-Lusty A, Shackleton N J, Backman J (1992) Upper Pliocene Discoaster abundance variations from the Atlantic, Pacific and Indian Oceans: the significance of productivity pressure at low latitudes. Memorie Sci geol 43: 357-373. Cros L (2002) Planktic coccolithophores of the NW Mediterranean. PhD thesis, Universitat de Barcelona (ISBN 84-475-2680-1), p. 181 Cros L, Fortuno J-M (2002) Atlas of Northwestern Mediterranean Coccolithophores. Scientia Marina 66: 7-182 Cros L, Kleijne A, Zeltner A, Billard C, Young J R (2000) New examples of holococcolith- heterococcolith combination coccospheres and their implications for coccolithophorid biology. Mar Micropaleontol 39: 1-34 de Vargas C, Norris R, Zaninetti L, Gibb S W, Pawlowski J (1999) Molecular evidence of cryptic spéciation in planktonic foraminifers and their relation to oceanic provinces. Proc natn Acad Sci USA 96: 2864-2868 Edvardsen B, Eikrem W, Green J C, Andersen R A, Yeo Moon-van der Staay S, Medlin L K (2000) Phylogenetic reconstructions of the Haptophyta inferred from 18S ribosomal DNA sequences and available morphological data. Phycologia 39: 19-35 Eldredge N (1971 ) The allopatric model and phylogeny in Paleozoic invertebrates. Evolution 25:156-167 Eldredge N, Gould S J (1972) Punctuated equilibria: an alternative to phyletic gradualism. In: Schopf T J M (ed) Models in Paleobiology. Freeman, Cooper, and Co., San Francisco, pp 82-115 Findlay C S, Giraudeau J (2001) Extant calcareous nannoplankton in the Australian sector of the Southern Ocean (austral summers 1994 and 1995). Mar Micropaleontol 40: 417- 439 Fisher N S, Honjo S (1989) Intraspecific differences in temperature and salinity responses in the coccolithophore Emiliania huxleyi. Biol Oceanogr 6: 355-361 Fresnel J (1994) A heteromorphic life cycle in two coastal coccolithophorids, Hymenomonas lacuna and Hymenomonas coronata (Prymnesiophyceae). Can J Bot 72: 1455-1462 Fujiwara S, Tsuzuki M, Kawachi M, Minaka N, Inouye I (2001) Molecular phylogeny of the haptophyta based on the rbcL gene and sequence variation in the spacer region of the RUBISCO operon. J Phycol 37: 121-129 Gartner S (1988) Paleoceanography of the Mid-Pleistocene. Mar Micropaleontol 13: 23-46 Gayral P, Fresnel J (1983) Description, sexualité et cycle développement d'une nouvelle coccolithophoracée (Prymnesiophiceae): Pleurochrysis pseudoroscoffensis sp. nov. Protistologica 19: 245-261 Geisen M (2003) Studies on the species level variation of selected coccolithophores. PhD thesis. University College London, p. 199 4 Discussion 162

Geisen M, Billard C, Broerse A T C, Cros L, Probert I, Young J R (2002) Life-cycle associations involving pairs of holococcolithophorid species: intraspecific variation or cryptic spéciation? Eur J Phycol 37: 531-550 Geitzenauer K R, Roche M B, McIntyre A (1977) Coccolith biogeography from North Atlantic and Pacific surface sediments. Acad. Press, London, New York, San Francisco Gould S J, Eldredge N (1977) Punctuated equilibria: the tempo and mode of evolution reconsidered. Paleobiology 3: 115-151 Hagino K, Okada H (2001) Morphological observations of living Gephyrocapsa crassipons. J nannoplankton Res 23: 3-7 Hay W W, Beaudry F M (1973) Calcareous nannofossils - Leg 15. In: Edgar N T, Saunders J B (eds) Init Repts DSDP, Leg 15, pp 625-684 Henderiks J, Renaud S (2001) Coccolith size increase of Calcidiscus leptoporus off­ shore Morocco during the last glacial maximum: An expression of enhanced glacial productivity? In: Henderiks J (ed) Coccolith studies in the Canary Basin: Glacial-inter- glacial paleoceanography of the eastern boundary current system PhD thesis, ETH- Zurich Holligan P M, Fernandez J A, Balch W M, Boyd P, Burkill P H, Finch M, Groom S B, Malin G, Müller K, Purdie D, A., Robinson C, Trees C C, Turner S M, Wal P v d (1993) A biogeochemical study of the coccolithophore, Emiliania huxleyi, in the North Atlantic. Global Biochemical Cycles 7: 879-900 Iglesias-Rodrfguez M D, Brown C W, Doney S C, Kleypas J A, Kolber D, Kolber Z, Hayes P K, Falkowski P G (2002a) Representing key functional groups in ocean carbon cycle models: Coccolithophorids. Global Biochemical Cycles 16: 47/41-47/20 Iglesias-Rodrfguez M D, Saez A G, Groben R, Edwards K J, Batley J, Medlin L K, Hayes P K (2002b) Polymorphic microsatellite loci in global populations of the marine coccolithophorid Emiliania huxleyi. Molecular Ecology Notes 2: 495-497 Inouye I, Pienaar R N (1984) New observations on the coccolithophorid Umbilicosphaera sibogae van foliosa (Prymnesiophyceae) with reference to cell covering, cell structure and flagellar apparatus. Br phycol J 19: 357-369 Inouye I, Pienaar R N (1988) Light and electron microscope observations of the type species o i Syracosphaera, S. pulchra (Prymnesiophyceae). Br phycol J 23: 205-217 Jafar S A, Martini E (1975) On the validity of the calcareous nannoplankton genus Helicosphaera. Senckenberg leth 56: 381-397 Jordan R W, Chamberlain A H L (1997) Biodiversity among haptophyte algae. Biodiversity Conserv 6: 131-152 Jordan R W, Green J C (1994) A check-list of the extant haptophyta of the world. J mar biol A ss U K 74: 149-174 Jordan R W, Kleijne A (1994) A classification system for living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 83-105 Jordan R W, Young J R (1990) Proposed changes to the classification system of living Coccolithophorids. Int Nannoplankton Assoc Newsl 1: 15-18 Kamptner E (1927) Beitrag zur Kenntnis adriatischer Coccolithophoriden. Arch Protistenkd 58: 173-184 Kamptner E (1943) Zur Revision der Coccolithineen-Spezies Pontosphaera huxleyi Lohm. Anz Akad Wiss Wien 80: 73-49 Kamptner E (1956) Das Kalkskelett von Coccolithus huxleyi (Lohmann) Kamptner und Gephyrocapsa oceanica Kamptner (Coccolithineae). Arch Protistenkd 101: 171-202 4 Discussion 163

Kamptner E (1963) Coccolithineen-Skelettreste aus Tiefseeablagerungen des Pazifischen Ozeans. Annln naturh Mus Wien 66; 139-204 Kleijne A (1991) Holococcolithophorids from the Indian Ocean, Red Sea, Mediterranean Sea and North Atlantic Ocean. Mar Micropaleontol 17; 1-76 Kleijne A (1993) Morphology, taxonomy and distribution of extant coccolithophorids (Calcareous nannoplankton). PhD thesis. Free University Amsterdam (ISBN 90- 9006161-4), p. 321 Knappertsbusch M (1997) Morphologic variability of the coccolithophorid Calcidiscus leptoporus in the plankton, surface sediments and from the Early Pleistocene. Mar Micropaleontol 30; 293-317 Knappertsbusch M (2000) Morphological evolution of the coccolithophorid Calcidiscus leptoporus from the Early Miocene to recent. J Paleontol 74; 712-730 Knowlton N (1993) Sibling species in the sea. A rev ecol syst 24; 189-216 Loeblich A R, Tappan H (1966) Annotated index and bibliography of the calcareous nannoplankton. Phycologia 5; 81-216 Lohmann H (1902) Die Coccolithophoridae, eine Monographie der Coccolithen bildenden Flagellaten, zugleich ein Beitrag zur Kenntnis des Mittelmeerauftriebs. Arch Protistenkd 1; 89-165 Matsuoka H, Okada H (1989) Quantitative Analysis of Quaternary Nannoplankton in the Subtropical Northwestern Pacific Ocean. Mar Micropaleontol 14; 97-118 Matsuoka H, Okada H (1990) Time-progressive morphometric changes of the genus Gephyrocapsa in the Quaternary sequence of the tropical Indian Ocean, Site 709. In; Duncan R A, Backman J, Peterson L C (eds) Proc ODP, Sci Results, College Station, pp 255-270 McIntyre A, Bé A W H (1967) Modem Coccolithophoridae of the Atlantic Ocean - I. Placoliths and Cyrtholiths. Deep-Sea Res 14; 561-597 Medlin L K, Barker G L A, Green J C, Hayes D E, Marie D, Wrieden S, Vaulot D (1996) Genetic characterization of Emiliania huxleyi (Haptophyta). J Mar Syst 9; 13-32 Medlin L K, Kooistra W H C F, Potter D, Saunders J B, Andersen R A (1997) Phylogenetic relationships of the "golden algae" (haptophytes, heterokont chromophytes) and their plastids. PI Syst Evol Suppl 11 ; 187-219 Mueller U G, LaReesa Wolfenbarger L (1999) AFLP genotyping and fingerprinting. Trends Ecol Evol 14; 389-394 Nishida S (1971) Nannofossils from Japan IV. Calcareous nannoplankton fossils from the Tonohama Group, Shikoki, southwest Japan. Trans Proc palaeont Soc Japan 83; 143- 161 Nishida S (1979) Atlas of Pacific Nannoplanktons. News Osaka Micropaleontol Special Paper; 1-31 Okada H, Honjo S (1973) The distribution of oceanic coccolithophorids in the Pacific. Deep- Sea Res 20;355-374 Okada H, McIntyre A (1977) Modem coccolithophores of the Pacific and North Atlantic Oceans. Micropaleontology 23; 1-55 Okada H, McIntyre A (1979) Seasonal distribution of modem coccolithophores in the Westem North Atlantic Ocean. Mar Biol Berlin 54; 319-328 Paasche E (2002) A review of the coccolithophorid Emiliania huxleyi (Prymnesiophyceae), with particular reference to growth, coccolith formation, and calcification-photosynthesis interactions. Phycologia 40; 503-529 4 Discussion 164

Paasche E, Brubak S, Skattebol S, Young J R, Green J C (1996) Growth and calcification in the coccolithophorid Emiliania huxleyi (Haptophyceae) at low salinities. Phycologia 35: 394-403 Parke M, Adams I (1960) The motile (Crystallolithus hyalinus Gaarder & Markali) and non- motile phases in the life history of Coccolithus pelagicus (Wallich) Schiller. J mar biol Ass U K 39: 263-274 Pearson P N (1993) A lineage phylogeny for the Paleogene planktonic foraminifera. Micropaleontology 39: 193-232 Perch-Nielsen K (1985) Cenozoic calcareous nannofossils. In: Bolli H M, Saunders J B, Perch-Nielsen K (eds) Plankton Stratigraphy. Cambridge University Press, Cambridge, pp 427-555 Pujos A (1985) Quaternary nannofossils from the Goban Spur, eastern North Atlantic Ocean DSDP Holes 548-549A. Initial Rep deep Sea Drilling Proj U.S. Government Printing Office, Washington, pp 767-792 Quinn P, Thierstein H R, Brand LE, Winter A (2003) Experimental evidence for the species character of Calcidiscus leptoporus morphotypes. J Paleontol Rayns D G (1962) Alternation of generations in a coccolithophorid, Cricosphaera carterae (Braarud & Fragerl.) Braarud. J mar biol Ass U K 42: 481-484 Renaud S, Klaas C (2001) Seasonal variations in the morphology of the coccolithophore Calcidiscus leptoporus off Bermuda (N. Atlantic). J Plankton Res 23: 779-795 Renaud S, Ziveri P, Broerse A T C (2002) Geographical and seasonal differences in morphology and dynamics of the coccolithophore Calcidiscus leptoporus. Mar Micropaleontol 46: 363-385 Rowson J D, Leadbeater B S C , Green J C (1986) Calcium carbonate deposition in the motile {Crystallolithus) phase o f Coccolithus pelagicus (Prymnesiophyceae). Br phycol J 21: 359-370 Sâez A G, Probert I, Geisen M, Quinn P, Young J R, Medlin L K (2003) Pseudo-cryptic spéciation in coccolithophores. Proc natn Acad Sci USA 100: 7163-7168 Samtleben C (1980) Die Evolution der Coccolithophoriden-Gattung Gephyrocapsa nach Befunden im Atlantik. Palaeont Z 54: 91-127 Samtleben C, Bickert T (1990) Coccoliths in sediment traps from the Norwegian Sea. Mar Micropaleontol 16: 36-64 Samtleben C, Schroder A (1992) Living coccolithophore communities in the Norwegian- Greenland Sea and their record in sediments. Mar Micropaleontol 19: 333-354 Smetacek V (2001) A watery arms race. Nature 411: 745 Theodoridis S (1984) Calcareous nannofossil biostratigraphy of the Miocene and revision of the helicoliths and discoasters. Utrecht micropaleont Bull 32: 1-271 Thierstein H R, Geitzenauer K R, Molfino B, Shackleton N J (1977) Global synchroneity of late Quaternary coccolith datum levels: Validation by oxygen isotopes. Geology 5: 400-404 Wallich G C (1877) Observations on the coccosphere. Ann Mag nat Hist 19: 342-350 Wei W (1993) Calibration of Upper Pliocene - Lower Pleistocene nannofossil events with oxygen isotope stratigraphy. Paleoceanography 8: 85-99 Wei W, Shilan Z (1996) Taxonomy and magnetobiochronology of Tribrachiatus and Rhomboaster, two genera of calcareous nannofossils. J Paleontol 70: 7-22 4 Discussion 165

Westbroek P, Brown C W, Bleijswijk J v, Brownlee C, B rummer G-J A, Conte M, Egge J K, Fernandez E, Jordan R W, Knappersbusch M, Stefels J, Veldhuis M J W, Wal P v d, Young J R (1993) A model system approach to biological climate forcing. The example o f Emiliania huxleyi. Global and Planetary Change 8: 27-46 Westbroek P, Hinte J E v, Brummer G-J A, Veldhuis M J W, Brownlee C, Green J C, Harris R, Heimdal B R (1994) Emiliania hwcleyi as a key to biosphere-geosphere interactions. In: Green J C, Leadbeater B S C (eds) The Haptophyte Algae. Clarendon Press, Oxford, pp 321-334 Westbroek P, Young J R, Linschooten K (1989) Coccolith production (Biomineralization) in the marine alga Emiliania huxleyi. J Protozool 36: 368-373 Winter A, Jordan R W, Roth P H (1994) Biogeography of living coccolithophores in ocean waters. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 161-177 Winter A, Siesser W G (1994) Atlas of living coccolithophores. In: Winter A, Siesser W G (eds) Coccolithophores. Cambridge University Press, Cambridge, pp 107-159 Young J R (1991) A Quarternary nannofossil range chart. Int Nannoplankton Assoc Newsl 13: 14-17 Young J R (1994) Variation in Emiliania huxleyi coccolith morphology in samples from the Norwegian EHUX experiment, 1992. Sarsia 79: 417-425 Young J R (1998) Neogene. In: Bown P R (ed) Calcareous Nannofossil Biostratigraphy. Chapman & Hall, London, pp 225-265 Young J R, Bown P R (1994) Palaeontological perspectives. In: Green J C, Leadbeater B S C (eds) The Haptophyte Algae. Clarendon Press, Oxford Young J R, Davis S A, Bown P R, Mann S (1999) Coccolith ultrastructure and biomineralisation. J struct Biol 126: 195-215 Young J R, Didymus J M, Bown P R, Prins B, Mann S (1992) Crystal assembly and phylogenetic evolution in heterococcoliths. Nature 356: 516-518 Young J R, Westbroek P ( 1991 ) Genotypic variation in the coccolithophorid species Emiliania huxleyi. Mar M icropaleontol 18: 5-23 5 Appendix 166

5 Appendix

5.1 Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton

This is a methodological paper, which was published in Micropaleontology, vol. 45 (1999), pp. 473-442. Due to a mistake during the final stages of the production one table was omitted from the original version. Here we have corrected this mistake and have included these data. The first author has constructed the settling device used, has conducted the experiments, analysed the data and written the manuscript. Additional palaeontological data (Table 2) was made available by Jens Herrle. Jorg Bollmann provided the micro beads used as spiking material and sample material. Jeremy Young assisted with data interpretation. 5 Appendix 167

Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton

Markus Geisen Jorg Bollmann \ Jens O. Herrle \ Jorg Mutterlose ‘ and Jeremy R. Young ' The Natural History Museum, Palaeontological Dept., London SW7 5BD, Great Britain ' Eidgenossische Technische Hochschule, Geological Institute, Soimeggstr. 5,8092 Zurich, Switzerland ■ Universitat Tubingen, Institut und Museum fiir Geologie und Palaontologie, Sigwartstr. 10, 72076 Tübingen, Germany 'Ruhr-Universitat Bochum, Institut fur Geologie, UniversitatsstraBe 150, 44801 Bochum, Germany email: [email protected]

ABSTRACT: We describe a device for random settling preparation o f calcareous nannoplankton (coccolith) samples. The device allows easy draining and cleaning, as well as sedimentation at different heights of the water column through the use o f different slide carriers. Reproducibility and accuracy of the device has been tested with standardised microbeads as spiking material. In applying counting techniques with different equa­ tions for determining absolute abundances, we discovered major anomalies in the calculated results that we interpreted as being due to the effect of convection currents within the device and the usage o f elevated cover slides. A modified formula that corrects for the influence o f elevated cover slides in random settling experiments is therefore proposed. Although the settling method is more time-consuming than standard smear-slide techniques, additional information is gained about the spatial and temporal distribution o f coccoliths. These are important for palaeoecological and palaeobiogeographical interpretations.

INTRODUCTION the method is discussed in some detail here and the count data The calculation of absolute abundances of coccoliths is valuable are presented in detail. for comparison of spatial and temporal sample sets. Standard smear slide preparation techniques used for nannofossils do MATERIAL not provide this information since the quantity of sediment on the slide is not known. In recent years, various authors (Moore, Random settling device 1973; Beaufort, 1991; Williams & Bralower, 1995; Su, 1996; The random settling device consists of an acrylic cylinder 80 Flores & Sierro, 1998; Baumann et al., 1999) have applied mm high and 80 mm wide. It is fixed on an acrylic base, which random settling techniques to allow the calculation of absolute is penetrated by a boring allowing the drainage of the water abundances of coccoliths. (Text-fig. 1,2). An o-ring is used to seal the base. A fixed mark in the upper part of the cylinder allows precise control of the The basic principle of the random settling technique is that the amount of water being used. A small hole in the centre of the nannofossils are allowed to settle out from a dilute suspension base is used to fix the slide carriers. We used slide carriers of onto a cover slip. The abundance of the nannofossils can then be different heights to attain water columns, varying from 5 -4 0 calculated from the weight of sediment put into suspension and mm in 5 mm increments. Stainless steel is used for the clip on the volume of the water column above the cover slip. Beaufort top of the slide carriers to avoid corrosion. The clip holds the (1991), Williams & Bralower (1995) and Su (1996) have de­ cover slide in such a way that it is possible to remove the slide scribed applications of this technique using standard laboratory carriers without touching the cover slide. equipment. We encountered problems in applying their methods due to dissolution of coccoliths during sedimentation, and an Spiking material overall lack of precision in the design of the device used. As spiking material borosilicate microbeads with a mean diam­ eter of 5.1 pim (standard deviation 0.8 pcm) and a specific grav­ The initial objective of this study was to develop a convenient ity of 2.5 g/cm' have been used. The microbeads are produced device in order to carry out random settling preparations and by Duke Scientific, Product Number 9005, Lot No. 19324 (J. calculate absolute abundances of coccoliths (coccoliths per unit Bollmann, personal communication). mass sediment). Design features of the device we developed, as described below, include a drain hole to allow removal of the Sediment samples water without evaporation, and cover slip holders designed to In order to test the settling method we used the same sediment ensure convenient use and precise positioning. The advantage (fine fraction <38 fim of DSDP 607,2-2,29-32 cm). This is a mid of the design described here is that it allows easy draining Pleistocene DSDP sample from the upper westem flank of the and cleaning. Additionally, the device allows preparations to Mid-Atlantic ridge, rouhgly 240 nautical miles northwest of the be made at different heights in the water column. To test the Azores (Ruddiman et al., 1987). The assemblage is dominated reproducibility of results using this device repeat preparations (>90%) by G. oceanica. were made of a single sample. In addition the accuracy of the For further tests whole rock samples from the middle Aptian of calculated results was checked in two ways: First by using a South-East France (Serre Chaitieu section, cf. Thierstein, 1973) sample for which absolute abundances had been calculated for have been used (PK74, PK84, PK85, PK86, PK87). another paper (Bollman etal., 1999); and second by spiking the sample with a known weight of borosilicate-glass microbeads. Sample preparation This test produced rather surprising results. Since this is of sig­ 1) A small amount of sediment is chopped off using a scalpel nificance for other applications of the random settling technique, and dried in an oven for 24 hours at 50° C. Alternatively the fine

reproduced and amended from micropaleontology, vol. 45, no. 4, pp. 437-442, text-figures 1-3, tables 1-2, 1999 5 Appendix 168

Markus Geisen el. al.: Calibration of the random settling^ technique for calculation of absolute abundances of calcareous nannoplankton

TEXT-FIGLIRE 1 Random settling device (left) and four slide carriers (right).The latter are used to attain different water column heights. fraction of sieved and filtered samples (<38 pm) can be used. from the device and so were avoided.

2) With bulk rock samples 10-50 mg of sediment is weighed using a microbalance (Mettler AE 260 with a precision of Counting 10 g). The amount of material used depends on the coccolith When examining the slide under the light microscope or the abundance in each particular sample. With bulk rock samples scanning electron microscope the number of studied fields of this can usually be estimated by using the sediment colour as view and the number of coccoliths in each field of view are an indicator. With fine fraction samples 1-10 mg of sediment noted. 600 fields of view per preparation were automatically is sufficient for both light- and scanning electron microscopy captured at a magnification of 3000x using a scanning electron analyses. For control experiments 2-5 mg of borosilicate mi­ microscope (Philips XC30) and written on CD. Subsequently crobeads are added to the sample. coccoliths were counted using a computer and the public domain image analysis program NIFl-lmage (developed at the U.S. Na­ 3) The sample is transferred into a sealed tube. After adding a tional institutes of health and available on the Internet at http: small amount of water, it is ultrasonicated until all particles are in suspension. We used normal tap water buffered with NFlj (pFl=8.5) to prevent etching and with a small amount of Triton To gain comparable results only complete coccoliths and X I00 detergent added to remove surface tension. fragments of more than a half coccolith were counted, also Florispliaera profunda nannoliths were excluded in the SEM 4) The suspended sample is transferred to a volumetric flask and counts. The total number of coccoliths was calculated using the diluted to 1000 cm-. Afterwards the suspension is homogenised following equation (Williams & Bralower, 1995): by a magnetic stirrer for several minutes followed by 4 inver­ sions of the flask. X=(NxV)/(MxExAxH)(l)

5) The suspension is poured into the settling device and left to where settle for 24 hours. This settling time was based on the height X=particles per gram of sediment |n/g| of the water column. According to Stokes' Law (cf. Walsby & N=number of particles counted Reynolds, 1980; Young, 1994) a 2 pm diameter calcite sphere V=volume of water used for dilution [ml] will sink at approximately 0.2 mm/minute, i.e. 300 mm/day. M=grams of sediment added |g| So 24 hours should ensure complete sedimentation of particles F=number of fields of view observed through the 50 mm water column. A=surface area of one field of view [cm^l FI=height of water column above slide |cm] 6) The water is drained carefully using the drain valve. This procedure eliminates the need for evaporating the water and so This equation is based on the assumption that the number of reduces the preparation time to an acceptable minimum. Once particles, whether coccoliths or microbeads, collected per unit the remaining water on the cover slide has air dried the slides area of the cover slip is proportional to the volume of suspension are mounted in the usual way. originally present above the cover slip.

7) To avoid contamination of samples the device is cleaned thor­ If the samples have been spiked by addition of a known number oughly after each usage. Acids proved hard to remove entirely of micro beads then the total number of coccoliths can be cal- reproduced and amended from micropaleontolo^y, vol. 45, no. 4. pp. 437-442, te.xt-fi^ures 1-3, tables 1-2, 1999 5 A ppendix 169

micropaleontology, vol. 45, no. 4,1999

acrylic cylinder mark (water surface)

% cover slide holder (stainless steel) slide carrier (PTFE) % locating lug at base % of slide holder acrylic baseplate o-ring seal

drain (stainless steel)

TEXT-nGURE2 Sketch of the random settling device used in this study. culated independently, by assuming that the relative abundance ing a Zeiss photomicroseope at a magnification of 1600X to counts of the spheres and coccoliths are proportional to their compare the scanning electron microscope counts with light absolute abundances. It is also necessary to calculate the number microscope counts. of microbeads per gram: The average number of coccoliths calculated from the number X = 6 X lO '- jJt X p xd^)j,| of observed particles per area (equation I) was 1.17*10" coc­ coliths per gram with a coefficient of variation of ±5.5% (Tab. 1) for the scanning electron microscopy counts and 1.12*10" X=number microbeads /gram coccoliths per gram with a coefficient of variation of ±4.22% p=density of sphere (g/cmO (Tab.l) for the light microscopy counts. The low variation in d=mean diameter (/

RESULTS To investigate the anomaly, we first compared the known weight The settling technique was tested by making 13 repeat prepara­ of the microbeads added to the samples, with the estimated tions of the DSDP sample spiked with borosilicate microbeads. weight from the microbeads counted. The estimated weight DependinglO‘° on the density of the preparations 100 or 150 was approximately 2.5 times higher than the known weight fields of view were counted, one of the preparations (76 & 76a) (Tab.l, equation 2). We then compared our estimates of cocco­ was counted twice to show variation on one microscope stub lith abundance in the sample with those calculated for the same (Text-fig. 3). Additionally five samples have been counted us­ sample in Bollmann et al. (1999). They estimate: 5.52* 10*°total reproduced and amended from micropaleontology, vol. 45, no. 4, pp. 437-442, text-figures 1-3, tables 1-2, 1999 5 A ppendix 170

Markus Geisen et. al.: Calibration of the random settling technique for calculation of absolute abundances of calcareous ruinnoplankton

column. Therefore we suspect that this enrichment is due to convection currents in the settling device. Close examination of settling suspensions did indeed reveal quite perceptible turbulent flow of particles in the suspension, continuing long after initial 75 81 addition of the suspension. These flows were estimated by eye to have velocities in the order of mm per second. By contrast Stokes' Law sinking rates for coccoliths and microbeads should be orders of magnitude lower: 4 pm per second, 0.2 mm per minute for a 2 pm calcite sphere or 20 /

CONCLUSIONS The results obtained with the commonly used particles per area 70 7 6a equation (equation 1 ) to a high degree of precision are repro­ ducible but yielded anomalous high values for the number of coccoliths per gram sediment. Using the microbeads as a tracer an enrichment of approximately 2.5 times for the settling device used could be demonstrated. We suspect that this apparent en­ 69 76 richment is due to convection currents in our new settling device. We suspect that convection currents occur in other settling set­ ups although they might be influenced by room temperature, grain size and the geometry of the settling device.

1 5 0 100 150 1 50 100 150 These observations lead to the following conclusions: (1) Our new settling device produces reproducible results, al­ TEXT-HGURE 3 though (2) the commonly used particles per area equation yields Results for the SEM counts on repeated preparations of DSDP incorrect results. (3) The settling technique and the particles per sample 607,2-2,29-32 cm. Each graph represents the evolution area equation remains a valid method if: the cover slide is placed of the calculated number of coccoliths per gram. The broken on or near the bottom of the settling. Furthermore a microbead lines represents the mean (4.7* 10'°) and standard deviation spike test can be carried out to check the results. (2.66*10°) calculated from all 13 preparations using equation 1 (H=5 cm). Note that samples 76 and 76a are double counts ACKNOWLEDGEMENTS on different areas on the microscope slide. Numbers given in The authors appreciate the help of the mechanical workshop of each graph are counted specimens. Counts stabilise after ap­ the Ruhr-Universitat Bochum, who built the random settling proximately 50 counted fields of view (-500 specimens); in­ device. Prof. Hans Thierstein (Eidgenossische Technische Hoch­ stability prior to this is due to the (low) unevenness in number schule, Zurich) encouraged this work and kindly provided the of specimens per field of view. Variation in final values reflects DSDP sample, the microbeads and excellent SEM facilities as the imprecision of the preparation method. well as laboratory facilities. Dr. Steve Culver and Dr. Neale Monks carefully reviewed earlier versions of this manuscript. coccoliths per gram (including F. profunda)-, 5.13*10'° (only This work is a contribution of the EC-TMR research project Gephyrocapsa) and 5.3*10'° coccoliths excluding F. profunda CODENET. based on a spraying technique with a microbead spike; 4.27* 10'° total coccoliths per gram based on a filtration technique, using a particles per area equation; and 4.9* 10'° total coccoliths per gram based on a filtration technique with a microbead spike. Evidently our estimate of 4.59* 10'° coccoliths per gram based on the microbead spike is much more comparable to those from other techniques. It follows that both the micro beads and the coccoliths are approximately 2.5 times more abundant on the cover slips than we would expect based on the independent es­ timates and the assumption of simple settling through the water

reproduced and amended from micropaleontology, vol. 45, no. 4, pp. 437-442, text-figures 1-3, tables 1-2, 1999 5 Appendix 171

micropaleonlolof’y, vol. 45, no. 4, IW 9

TABLE I Dataset showing the results for 13 repeated preparations of one sample (DSDP607,2-2,29-32 cm) for scanning electron microscope (top) counts and light microscope counts (bottom). Field of view is 1.1631* 10 ’ cm’ for scanning electron microscopy observations and 1.13*10'' cm' for light microscopy observations. Note the great variation in the amount of sediment and microbeads put in the different preparations in contrast to the uniform results. Note that samples 76 and 76a are double counts on different areas on the mieroseope slide. H: height.

coccoliths per calculated weight ratio spheres coccoliths per coccoliths per sediment spheres fields of gram fine of spheres in weight / spheres gram fine coccoliths gram fine L weigtit in weight in fraction (using sample (using calculated (using fraction (using counted counted fraction mg mg counted equation 1 with equation 1 with equation Iwith equation 1 with (equation 3) H=2 cm) H= 2cm) in mg 1-1=2 cm) H=5 cm) I i . i

69 2 635 3,551 150 1128 193 1.26E+11 9,605 2.70 4 54E+10 5.08E+10 70 3.103 3957 150 1247 187 1.17E+11 9 306 2 35 490E+10 451E+10 71 3 535 3257 100 949 132 1.18Et11 9853 3.03 3 81E+10 4 85E+10 E 72 3 238 2716 150 1262 161 1.14E+11 8 012 2,95 3 79E+10 4 70E+10 73 1 3 334 3 105 100 932 129 1.24E411 9629 3 10 3 88E+10 501E+10 s 74 3 722 4 982 150 1475 254 1.16E+11 12,640 2 5 4 4 48E+10 4 80E+10 75 1 3 424 3927 150 1339 132 114Ev11 6 569 1,67 670E+10 4 48E+10 76 1 4 111 4 863 100 1012 155 1.08E+11 11,570 2.38 4 45E+10 4 37E+10 o 76a 4 111 4 863 150 1455 256 1 03E+11 12,740 2 6 2 387E+10 429E+10 è 77 2 715 3 723 150 1060 166 1.15E+11 8261 2 22 504E+10 4 26E+10 O 78 4 515 4 598 150 1809 219 1.18E+11 10,898 2,37 4 84E+10 4 82E+10 79 6 758 3909 100 1741 130 1.14E+11 9,704 2,48 4 46E+10 4 82E+10 80 5 3 154 100 1322 102 1.17E+11 7,614 2.41 4 71E+10 4 83E+10 81 3 946 3 125 150 1727 162 1.27E+11 8 062 2.58 4 86E+10 5 04E+10 1.17E+11 2 53 4 59E+10 4 70E+10 standard dev 622E+09 0 35 7 46E+09 2 66E+09 coefficient of variation ± 5 34% ± 16 2% ± 16 2% ± 5 66%

coccoliths per calculated weight ratio spheres coccoliths per coccoliths per 1 sediment fields of gram fine of spheres in weight / spheres gram fine coccoliths beads gram fine weight in weight in view fraction (using sample (using calculated (using fraction (using counted counted fraction mg counted equation 1 with equation 1 with equation Iwith equation 1 with (equation 3) H=2 cm) H= 2cm) in mg H=2 cm) H=5 cm) f 1 69 2635 3551 9 640 59 1,19E+11 5 032 1 42 8 42E+10 4 77E+10 70 3 103 3 957 9 719 81 1 14E+11 69 0 9 1.75 6 52E+10 4 55E+10 71 3 535 3257 9 804 62 1 12E+11 5288 1.62 688E+10 4 47E*10 72 3 238 2716 9 692 68 1 05E+11 5,715 2 10 4 92E+10 4 20E+10 78 4 515 4 598 9 1008 118 1 10E+11 10 065 2 19 5.01E+10 4 39E+10 1 12E+11 1 82 635E+10 4 48E+10 standard dev, 4 72E+09 0.29 1 45E+10 1 89E+09 coefficient of variation ± 4 22% ± 16% ± 22.9% ± 4,22%

TABLE 2 Dataset showing the assemblage composition (in percent) and absolute abundances for five mid Aptian samples as determined with a light microscope. Field of view is 1.77* 10^ cm’. For each sample two preparations, one at 2 cm watercolumn and one at 5 cm watercolumn have been examined. Note the low variation in the two preparations of each sample in both the assemblage composition and the absolute abundances. The absolute abundances were calculated with equation 1 corrected for 5 cm water column.

I; ; i f : 84 03 8.7 03 1 6

85 0.3 2.8 11,$ 0 6 0.6 0 3 .0 .3 0 3 0 3 0 3 0.6 1.3 03 80 1 0 68 16.1 06 0 4 06 09 03 09 60 06 60 13.5

SI i j 1.0 38 9 '«5 1-0 0.8 367 _m_Li __ 13 12.9 0.6 ISA 0.3

166 0 3 0 3 14 0 0.3 2 6 17.2 1.0 33 1 06 13,0 1.3 2.0 34 2 __

reproduced and amended from micropaleontolo^y, vol. 45, no. 4, pp. 437-442, text-figures 1-3, tables 1-2, 1999 5 Appendix 172

Markus Geisen et. al.: Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton

REFERENCES SU, X., 1996. Development of late Tertiary and Quaternary coc­ BAUMANN, ANDRULEIT, A. H. and XIN SU, 1999. colith assemblages in the northeast Atlantic. Geomar Comparison of different preparation techniques for Report, 48: 119 p. quantitative nannofossil studies. Journal of Nanno­ THIERSTEIN, H R. 1973. Lower Cretaceous calcareous nan­ plankton Research, 20 (2): 75-80. noplankton biostratigraphy. Abhandlungen der Geolo- BOLLMANN, J., BRABEC, B., CORTÉS, M. Y. and GEISEN, gischen Bundundesanstalt, 29: 52 p. M., 1999. Determination of absolute coccolith abun­ WALSBY, A.E., REYNOLDS, C.S. 1980. Chapter 10. Sinking dances by spiking with microbeads and spraying and floating. In: Morris, I. (Editor), The Physiologi­ (SMS-method). Marine Micropalaeontology, 38: cal Ecology of Phytoplankton. Studies in Ecology, 29-38. 7: 371-412. BEAUFORT, L., 1991. Adaptation of the Random settling WILLIAMS, J R., BRALOWER, T.J., 1995. Nannofossil as­ method for quantitative studies of calcareous nan­ semblages, fine fraction stable isotopes, and the pale­ nofossils. Micropaleontology, 37: 415-418. oceanography of the Valanginian-Barremian (Early FLORES, J.A., SIERRO, F.J., 1998. Revised technique for Cretaceous) North Sea Basin. Paleoceanography, 10 calculation of calcareous nannofossil accumulation (4): 815-839. rates. Micropaleontology, 43 (3): 321-324. YOUNG, JR . ( 1994). Functions of coccoliths. In: Winter, A. and MOORE, T.C., 1973. Method of randomly distributing grains Siesser, WG. (Editors), Coccolithophores, pp. 63-82. for microscopic examination. Journal of Sedimentary Cambridge: Cambridge University Press. Petrology, 43 (3): 904-906. RUDDIMAN, W.F., KIDD, R.B., THOMAS, E., et al., 1987. Manuscript received December 30, 1998 Initial Reports of the Deeps Sea Drilling Project. Manuscript accepted April 15, 1999 Volume 94, part 1: 1-613. Washington D C.: U.S. Government Printing Office.

reproduced and amended from micropaleontology, vol. 45, no. 4, pp. 437-442, text-figures 1-3, tables 1-2, 1999 5 Appendix 173

5.2 Determination of absolute coccolith abundances in deep-sea sediments by spiking with microbeads and spraying (SMS-method)

This is a methodological paper, which was published in Marine Micropaleontology, vol. 38 (1999), pp. 29-38. The last author has tested the method presented by preparing a large number of samples both for this manuscript and for general use as the communal sample set in CODENET. Additionally comparative data obtained by the author of this thesis with the settling method was included in this manuscript. 5 A ppendix 174

mRRinE mitROpniEonroLocv

ELSEVIER Marine Micropaleontology 38 (1999) 29-38 www.elsevier.com/locate/marmicro

Determination of absolute coccolith abundances in deep-sea sediments by spiking with microbeads and spraying (SMS-method)

Jorg Bollmann^*, Bernhard Brabec Mara Y. Cortés®, Markus Geisen

" Geological Institute, ETZ-Zurich, Sonneggstrasse 5, 8092 Zurich, Switzerland * Swiss Federal Institute for Snow and Avalanche Research, Fliielastrasse 11, 7260 Davos Dorf. Switzerland '' The Natural History Museum, Palaeontological Department, Cromwell Road, London SW7 5BD, Great Britain

Received 22 December 1998; accepted 15 June 1999

Abstract

A quick new method is described for the quantification of absolute nannofossil proportions in deep-sea sediments. This method (SMS) is the combination of Spiking a sample with A/icrobeads and Spraying it on a cover slide. It is suitable for scanning electron microscope (SEM) analyses and for light microscope (LM) analyses. Repeated preparation and counting o f the same sample (30 times) revealed a standard deviation o f ±10.5% . The application o f tracer microbeads with different diameters and densities revealed no statistically significant differences between counts. The SMS-method yielded coccolith numbers that are statistically not significantly different from values obtained from the filtration-method. However, coccolith counts obtained by the random settling method are three times higher than the values obtained by the SMS- and the filtration-method. © 1999 Elsevier Science B.V. All rights reserved.

Keywords: coccoliths; absolute abundance; microbeads; preparation technique

1. Introduction Sierro, 1998) and (b) a filtration technique described by Backman and Shackleton (1983) and modified by There is an increasing need for absolute calcare­ Andruleit (1996). The accuracy and reproducibility ous nannofossil counts in order to calculate coccolith of both techniques depends on three factors: ( 1) the fluxes from the photic zone into sediment traps or to even distribution of the particles on a filter or on a estimate the varying coccolith carbonate accumula­ cover slide, ( 2) the assumption that no size-depen­ tion rates in geological records. In the past, several dent fractionation of coccoliths occurred during the different techniques were developed, although there preparation procedure, and (3) that there is no loss are currently only two basic methods that are in or bias due to splitting procedures, filter funnels or use: (a) the random settling technique introduced by settling devices. Beaufort (1991) and modified by different authors Here, we demonstrate the application of the com­ (Williams and Bralower, 1995; Su, 1996; Flores and bination of two techniques that have been already applied or suggested for nannofossil analyses: (a) a * Corresponding author. Tel.: -f4I 1632 3684; Fax: -1-41 1632 spraying technique used by McIntyre et al. (1967) 1080; E-mail: [email protected] and (b) the addition of microbeads as tracer parti­

0377-8398/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PIT: 80377-8398(99)00032-8 5 A ppendix 175

30 J. Bollm ann et al. / Marine Micropaleontology 38 (1999) 29-38

cles (suggested by Okada, 1992). Furthermore, we <2%. These microbeads are made of polystyrene, compare the results obtained by this new method latex or borosilicate and they are available in many with the counts obtained by the filtration- and the different sizes, accuracy of diameter and physical settling-method. and chemical properties. In order to use these mi­ crobeads for calcareous nannofossil counting they 1.1. Background have to be stable in water suspensions with a pH of 8.5 or in alcohol. Furthermore, they have to be resis­ McIntyre et al. (1967) used a spray gun in order to tant to ultrasonic treatment and they should not form distribute coccoliths of various sizes homogeneously large aggregates during the preparation procedure. on a slide and Okada (1992) suggested adding a We used two different kinds of microbeads that known weight of microbeads (Potter Ballotini Inc., meet the above-mentioned requirements in our new type MB 10) to a sediment sample in order to cal­ method: (a) polystyrene microbeads (mean diameter culate absolute abundances of nannofossils. Okada 4.0 p,m ± 0.06 |xm standard deviation; density = (1992) assumed a constant number of microbeads 1.09 g/cm^) and (b) borosilicate microbeads (mean per unit weight without knowing the real number of diameter 5.1 |xm ± 0.8 p.m, density = 2.5 g/cm^; microbeads per weight unit. Therefore, his method for details see Table 1). enables only the counting versus a relative standard The number of microbeads per solid weight (e.g. but not the calculation of absolute nannofossil con­ 1 g) can be calculated as follows: tent per gram sediment. However, Okada (1992) also suggested calibrating the number of microbeads per M-xot = -7T 7 — -—— =------z------r ( 0 weight unit using a suitable standard if this standard P m X were available. where A / j o t = number of microbeads per gram; Adding tracer particles or tracer chemicals is a = weight of one microbead; Pm = density of one common method, especially in chemistry, to calcu­ microbead; Fm = volume of one microbead. late the unknown amount of a component. Particle tracers have been used by micropaleontologists since 2.2. Sediment samples the 1960’s (Benninghoff, 1962; Stockmarr, 1971). Benninghoff (1962) added a known number of Ly- In order to test the new method we used two dif­ copodium spores to a sediment sample in order to ferent sediment samples: (a) the fine fraction (<38 calculate the absolute abundance of pollen per gram p,m) of a pure coccolith ooze from the mid-Pleis- sediment. However, until now it was not possible tocene Gephyrocapsa dominance interval (Bollmann to apply this simple concept to coccolith counts be­ et al., 1998) of North Atlantic DSDP Hole 607 that is cause (a) no uniform tracer particles were available composed of 93% of Gephyrocapsa placoliths (607, in the size range of coccoliths and (b) it was difficult 2-2, 29-32 cm), and (b) a hemipelagic Holocene to calculate the number of tracer particles that had sample (bulk) with low carbonate content (Blagnac to be added to a sample with an acceptable standard sample: off Cape Verde; W 18°15.1; N 21°19.76; deviation. 2002 m water depth). This sample was already used for an international intercalibration experiment be­ tween thirteen nannofossil experts. Until now, the 2. Materials and methods results of this experiment have not been published and therefore, we show only our own determinations 2.7. Microbeads (Table 2).

Microbeads of uniform diameter have recently be­ 2.3. Spraying device come available commercially and thus allow a sim­ ple theoretical calculation of microbeads per solid We used an Effa Spray Mounter and small glass weight (instead of enumerating complete aliquots, capillaries (for details see Table 1). The distance see Stockmarr, 1971) with a standard deviation of between the target and the spray mounter was found 5 A ppendix 176

J. Bollmann et al./M arine Micropaleontology 38 (1999) 29-38 31

Table 1 Materials used for SMS preparation

Item Ordering No. Amount Price Supplier

Polystyrene Microbeads; Product number: I g $ 270,- Bangs Laboratories, Inc., 9025 Technology Drive, diameter: 4.0 ± 0.06 [im; density: PS055N/001168; Inv. Fishers, IN 46038-2886 USA, e-mail: 1.0® g/cm^ Nr. L970305E [email protected]

Borosilicate Microbeads; Product number: 1 g ~$ 150,- Duke Scientific Corp., 2463 Faber Pl., Palo Alto, diameter: 5.1 ± 0.8 |im; density: 9005; LOT Nr. 19324 P.O. Box 50005, CA 94303, USA, e-mail: 2.5 g/cm^ [email protected] Spray gun 11805 or 11800 1 $95,- Ernest Fullam., 900 Albany Shaker Rd. Latham, New York 12110, USA, e-mail: [email protected] Small glass capillary tubes 11802 100 $ 13,- Ernest Fullam., 900 Albany Shaker Rd. Latham, New York 12110, USA, e-mail [email protected] Triton X 100 detergent XlOO 100 ml S 28,- Sigma Chemical Company, P.O. Box 14508, St. Louis, Missouri 63178-9916, USA Laser granulometer, GALAl CIS 1 L.O.T. - Oriel GmbH, D-64295 Darmstadt, Germany

to be optimal at ca. 20 cm with respect to particle these microbeads. Borosilicate microbeads and sedi­ density and distribution. ment are suspended in 2 to 4 ml water. We buffered the water with NH 3 to a pH of 8.5 in order to avoid 2.4. Sample preparation dissolution or overgrowth of coccoliths. In addition, a very small amount of Triton X 100 detergent was 2.4.1. Spraying (Figs. 1 and 2) added to avoid coagulation of the particles. Step 1. Microbeads and dry sediment are weighed Step 3. The suspension is ultrasonicated for about with a high-precision balance (e.g. 1.0 mg with a 30 s at 35 kHz. Mettler balance AE 260 with a precision of 10~® g). Step 4. The suspension is sprayed onto a cover slip We suggest adjusting the ratio between the weight (5 to 10 times) using small glass capillaries (Figs. 1 of microbeads and the weight of sediment in order and 2). However, for routine work we recommend to avoid counting excessively large numbers of mi­ using a 5 ml syringe in combination with a glass crobeads with low numbers of coccoliths, or vice capillary (Fig. 2). The syringe is filled with the versa. For pure coccolith oozes a weight ratio of suspension and the suspension is sprayed onto the microbeads (polystyrene) to coccoliths (<38 |xm target. This offers the advantages of (a) spraying a fraction) of 2 to 1 is recommended and for samples complete suspension onto a target without refilling with low carbonate content (e.g. 40%) we recom­ capillaries and (b) a better control on how much mend to use a ratio of 1 to 4. If bulk samples are material is sprayed onto the target. used, we suggest increasing the amount of sediment Step 5. The dry cover slide is mounted on an to a minimum of 10 mg in order to avoid inhomo- aluminium stub for scanning electron microscope geneous samples. However, if fine-fraction samples (SEM) analysis or on a microscope slide for light are used about 1 mg sediment is sufficient in order to microscope (LM) analysis. guarantee homogeneous samples. Detailed ordering information (Table 1) and a Step 2. Polystyrene microbeads and sediment are full description of the spraying setup is given on: suspended in 2 to 4 ml of denatured alcohol. We http://www.geology.ethz.ch/mp/data/microbeads/ used alcohol in order to provide a homogeneous beads main.html mixing between polystyrene microbeads and sedi­ ment because of the comparatively low density of T a b l e 2 T est for rep rod u cib ility and p recision o f the S M S -m eth od

Sample Method Microbead type Microbeads S e d i m e n t M icrobead counts C occolith counts C o c c o l i t h s E. huxleyi Gephyrocapsa E pm funda O t h e r s ( g X 1 0 - “ ) ( g X 1 0 - “ ) ( x l O ' V g ) ( x l O ' O / g ) spp. (x IO '°/g) ( x l O ' V g ) ( X l O ' V g ) (%)

D S D P 6 0 7 - I SMS P 7.68 26.15 203 1228 4.87 4 . 5 3 0 . 2 0 0 . 1 3 8 . 0 1 D S D P 6 0 7 - 2 SMS P 5.44 15.03 180 937 5.16 - 4 . 8 9 0 . 1 4 0 . 1 3 8 . 5 4 DSDP607-3 SMS P 5 . 4 4 1 5 . 0 3 2 0 1 1 0 5 4 5 . 2 0 - 4.79 0.19 0.22 8.13 D S D P 6 0 7 - 4 SMS P 3 . 7 8 1 1 . 1 9 2 1 2 1 0 4 5 4 . 5 6 - 4 . 2 0 0.21 0.14 7.97 D S D P 6 0 7 - 5 SMS P 8 . 1 4 2 5 . 7 2 2 0 1 1 1 2 3 4 . 8 4 - 4.56 0.13 0.16 8.09 DSDP607-6 SMS P 6 . 0 4 13.24 201 812 5 . 0 5 - 4 . 8 4 0 . 0 9 0 . 1 2 8 . 3 0 DSDP607-7 SMS P 9.5 34.22 153 1025 5 . 0 9 - 4 . 7 3 0 . 1 6 0 . 2 0 9 . 0 5 DSDP607-8 SMS P 4 . 6 8 1 4 . 9 0 2 0 3 1 1 8 0 5 . 0 0 - 4 . 6 9 0 . 2 1 0 . 1 0 8 . 0 3 DSDP607-9 SMS P 8.1 16.35 2 0 5 8 1 0 5 . 3 6 - 5 . 0 7 0 . 1 7 0 . 1 3 8 . 2 4 DSDP607-I0 SMS P 5 . 9 4 5 . 9 3 3 0 7 6 8 1 6 . 0 9 - 5 . 6 4 0.25 0.20 7.35 DSDP607-II SMS P 5.49 4.46 3 0 9 5 2 5 5 . 7 3 - 5 . 2 8 0 . 2 9 0 . 1 5 7 . 6 3 D S D P 6 0 7 - I 2 SMS P 1 0 . 7 8 1 8 . 1 9 2 0 9 5 9 0 4 . 9 9 - 4 . 7 1 0 . 1 4 0 . 1 4 8 . 4 6 DSDP607-I3 SMS P 1 5 . 9 4 1 7 . 4 4 3 0 3 7 4 4 6 . 1 5 - 5.72 0.24 0.19 7.29 D S D P 6 0 7 - 1 4 SMS P 1 1 . 5 5 1 4 . 5 2 3 0 1 7 7 7 5 . 6 2 - 5.24 0.24 0.14 7.27 DSDP607-I5 SMS P 12.62 14.76 203 5 3 3 6 . 1 5 - 5.65 0.33 0.16 8 . 6 5 DSDP607-16 SMS P 1 3 . 1 4 2 0 . 8 8 2 0 4 6 2 2 5 . 2 6 - 4 . 8 3 0.25 0.17 8.48 DSDP607-17 SMS P 2 1 . 0 6 3 1 . 0 6 2 7 4 8 4 2 5 . 7 1 - 5.29 0.24 0.17 7.42 DSDP607-18 SMS P 9 . 6 1 1 2 . 2 1 2 6 4 7 6 1 6 . 2 1 - 5 . 7 1 0 . 2 9 0 . 2 1 7 . 6 0 D S D P 6 0 7 - I 9 SMS P 8 . 4 9 7 . 9 4 2 0 4 4 3 2 6 . 2 0 - 5 . 6 9 0 . 3 0 0 . 2 2 8 . 8 8 DSDP607-20 SMS P 8 . 5 3 6 . 2 9 3 0 6 5 0 1 6 . 0 8 - 5 . 4 9 0 . 3 9 0 . 2 1 7 . 7 1 D S D P 6 0 7 - 2 I SMS P 9 . 5 2 1 9 . 0 8 2 0 6 9 1 8 6 . 0 9 - 5.65 0.33 0.11 8 . 1 4 D S D P 6 0 7 - 2 2 SMS P 6 . 9 1 2 . 4 8 2 0 4 9 0 4 6 . 7 1 - 6 . 3 2 0 . 2 0 0 . 1 9 8 . 1 8 DSDP607-23 SMS P 7 . 9 8 12.74 212 741 6 . 0 0 - 5 . 5 8 0 . 2 3 0 . 1 9 8 . 2 1 D S D P 6 0 7 - 2 4 SMS P 18.5 13.68 3 1 5 5 0 0 5 . 8 8 - 5 . 4 1 0 . 1 5 0 . 3 2 7 . 6 5 DSDP607-25 SMS P 6.19 11.77 2 0 1 7 1 4 5 . 1 2 - 4.74 0.17 0.20 8.40 DSDP607-26 SMS P 1 3 . 9 1 6 . 6 9 3 0 9 8 3 0 6 . 1 3 - 5.79 0.13 0.21 7.15 DSDP607-27 SMS P 10.4 13.16 301 619 4.45 - 4 . 2 0 0 . 1 4 0 . 1 2 7 . 4 9 D S D P 6 0 7 - 2 8 SMS P 8.65 15.86 3 0 2 9 8 1 4 . 8 5 - 4 . 5 4 0 . 1 6 0 . 1 5 7 . 0 8 DSDP607-29 SMS P 10.17 10.10 3 0 7 6 0 7 5 . 4 5 - 5 . 0 4 0 . 2 5 0 . 1 6 7 . 4 7 DSDP607-30 SMS P 7 . 1 8 9 . 4 1 2 0 9 5 6 7 5 . 6 7 - 4 . 9 8 0 . 2 7 0 . 4 2 8 . 5 0 A v e r a g e 5 . 5 2 7 . 9 8 Standard deviation 0.58 % Standard deviation 1 0 . 5 3

DSDP607-31 SMS B 7 7 . 4 6 3 3 . 6 5 2 0 0 8 4 4 5 . 6 0 - 4 . 9 7 0.31 0.31 0.28 DSDP607-32 F-t-MB B 5 9 . 6 8 2 5 . 6 8 2 8 3 1 0 9 3 5 . 1 7 - 4 . 9 0 0.12 0.15 0.28 DSDP607-32 F B 0 25.68 1093 4.50 - 4.27 0.10 0.13 D S D P 6 0 7 - 3 3 SE 4 -MBB 3 5 . 5 1 2 6 . 3 5 220 1133 4.00 - 3.87 0.05 0.08 0.28 DSDP607-33 SE B 0 2 6 . 3 5 1 1 3 3 1 2 . 3 4 - 1 1 . 9 5 0 . 1 5 0 . 2 4 B l a g n a c - 1 F B 0 1 0 2 . 0 0 1276 0.27 0.083 0.08 0 . 0 4 0 . 0 6 B l a g n a c - 2 SMS B 7 2 . 0 3 2 9 3 . 6 3 2 1 0 4 3 8 0 . 2 9 0 . 0 9 6 0 . 1 1 0 . 0 5 0 . 0 4 0 . 2 8

T h e resu lts o f the rep eated preparations an d cou n tin g o f D S D P sam p le 6 0 7 -2 -2 , 2 9 -3 2 cm and the B lagn ac sam p le are sh ow n . In ad d ition , the cou n ts o f th e sa m e sam p les ob tain ed b y th e filtration and settlin g m eth od are given . T h e abbreviations in colu m n ‘M eth od ’ are: S M S = estim ates ob tain ed b y sp ik in g w ith m icrob ead s and spraying; F = estim ates ob tain ed b y filtration b ased on particles p er area; F + M B = estim ates ob tain ed by m icrob ead s ad d ed to filtered sam p le (ratio m icrob ead s/coccolith s); S E = estim ates ob tain ed b y settlin g b ased on particles per area (after W illiam s an d B ralow er, 1995). S E + M B = estim ates ob tain ed b y m icrob ead s ad d ed to a settled sam p le (ratio m icrob ead s/coccolith s). T h e typ e o f m icrob ead is ind icated b y P = p olystyren e m icrob ead s, an d B = b orosilicate m icrob ead s. N ote that th e n u m b ers o f coccolith s p er gram sed im en t h ave to b e m u ltip lied b y I0 '°. T h e error valu e is calcu lated from the n u m b er o f m icrob ead s and coccolith s coim ted and th e varian ce in m icrob ead sizes. S ee text for d eta ils. 5 Appendix 178

J. Bollmann et al. /Marine Micropalcontology 3B (1999) 29-3H 33

air gun (cross-section) cover slide capillary tube with 'suspended preparation

20 cm working distance

Fig. 1. Sketch oF the spraying set-up. A distance o f 20 cm between gun and target was found to be optimal with respect to particle distribution. Cover slides and SEM stubs can be mounted with a small strip o f double-sided tape or a small strip o f spray glue.

Spray gun (cross-section)

c

Fig. 2. Close-up o f the spraying device: A = spray gun nozzle; B = capillary tube filled with sediment suspension; C = syringe with suspension; D = pipette tip; E = air pressure supply; F = underpressure. Note, for routine work we recommend using a syringe (C) in combination with glass capillaries (B). This offers the advantage (a) to spray a complete suspension onto a target without refilling capillaries and (b) o f a better control on how much material is sprayed onto the target.

2.4.2. Filtration two by pouring the suspension repeatedly from one Both samples (DSDP607 and the Blagnac sample) glass beaker to another until the same amount of were prepared once using the filtration technique af­ suspension was in both beakers (25 ml ± 1 ml). ter Andruleit ( 1996). In contrast to Andruleit ( 1996), One of the beakers was filled to 50 ml with buffered we split the sample manually (see step 3) and not water and split again as described. This procedure with a rotary splitter. In order to obtain an indepen­ was repeated four to five times in order to obtain dent measure of the accuracy of this method, we an appropriated amount of material to be filtered for added a known amount of borosilicate microbeads to optimal particle density on the filter membrane. the DSDP sample. Step 4; the last split was filtered on a Nucleopore Step 1 and 2 same as SMS-method. PC 0.8 p.m m embrane using a Gelm an 1119 inline Step 3: the suspension was manually split in filter gasket. 5 Appendix 179

34 J. Bollmann et al. /Marine Micropaleontology 38 (1999) 29-38

2.4.3. Settling (after Williams and Bralower, 1995) follows (from Andruleit, 1996): We used the same settling box as Geisen et a). F X C X S Ctot = (3) (1999) and we placed the slides at water column A X W height of 2 cm. In order to obtain an independent where Ctot = number of coccoliths in a sample; F measure for the accuracy of this method we added = filtration area; C = number of coccoliths counted; also a known amount of borosilicate microbeads to A = analysed filtration area; W = sample weight; S the DSDP sample. = split factor. Step 1, 2 and 3 as SMS-method. Step 4: the suspended sample was transferred 2.5.3. Settling into a volumetric flask and diluted to 1000 cm^ The analyses of the random settling sample were and homogenised by a magnetic stirrer for several done according to Williams and Bralower (1995). minutes followed by four inversions of the flask. Calculations are based on the number of particles per Step 5: the suspension was poured into the settling observed area and the number of coccolith per solid device and left to settle for 24 h. weight can be calculated as follows (from Williams Step 6: the water was drained carefully using and Bralower, 1995): a drain valve at the bottom of the settling device (for details of the settling device see Geisen et al., ~ W x F xAxH 1999). Once the remaining water on the cover slide where Ctot = number of coccoliths in a sample; C had air-dried the slide was mounted on a SEM stub = number of coccoliths counted; V = volume of (same as SMS-method). water used for dilution; W — sample weight; F = number of fields of view observed; A = area of one 2.5. Counting field of view; H = height of water column above slide. All counts were done using scanning electron microscopes (HITACHI S2300 and PHILIPS XL30) 2.6. Error estimates applying the SMS-method at magnifications of 3000 x and 6000x. The variance of the estimates of coccoliths/g sediment can be derived from information on the 2.5.1. Spraying variance (a) of weighing microbeads, (b) of weigh­ All coccoliths and microbeads were counted in ing sediment, (c) of calculating number microbeads each field of view. The counting was terminated per unit weight, (d) the number of coccoliths counted when at least 100 to 300 microbeads were counted and (e) the number of microbeads counted. In order (Table 2). From the ratio between counted and added to calculate the variance of coccoliths per weight microbeads the number of coccoliths per weight can sediment, ct^(Ctot), we used the Gaussian law of be calculated as follows: error propagation (for details see Appendix A). The

C count M t o T estimates of the standard deviation obtained for 100 C t o t = — ------x (2) Mr, Wt microbeads and 100 coccoliths using the polystyrene microbeads are ±14.4%. Additional counting up to where Ctot = number of coccoliths in a sam­ 1000 microbeads and 1000 coccoliths reduces the ple; Ccount = number of coccoliths counted; Mjot standard deviation to ±5.5% (Fig. 3). However, ad­ = number of microbeads added; Wjot = sample ditional counting cannot decrease the minimum stan­ weight; Mcount = number of microbeads counted. dard deviation to less than ±3% for the polystyrene microbeads and to less than 27.7% for the borosil­ 2.5.2. Filtration icate microbeads because of the error on the num­ Counting on filter membranes was done according ber of microbeads per weight unit that is given by to Cortés (1998). Calculations are based on the num­ the variance of the diameter of these microbeads ber of particles per observed area and the number of (±0.0036 p,m for polystyrene microbeads and ±0.64 coccoliths per solid weight ( C t o t ) was calculated as p,m for borosilicate microbeads). 5 A ppendix 180

J. Bollmann et al. /Marine Micropaleontology 38 (1999) 29-38 35

100

100,00

^ 10

100,00 O

100 1000 10000 Number of coccoliths counted (log scale) Fig. 3. Error estimates for the calculated number of coccoliths in a sample. The error is given in percent. The upper curves show the total error when 100 to 10,000 borosilicate microbeads were coimted. The lower curves show the same estimates for the polystyrene microbeads with a smaller error on the calculated number of microbeads per weight unit.

3. Results and 1.78 x 10'° microbeads using the particle per area based Eq. 4. This is three times the number The SMS-technique was tested with 30 repeated of coccoliths calculated from the coccoliths to mi­ preparations of the DSDP sample using polystyrene crobeads ratio for the same sample using Eq. 2 microbeads. The average number of coccoliths per (4.00 X 10'° coccoliths/g, in sample NR. 33; Ta­ gram fine-fraction sediment was 5.52 x 10'° with a ble 2), 2.2 times the average number of coccol­ standard deviation of ±0.58 x 10'° (Table 2). A sin­ iths calculated from all SMS-preparations and three gle preparation of the same sample using borososili- times the theoretical number of microbeads as calcu­ cate microbeads revealed 5.60 x 10'° coccoliths per lated from Eq. 1. gram fine fraction. The analyses of the hemipelagic Blagnac sample The filtration of the same DSDP sample (spiked revealed 0.295 x 10'° coccoliths using borosilicate with borosilicate microbeads as an independent mea­ microbeads and 0.265 x 10'° coecoliths using the sure for accuracy) revealed 4.5 x 10'° coccoliths filtration method. This estimate is 11% smaller than per gram sediment and 5.02 x 10'° borosilicate mi­ the value obtained with the SMS-method (Table 2). crobeads (calculated from the number of particles per observed area, see Eq. 3). These measurements produced 13% fewer coccoliths than estimated from 4. Discussion the coccoliths to microbeads ratio of the same sam­ ple applying Eq. 2 (5.17 x 10'° coccoliths/g, sample Coccolith counts obtained by the random settling Nr.32 in Table 2) and 9% fewer than the aver­ method applying the formula given by Williams age number of coccoliths calculated from all SMS- and Bralower (1995) are three times higher than preparations. In addition, the number of microbeads the values obtained by the SMS-method and the is 13% smaller than the theoretical number of mi­ filtration method. These high values are apparently crobeads (see Eq. 1 ). reproducible in repeated preparations (Geisen et al., Applieation of the random settling method re­ 1999), suggesting the possibility of erroneous es­ vealed 1.2 X 10" coccoliths per gram fine fraction timates of the number of borosilicate microbeads for the DSDP sample (spiked with borosilicate mi­ per gram spiking material. Therefore, we checked crobeads as an independent measure for accuracy) the number of borosilicate microbeads per gram us- 5 A ppendix 181

36 J. Bollmann et al./Marine Micropaleontology 38 (1999) 29-38

ing a lasergranulometer (GALAI CIS 1). The result of microbeads and Dr. Laube, Chemie Uetikon Ag, of 5.86E X 10^ ±25% microbeads confirmed the provided different types of uniform zeoliths. Both theoretical calculation of 5.762 x 10^ ±27.7% mi­ of them have strongly supported the project at the crobeads per gram (Appendix B). We suspect that beginning. Dr. Jahn, Potters-Ballotini GmbH, kindly convection currents within the settling box might sent 500 g of microbeads for testing purpose. 1 cause the apparent enrichment of particles on the thank Prof. H. Okada for a split of the microbeads of elevated cover slide in the settling box. This is sup­ Potter Ballotini Inc., type MB 10 and J. Giraudeau for ported by an estimate of 4.94 x 10'° coccoliths per the Blagnac sample. The manuscript profited from gram when the number of coccoliths is calculated as­ reviews of L. Beaufort and K.-H. Baumann.This suming a total water column height of 5 cm instead work is a contribution to the EC-MAST program of 2 cm in Eq. 4. This corresponds to a slide position CANIGO, EC contract No. MAS3-CT96-0060, and on the bottom of the settling device instead of being the EC-TMR program CODENET, EC contract No. elevated 3 cm. ERB-FRMX-CT97-0113. It was funded by the Swiss The SMS-method, however, yields coccolith num­ Federal Office for Education (BBW Nr. 95.0355). bers that are not significantly different from values obtained from the filtration method. The application of microbeads with different densities also gave sim­ Appendix A ilar values. We conclude that both methods and both types of microbeads are suitable for coccolith anal­ In order to calculate the variance of coccoliths per gram sed­ yses. Microbeads with small diameter variation are iment, (t^(Ctot). we used the Gaussian law of error propagation (Hartung et al., 1993, p. 326) and the variance calculations were preferable because of the resulting reduction of the done according to Hartung et al. (1993, p. 117). The variance of error estimates. the estimates of coccoliths/g sediment can be derived from infor­ mation on the variance of: (a) weighing microbeads, and (b) sediment, (^Ctot)- These errors are assumed to be negligible (<1%) because the accuracy of the Mettler balance is 5. Conclusions 10“^ g. (c) Calculating the number of microbeads per unit weight that has been added to a sample, (A/ t o t ) , which is due to vari­ Microbeads can be used in order to calibrate abso­ ance in the weight of a single microbead, is a lute abundances in combination with any commonly function of the variance of the density of the microbeads and the used method, e.g. filtration or settling. However, variance of the volume of one microbead, The variance the combination of microbeads with spraying (SMS- of the density is assumed to be zero, fr^(pM) — 0 and o^CKm) is a function of the variance of the radius, o^(r), that is given by method) is a uniquely fast, accurate and reproducible the manufacturer for different microbeads. The variance for the technique for quantitative nannoplankton analyses radius of the microbeads used is ±0.0009 tim for the polystyrene in sediment samples. It is possible to prepare sam­ microbeads and ±0.16 |im for the borosilicate microbeads. ples for SEM and LM analyses at the same time From this, the variance of the calculated number of mi­ and the distribution of particles is homogeneous and crobeads per unit weight can be derived as follows: size-independent. Additionally, potential loss of sed­ iment and thus underestimation of coccoliths is of minor importance because there is no need for sam­ 1 (fVj X a (ITm) ple splitting. Furthermore, the potential dissolution (5) of coccoliths is reduced because the contact between where liquid and sediment is very short (a few minutes). O’ (^m ) = (Pm) X a (Vm) (5.1) where:

Acknowledgements o^(Fm) = Q X jr X X 7T^ x a^(r^) (5.2)

We would like to thank Hans R. Thierstein for where: his stimulating discussions during this project. Dr. o^(r^) = o^(r X r^) = r"* x o^(r) ± x a^(r^) = Marquard, Merck Darmstadt, provided several types X CT^(r) ± r^ X (2 X X o^(r)) = 3 x r'* x cr^ir) (5.3) 5 A ppendix 182

J. Bollmann et al. /Marine Micropaleontology 38 (1999) 29-38 37

(d) The variance related to the number of coccoliths counted, Appendix B (continued) o'^(Qount). It is assumed to follow a Poisson distribution with Xc. (e) The variance related to the number of microbeads Sample Weight Microbeads Estimated number counted, cr^(Afcount)- It is assumed to follow a Poisson distri­ (g X 10" ■'') (xlO^/ml) of microbeads bution with Xm. (xlO»/g) From this the total variance of coccoliths per gram sediment, ^^(C tot), can be calculated by: 3 3.74 6.50 5.21 3 3.74 9.40 7.54

^ / a /tot \ " 3 3.74 7.80 6.26 3 3.74 6.70 5.37 3 3.74 5.90 4.73 / Q o u n t \ ( Ccount \ ^ ^2 / ^TOT \ 3 3.74 10.00 8.02 (6) \ A/count / \ A/count / \ ^CjoT J 3 3.74 10.00 8.02 3 3.74 13.00 10.43 where: 3 3.74 9.60 7.70 4 Ccount \ ^ 1 4.00 8.90 6.68 ( 4 4.00 7.50 5.63 Afcount / ^count 4 4.00 8.00 6.00 4 4.00 7.10 5.33 4 4.00 8.20 6.15 (Ccount) + ( 1 X (T ^(M c, \ Account / 4 4.00 7.50 5.63 4 4.00 8.70 6.53 4 4.00 6.70 Xc + X Xm (6 . 1) 5.03 (^) Average 5.86 and: Standard Deviation 1.47 % 25 2 / A/tot

Ctot / AfroT Y a (M tot) X (6 .2) References

Andruleit, H., 1996. A filtration technique for quantitative studies of coccoliths. Micropaleontology 42, 403-406. Backman, J., Shackleton, N.J., 1983. Quantitative biochronol­ ogy of Pliocene and Early Pleistocene calcareous nannofossils Appendix B from the Atlantic, Indian and Pacific Oceans. Mar. Micropale- ontol. 8, 141-170. Number of borosilicate microbeads per gram solid weight esti­ Beaufort, L., 1991. Adaptation of the random settling method for mated with a laser granulometer. Three samples of the borosili­ quantitative studies of calcareous nannofossils. Micropaleon­ cate microbeads were measured several times. tology 37, 415-418. Sample Weight Microbeads Estimated number Benninghoff, W.S., 1962. Calculation of pollen and spores den­ (g X 10-4) (x 10^/ml) of microbeads sity in sediment by addition of exotic pollen in known quanti­ (xioVg) ties. Pollen Spores 4, 332-333. Bollmann, J., Baumann, K.-H., Thierstein, H.R., 1998. Global 1 5.54 8.20 4.44 dominance of Gephyrocapsa coccoliths in late Pleistocene: se­ 1 5.54 8.70 4.71 lective dissolution, evolution, or global environmental change? 1 5.54 8.10 4.39 Paleoceanography 13, 517-529. 1 5.54 8.10 4.39 Cortés, M.Y., 1998. Coccolithophores at the Time Series Sta­ 1 5.54 8.10 4.39 tion Aloha, Hawaii: Populations Dynamics and Ecology. PhD 1 5.54 8.10 4.39 Dissertation, University of Zurich, 176 pp. 1 5.54 11.00 5.96 Flores, J.A., Sierro, F.J., 1998. Revised technique for calculation 1 5.54 8.60 4.66 of calcareous nannofossil accumulation rates. Micropaleontol­ 1 5.54 9.70 5.25 ogy 43, 321-324. 1 5.54 8.10 4.39 Geisen, M., Bollmann, J., Herrle, J.O., Mutterlose, J., Young, 1 5.54 14.00 7.58 J R., 1999. Calibration of the random settling technique for 1 5.54 9.70 5.25 calculation of absolute abundances of calcareous nannoplank- 5 A ppendix 183

38 J. Bollmann et al. /M arine Micropaleontology 38 (1999) 29-38

ton. J. Micropaleontol. (in press). Stockmarr, J., 1971. Tablets with spores used in absolute pollen Hartung, J., Elpelt, B., Klosener, K.-H., 1993. Statistik: Lehr- analysis. Pollen Spores 13, 615-621. und Handbuch der angewandten Statistik mit zahlreichen, voll- Su, X., 1996. Development of late Tertiary and Quaternary standig durchgerechneten Beispielen. Oldenbourg, München, coccolith assemblages in the northeast Atlantic. Geomar Rep. 975 pp. 48, 1-120. McIntyre, T.C., Be, A.W.H., Prekistas, R., 1967. Coccoliths Williams, JR., Bralower, T.J., 1995. Naimofossil assemblages, and the Pliocene-Pleistocene boundary. In: Sears, M. (Ed.), fine fraction stable isotopes, and the paleoceanography of the Progress in Oceanography 4, Pergamon, New York, pp. 3-25. Valanginian-Barremian (Early Cretaceous) North Sea Basin. Okada, H., 1992. Use of microbeads to estimate the absolute Paleoceanography 10, 815-839. abundance of nannofossils. INA Newsl. 14, 96-97. 5 Appendix 184

5.3 Xenospheres - Associations of coccoliths resembling coccospheres

This is a short research report, which was published in the Journal of Nannoplankton Research, vol. 24 (2002), pp. 27-35. As the interpretation of rare combination coccospheres is crucial to understand life-cycles of coccolithophores and to interpret their biodiversity the need for caution is highlighted in this publication with examples of accidental post-mortem associations. The author of this thesis has provided a number of scanning electron micrographs used for this manuscript and has assisted with the interpretation of the examples. 5 A ppendix 185

J. R. Young & M. Geisen: Xenospheres-associations of coccospheres...., p. 27-35. Journal of Nannoplankton Research, 24, 1,2002.

XENOSPHERES - ASSOCIATIONS OF COCCOLITHS RESEMBLING COCCOSPHERES Jeremy R. Young & Markus Geisen Palaeontology Dept., The Natural History Museum, Cromwell Road, London, SW7 5BD, UK; [email protected]

Key words: xenospheres, coccospheres, nannoplankton, biology, taxonomy

Abstract: A set of images of xenospheres are presented. These are anomalous coccospheres bearing coccoliths of two or more coccolithophore species. Unlike combination coccospheres, these are almost certainly the products of post mortem processes. Possible mechanisms of formation are discussed, and comparative examples of unambiguous agglutination by tintinnids and foraminifera, and in pellets, are illustrated. The need for caution in interpreting putative combination coccospheres is highlighted.

Introduction either in the water-column {e.g. by tintinnids: Broerse, 2000; There have been occasional records of anomalous coccospheres Winter et al., 1986) or in the sediment {e.g. by agglutinating bearing coccoliths of two or more apparently discrete species foraminifera: Murray, 1991 ; Widmark & Henriksson, 1995); (3) almost since the beginning of study of coccospheres (Lohmann, incorporation of coccoliths 'm faecal pellets. We illustrate here a 1902; Kamptner, 1941; Lecal-Schlauder, 1961). These have at­ few examples of xenospheres, both as curious anomalies which tracted a range of speculations on possible causes but recently may yet provide information on certain processes, and as exam­ it has become clear that many examples record life-cycle tran­ ples of the real need for caution. In addition, we illustrate a few sitions. Most of these life-cycle transitions record change be­ tintinnids and faecal pellets and an agglutinating foraminifera tween heterococcolith-bearing and holococcolith-bearing stages for comparative purposes. (Kleijne, 1991;Thomsen e ta i, 1991 ;Cros g/a/., 2000a; Geisen et al., submitted), whilst others combine heterococcoliths and nannoliths (Cros etal., 2000b; Sprengel & Young, 2000). These Examples Emiliania huxleyi and Gephyrocapsa oceanica combination coccospheres are providing invaluable evidence of Plate 1, Figures 1-2 life-cycle associations and so valuable insights into the ecol­ This specimen was found in a plankton sample from the Alboran ogy, phylogeny and fine-scale taxonomy of coccolithophores. Sea, western Mediterranean. It consists of numerous specimens However, as Cros et al. (2000a) noted, there are other ways in of E. huxleyi and a single specimen of G. oceanica. Although which different coccoliths can end up on single coccosphere. this looks at first glance like a regular coccosphere, two features So, careful assessment of each individual case is needed before support the interpretation of it as a xenosphere. Firstly, the E. an interpretation of a life-cycle association is accepted. huxleyi coccoliths show an anomalously wide range of varia­ tion in degree of calcification, size, and central-area structure. The phenomenon was discussed during the terminology work­ Secondly, the coccosphere also includes a piece of tubular shop at the 1991 INA Conference in Prague, where the term debris which underlies one of the E. htixleyi coccoliths and ‘xenosphere’ was proposed (by Jackie Burnett) and consequently extends between the shields of the G. oceanica coccolith. It is formally recommended in the terminology guide which eventu­ not possible to determine whether this is a coccosphere which ally followed (Young etal., 1997). It was defined there in (p.877) has accumulated additional material, or a pseudo-coccosphere as follows: ‘‘‘'Xenosphere {new, from Greek xenos, stranger} produced by some other organism, but it seems very unlikely to - anomalous coccosphere containing coccoliths normally re­ be a true coccosphere, i.e. a sphere of coccoliths produced by garded as forming on quite discrete species (e.g. Emiliania a single coccolithophore cell, or inherited from the cell which huxleyi and Gephyrocapsa oceanica'. Winter et al. 1979). N.B. gave rise to it. These are very probably artefacts, the term is suggested specifi­ cally to suggest the abnormal nature of these structures.” This A few other spheres have been illustrated, including coccoliths definition does not clearly exclude combination coccospheres, of these two species (Clocchiatti, 1971; Winter et al., 1979). so xenospheres may be better redefined as “specimens resem­ Since these are closely-related species, and since this associa­ bling coccospheres but which include coccoliths of discrete tion has been found a few times, a biological cause has been species which are unlikely to have been produced as a result of suggested. One suggested possibility was hybridisation. This a life-cycle change or hybridisation event”. By contrast, true can be categorically ruled out since hybridisation involves coccospheres may be defined as “an association of coccoliths fusion of gametes, haploid-phase cells. Flow cytometry has produced by a single coccolithophore cell, or inherited from the shown that heterococcoliths are produced on the diploid phase cells which gave rise to it”. This concept includes combination of E. huxleyi (Green et al., 1996), whilst the haploid phase is coccospheres. non-calcifying. So, even though hybridisation between G. oce­ anica and E. huxleyi is conceivable, it would give rise to a cell Possible causes for xenospheres include: (1) accidental incor­ containing organic scales characteristic of these two species, poration of loose coccoliths onto a genuine coccosphere. This not heterococcoliths. might occur in the water-column if there are large numbers of loose coccoliths. It could also obviously occur during sample An alternative suggestion is that, since E. huxleyi has evolved preparation, especially in high-density samples; (2) agglutina­ relatively recently from Gephyrocapsa (c.250kyr ago: Thierstein tion of coccoliths by a small protist. This can potentially occur et a i. 19771. the eenotvne might somehow have retained the 5 A ppendix 186

Journal of Nannoplankton Research, 24, 1, 2002. J. R. Young & M. Geisen: Xenospheres-associations of coccospheres...., p. 27-35. potential to produce coccoliths with the ancestral morphology. into the coccosphere. Thirdly, two discrete holococcolith types This type of phenomenon is known in dogs, pigeons, and other occur. Fourthly, the two main coccolith morphotypes involved intensively-bred domesticated species. However, in those cases, are known to form other associations - A. quattrospina with a time-scales are tens of years rather than hundreds of thousands previously undescribed holococcolith, and S. bannockii HOL of years, so this is biologically a rather bizarre suggestion. with a previously undescribed Syracosphaera species (Cros et Moreover, although E. huxleyi has been cultured in numerous al., 2000a). This set of anomalies leads us to conclude that this laboratories, and enormous numbers of coccoliths observed in specimen should be regarded as a xenosphere rather than as a LM, SEM and TEM, no examples of Gephyrocapsa coccolith- combination coccosphere. A possible explanation of the speci­ production have been recorded. men is that it is a small faecal pellet.

So, we do not think these examples are likely to be true coc­ Helicosphaera carieri and Syracolithus dalmaticus cospheres. We conclude that the most likely reason for the Plate I, Figure 7 recurrence of this association is simply that E. huxleyi and G. This specimen was found in a plankton sample from the Gulf of oceanica are the two most common coccolithophores, and so Mexico. It was sampled by Vita Pari en te and imaged by Claire are the most likely two species to co-occur as a result of artefact. Findlay. It consists predominantly of H. carter/heterococcoliths They also often co-occur on, for instance, tintinnids. and S. dalmaticus holococcoliths. Since H. carter/ has previously been shown to form unambiguous associations with Syracolithus Cribrocentrum reticulatum and Coccolithus pelagicus catilliferus and S. confusus (Cros et al., 2000a; Geisen et al., Plate I, Figure 3 submitted), it would not be surprising if it also formed associa­ This coccosphere was found in a Late Eocene DSDP sample tions with the very closely similar holococcolith, S. dalmaticus. from the Indian Ocean, studied in collaboration with Sivara- However, there are several reasons for regarding this as an un­ makrishnan Rabindranath. It resembles a normal coccosphere convincing combination coccosphere. Firstly, other coccoliths but contains heterococcoliths from different families, the Coc- occur on this specimen, including an E. huxleyi heterococcolith, colithaceae and Noelaerhabdaceae, which are interpreted on several Calciosolenia heterococcoliths, and one unidentifiable both conventional stratophenetic grounds and from molecular holococcolith. Secondly, diatom and other debris also occurs genetic research as only distantly related (Perch-Nielsen, 1985; on the specimen. Thirdly, the H. carter/ and S. dalmaticus coc­ Young, 1998; Edvardsen et al., 2000; Fujiwara et al., 2001). coliths are not in direct contact but separated by debris. So, there There is some breakage of the coccoliths but no other obvious are clearly at least two alternative interpretations of this speci­ evidence to support an origin by accidental means. In particular, men; either it is a combination coccosphere onto which a range it is noticeable that the coccoliths imbricate tightly. No other of other material has fallen during sample collection, or that it specimens showing this association have been reported or were is an entirely accidental agglomeration of heteromict material. found in the sample, indeed no other coccospheres were found Given this ambiguity, the specimen cannot be used as evidence in the sample at all. of an association of H. carter/ and S. dalmaticus.

Obviously, it is not possible to come up with a definitive explana­ Syracosphaera noroitica and Helladosphaera comifera tion of this xenosphere but one possibility is that it is the prolocu- Plate 1, Figure 8 lus (first chamber) of an agglutinating foraminifera. Agglutina­ This specimen was found in a plankton sample off the Canary tion of coccoliths by foraminifera is not an especially common Islands. It consists of about eight exothecal coccoliths of S. phenomenon but many examples have been documented (Wal- noroitica and ten or more holococcoliths of H. corni/era. No lich, 1877; Murray, 1991 ; Widmark & Henriksson, 1995). The other coccoliths are definitely included in the specimen, although coccosphere illustrated by Gard (1987) seems comparable, and single E. huxleyi and Discosphaera tubifera coccoliths lie a similar origin might be suggested for it. nearby. However, there are problems with accepting this as a combination coccosphere. Firstly, the S. noroitica coccoliths are Acanthoica quattrospina, Syracosphaera bannockii (in the all exothecal coccoliths whilst all definitive combination coc­ holococcolithophorid phase of the life-cycle, or HOL), and cospheres of Syracosphaera include endothecal heterococcol­ ICalyptrolithophora gracillima iths. Secondly, both the S. noroitica heterococcoliths and the H. Plate 1, Figures 4-6 corni/era holococcoliths are chaotically arranged. Thirdly, al­ This specimen was found in a plankton sample from off the though the two coccoliths types are closely associated on this Canary Islands. It consists of: (1) numerous heterococcoliths specimen, they do overlap each other and are not interspersed. of A. quattrospina, including both coccoliths and apical, spine- It is conceivable that this specimen is a combination coccosphere bearing coccoliths; (2) numerous holococcoliths of S. bannockii. but the alternative possibility, that it is an accidental association, NB These coccoliths were previously assigned to the holococ­ is at least equally possible, hence in the absence of other evi­ colithophorid species, Zygosphaera bannockii, but following dence it cannot be used to infer a life-cycle association between observation of several combination coccospheres with a previ­ these species. ously undescribed Syracosphaera species, the combination S. bannockii has been proposed (Cros et al., 2000a); (3) a few specimens of another holococcolith (on the lower left part of the Tintinnids specimen). These coccoliths show hexagonal, perforate wall­ Plate 2, Figures 1-6 structure and a flat top. They most closely resemble C. gracillima Tintinnids, marine protozoans with an external organic test but might alternatively be Calyptrolithophora papillifera. (lorica) between 45 and 1000//m long, are able to agglutinate particles such as coccoliths onto their loricae. In addition to the The specimen superficially resembles a holococcolith-het- examples given by Broerse (2000) and Winter et al. (1986), we erococcolith combination coccosphere but there are several picture here four clear (with parts of the tintinnid lorica visible reasons for doubting this interpretation. Firstly, two discrete in each case: Figures 1,3-5) and one ambiguous (tintinnid lorica holococcolith types are included. Secondly, the heterococ­ not visible, but size and shape comparable: Figure 2) examples coliths are rather irregularly arranged. In particular, three of of tintinnids covered with different heterococcoliths from dif­ the spine-bearing coccoliths are arranged with spines directed ferent plankton samples. reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 A ppendix 187

J. R. Young & M. Geisen: Xenospheres-associations of coccospheres...., p. 27-35. Journal of Nannoplankton Research, 24, 1, 2002.

Faecal pellets (Coccolithophoridés). C. r. hebd. Sednc. Acad., Paris. Série D, Plate 2, Figures 7, 8 273 : 318-321. A small (Figure 7) and a medium-sized (Figure 8) faecal pellet Cros, L., Kleijne, A. & Young, J R. 2000b. Coccolithophorid diversity in are illustrated in Plate 2. The medium-sized pellet includes a the genus Polycrater and possible relationships with other genera. J. Nannoplankton Res., 22(2): 92. diverse range of coccoliths, including, Helicosphaera carteri Cros, L., Kleijne, A., Zeltner, A., Billard, C. & Young, J.R. 2000a. New holococcoliths {‘Syracolithus catilliferus'), Rhabdosphaera examples of holococcolith-heterococcolith combination cocco­ clavigera, Syracosphaera pirus, Umbellosphaera tenuis, Cera- spheres and their implications for coccolithophorid biology. Mar. tolithus cristatus planoliths {‘Neosphaera coccolithomorpha') Micropaleontol, 29{\A)-. 1-34. and Discosphaera tubifera. There is also a dinoflagellate and Edvardsen, B., Eikrem, W., Green, J.C., Andersen, R.A., Yeo Moon-van other debris. The range of material in the pellet makes it obvious der Staay, S. & Medlin, L.K. 2000. Phylogenetic reconstructions of that it is a post mortem association but it is noticeable that the the Haptophyta inferred from 18S ribosomal DNA sequences and available morphological data. Phycologia, 39( 1): 19-35. preservation is very good - a small portion of this pellet could Fujiwara, S., Tsuzuki, M., Kawachi, M., Minaka, N. & Inouyé, 1.2001. easily cause confusion. The smaller pellet illustrates this point Molecular phylogeny of the haptophyta based on the rbcL gene and even more clearly, consisting of a large group of Helladosphaera sequence variation in the spacer region of the RUBISCO operon. spinosa coccoliths, a single inverted E. huxleyi coccolith, and a J. PhycoL, 37 : 121-129. range of unidentifiable debris. Gard, G. 1987. Observation of a dimorphic coccosphere. Abh. geol. Bundesanst., 39: 85-87. Geisen, M., Billard, C., Broerse, A.T.C., Cros, L., Probert, I. & Young, Agglutinating foraminifera J.R. submitted. Life-cycle associations involving jjairs of holococ­ Plate 3, Figures 4-6 colithophorid species: intraspecific variation or cryptic spéciation? This is a specimen of a textulariid foraminifera which has formed Eur. J. Phycol. its test almost exclusively from C. pelagicus coccoliths. Note Green, J.C., Course, P.A. & Tarran, G.A. 1996. The life-cycle of E m il­ that the initial chambers are similar in size to coccospheres iania huxleyi: A brief review and a study of relative ploidy levels analysed by flow cytometry. J. Mar. Syst., 9: 33-44. (compare coccosphere in Figures 1-3). An isolated first cham­ Kamptner, E. 1941. Die Coccolithineen der SUdwestkiiste von Istrien. ber (proloculus) would be very difficult to distinguish from a Ann. Naturh. Mus. Wien, 51: 54-149. coccosphere, unless the aperture was visible, or it included a Kleijne, A. 1991. Holococcolithophorids from the Indian Ocean, Red range of material. Sea, Mediterranean Sea and North Atlantic Ocean. Mar. M icro­ paleontol., 17: 1-76. This specimen is also of some historical interest, in that it is Lecal-Schlauder, J. 1961. Anomalies dans la composition des coques de from a slide in the G.C. Wallich collection in the Natural History flagelles calcaires. Bull. Soc. Hist. nat. Afr. N., 52: 63-66. Lohmann, H. 1902. Die Coccolithophoridae, eine Monographie der Coc- Museum. It was certainly one of the specimens he observed (the colithen bildenden Flagellaten, Zugleich ein Beitrag zur Kenntnis slide is labelled) and possibly one he illustrated (Wallich, 1861, des Mittelmeerauftriebs. Arch. Protistenk., 1: 89-165. 1877). As discussed in Siesser (1994), specimens of this type Murray, J.W. 1991. An atlas of British recent foraminiferids. Heinemann led Wallich into speculating that coccospheres were the larval Educational Books Ltd., London: 244 pp. stage of foraminifera. Perch-Nielsen, K. 1985. Cenozoic calcareous nannofossils. tn: H.M. Bolli, J.B. Saunders & K. Perch-Nielsen (Eds). Plankton Stratig­ raphy. Cambridge University Press, Cambridge: 427-555. Conclusions Siesser, W.G. 1994. Historical background of coccolithophore studies. The specimens illustrated here are all intriguing, and have indeed In: A. Winter & W.G. Siesser (Eds). Coccolithophores. Cambridge diverted us into interesting speculations. However, in each case University Press, Cambridge: 1-11. the balance of evidence suggests that they are unlikely to be true Sprengel, C. & Young, J.R. 2000. First direct documentation o f associa­ tions of Ceratolithus cristatus ceratoliths, hoop-coccoliths and coccospheres. This set of examples should serve as cautionary Neosphaera coccolithomorpha planoliths. Mar. Micropaleontol, warnings as we seek to identify combination coccospheres. 39(1-4): 39-41. Thierstein, H R., Geitzenauer, K.R., Molfino, B. & Shackleton, N.J. 1977. Global synchroneity o f late Quaternary coccolith datum Acknowledgements levels: Validation by oxygen isotopes. Geology, Boulder, Colo., These examples are drawn form research over an extended 5: 400-404. period in collaboration with numerous colleagues, including Thomsen, H.A., Ostergaard, J.B. & Hansen, L.E. 1991. Heteromorphic notably Lluïsa Cros (Inst, de Ciencias del Mar, Barcelona), life histories in Arctic coccolithophorids (Prymnesiophyceae). J. Claire Findlay (whilst working at the NHM), Annelies Kleijne P hycol, 27 : 634-642. (Free Univ., Amsterdam), Vita Pariente (Texas A&M Univ.), and Wallich, G.C. 1861. Remarks on some novel phases of organic life, and Sivaramakrishnan Rabindranath and Claudia Sprengel (Alfred on the boring powers o f minute aimelids, at great depths in the Wegener Inst, for Polar & Marine Research). Research funding sea. Ann. Mag. nat. Hist., 8: 52-58. Wallich, G.C. 1877. Observations on the coccosphere. Ann. Mag. nat. was provided by the EU (CODENET Project) and the UK’s Hist., 19: 342-350. Natural Environment Research Council. The two reviewers Widmark, J.G.V. & Henriksson, A S. 1995. The “orphaned” aggluti­ are also thanked. nated foraminifera - Gaudryina cribrosphaerellifera n. sp. from the Upper Cretaceous (Maastrichtian) Central Pacific Ocean. In: References M.A. Kaminski, S. Geroch & M.A. Gasinski (Eds). Proceedings Broerse, A. 2000. Coccolithophore export production in selected ocean o f the fourth international workshop on agglutinatedforaminifera. environments. Faculty of Earth Sciences, Amsterdam: 185 pp. Grzybowski Foundation Special Publication: 293-301. Clocchiatti, M. 1971. Sur l’existence de coccospheres portant coc- Winter, A ., Reiss, Z. & Luz, B. 1979. Distribution o f Living Coc­ colithes de Gephyrocapsa oceanica et de Emiliania huxleyi colithophore Assemblages in the Gulf of Eilat (‘Aqaba). Mar. Micropaleontol, 4: 197-223. Winter, A., Stockwell, D. & Hargraves, P.E. 1986. Tintinnid agglutina­ tion o f coccoliths: A selective or random process? Mar. Micro­ paleontol., 10: 375-379. 5 Appendix 188

Journal of Nannoplankton Research, 24, 1, 2002. J. R. Young & M. Geisen: Xenospheres-associations of coccospheres p. 27-35.

Young, J.R. 1998. Neogene. In: PR. Bown (Ed.). Calcareous Nan- Young, J R., Bergen, J.A., Bown, PR., Burnett, J.A., Fiorentino, A., nofossil Biostratigraphy. British Micropalaeontological Society Jordan, R.W., Kleijne, A., Niel, B E. van, Romein, A.J.T. & Salis, Publications Series. Chapman & Hall/Kluwer Academic, London: K. von 1997. Guidelines for coccolith and Calcareous nannofossil 225-265. terminology. Palaeontology, 40: 875-912.

Plate captions All SEM images, with the exception of Plate I, Figure 3, were taken with a Philips XL-30 fieid-emission digital SEM in the elec­ tron microscopy and mineral preparation unit (EMMA) of the Dept, of Mineralogy at the Natural History Museum, London. For Plate 1, Figure 3, a Hitachi S800 field-emission SEM was used. The files of the images have been archived and are accessible via a database; the BMNH reference numbers are given in the captions. The LM images have been captured using a Hamamatsu video camera attached to a Zeiss Axioplan light-microscope.

Plate 1 Figs 1-2: Xenosphere of Emiliania huxleyi and Gephyrocapsa oceanica. We do not have any definite explanation for this speci­ men but believe it is likely to be an artefact. Cruise MATER II, R/VHesperides, station 56 at 37.16°N, 1.19°W, depth 34m. Fig.l: BMNH 119-33; Fig.2: BMNH 119-34.

Fig.3: Xenosphere of Cribrocentrum reticulatum and Coccolithus pelagicus. This specimen is possibly the proloculus of an ag­ glutinating foraminifera. DSDP sample 220-11-1,70cm, Indian Ocean, Late Eocene (NP18). BMNH-088589.

Figs 4-6; Xenosphere of Acanthoica quattrospina, Syracosphaera bannockii (in hoiococcolithophore phase of life-cycle), and ICalyptrolithophora gracillima. We suspect this is a faecal pellet. Plankton sample from cruise P233b, R/V Poseidon, station 2 at 29.75°N, 17.93°W, depth 50m. Fig.4: BMNH 167-10; Fig.5: BMNH 167-11 ; Fig.l 1: BMNH 167-12. Collected by C. Sprengel.

Fig.7: Xenosphere of Helicosphaera carteri and Syracolithus dalmaticus. This specimen is probably an accidental association produced during sample collection. Plankton sample from the Gulf of Mexico, cruise 93-G-Ol, R/V Gyre, station 5c at 26.68®N, 95.12°W, depth 20m. BMNH CSF0195. Collected by V. Pariente.

Fig.8: Xenosphere of Syracosphaera noroitica and Helladosphaera corni/era. Again, this specimen is probably an accidental associa­ tion produced during sample collection, but conceivably could be a combination coccosphere. Plankton sample from off the Canary Islands, cruise P233b, R/V Poseidon, station 2 at 29.75°N, 17.93°W, depth 50m. BMNH 126-20. Collected by C. Sprengel. reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 189

/. R. Yoiiiif’ M. Geisen: Xenospheres-associations of coccospheres p. 27-35. Journal of Nannoplankton Research. 24, I, 2002.

/ ; iïV^'

% .„v> .a»

V w "//UV' » < r f 5 Appendix 190

Journal of Nannoplankton Research, 24, 1, 2002. J. R. Young & M. Geisen: Xenospheres-associations of coccospheres p. 27-35.

Plate 2 Figs 1- 6: Tintinnids with agglutinated coccoliths, all from plankton samples. Figures W : South Atlantic Ocean, off Namibia, cruise M48-4b, R/V Meteor. Fig.l: station 11 at 20.6“S, 9.87°E, depth 5m; Figs 2-3: station 477 at 23.46°S, 12.62°E, depth 5m; Fig.4: station 44 at 30.15“S, 4.43°E, depth 5m. Figs 5-6: Alboran Sea, western Mediterranean, cruise MATER II, R/V Hesperides, station 69 at 37.43°N, 0.43°W, depth 50m. Fig.l: BMNH 136-35; Fig.2: BMNH 137-17; Fig.3: BMNH 137-18; Fig.4: BMNH 137-20; Fig.5: BMNH 145-27; Fig.6: BMNH 145-28.

Figs 7-8: Faecal pellets. Cruise P233b, R/V Poseidon, station 2 at 29.75°N, 17.93°W, depth 50m. Fig.7: BMNH 118-27; Fig.8: BMNH 126-16. Collected by C. Sprengel.

reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 191

J. R. Youn^ & M. Geisen: Xenospheres-associations of coccospheres...., p. 27-35. Journal of Nannoplankton Research, 24, 1, 2002.

f

'uytÊÊÊÊÊÊB

i t 5 Appendix 192

Journal of Nannoplankton Research, 24, 1, 2002. J. R. Young, & M. Geisen: Xenospheres associations of coccospheres p. 27-35. i

S Ÿ 10

Plate 3 Foraminifera agglutinating Coccolithus pelagicus. Surface sediment sample from south of Iceland, collected by G,C. Wallich dur­ ing Brr/A/rrg cruise, station 120, in I860.

Figs 1-3; Coccosphere of Coccolithus pelagicus. Fig. 1 : crossed-polars; Fig.2: bright field, surface focus; Fig.3: Bright field, median focus.

Figs 4-6: Agglutinating foraminifera. Fig.4: complete specimen; Fig.5: terminal chamber, surface focus, bright field; Fig.6: terminal chamber, surface focus, crossed- polars.

reproduced from journal of Nannoplankton Research, 24, I, 2002 5 A ppendix 193

5.4 Coccolithophores for exhibition : A note

In 2001 the author of this thesis won a first prize in the “science close up” category of a photography competition sponsored by Novartis and the Daily Telegraph. Past and present winning images can be seen at http://www.visions-of-science.co.uky. Subsequent to this the Natural History Museum in London decided to devote a temporary exhibition solely to coloured scanning electron microscope images of coccolithophores. This contribution to the public understanding of science has been planned and edited in collaboration with an exhibition designer and an editor at the NHM. These and further images are available in high resolution at the NHM picture library. Feel free to subsidise the author and buy some images at http:// www.nhm.ac.uk/services/piclib/index.html. The author of this thesis has captured the majority of the scanning electron micrographs, has false coloured the images and written the text in collaboration with Jeremy Young.. 5 A ppendix 194

M. Geisen, I. Probert & J. Young: Coccolithophores for exhibition, p. 3-7 Journal of Nannoplankton Research, 24, 1, 2002

COCCOLITHOPHORES FOR EXHIBITION: A NOTE Markus Geisen*, Ian Probertf & Jeremy R. Young* *Palaeontology Dept., The Natural History Museum, London, SW7 5BD, UK; flab, de Biologie & Biotechnologies Marine, Univ. de Caen Basse Normandie, Caen, France; [email protected]

Introduction a wide range of morphologies and some of the more interesting When we work with coccolithophores, we sometimes overlook structures was made. A TEM section was included to provide a their beauty, strangeness and visual impact. However, recently graphic illustration of the relationship between the coccosphere their visual qualities were amply confirmed, after one of us (MG) and the cell, and also a coccolithophore bloom image (courtesy spent an afternoon false-colouring an SEM image for a science of S. Groom, PME) to highlight the potential ecological impact photography competition: the resultant image won First Prize of coccolithophores. in the ‘Science Close Up’ category of the Novartis/Daily Telegraph Visions of Science competition. Further information The images were false-coloured both to increase the visual about the competition, and this winning image, can be viewed impact and to highlight specific features for text-reference. The at www.visions-of-science.co.uk. Through this medium, non­ images have been enlarged to 1 m across for the exhibition and scientists got their first vision of the beauty of these minute make a very attractive display, which the visitors seem to like. works of art. The plate captions presented here are essentially the same as those in the exhibition and reflect how we have tried to By a happy coincidence, The Natural History Museum has encourage public understanding of our science, although there recently devoted space to temporary exhibitions of artwork are limits to what you can do in the number of words allowed. related to natural history. Consequently, we were asked to The exhibition will be displayed in the public galleries until prepare a temporary exhibition of coccolithophore images for March/April and after that will probably move to our common it. In collaboration with an editor and exhibition designer, we room. If any other INA members would like an opportunity to selected 14 images of coccospheres. Selection was partly based use the exhibition for public display we would be pleased to on image quality and attractiveness but, also, an attempt to show hear from you.

reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 195

Journal o f Nannoplankton Research, 24, 1, 2002 M. Geisen, I. Probert & J. Young: Coccolithophores for exhibition, p. 3-7

Plate 1

Fig.l; Calcidiscusleptoporus Coccolithophores mostly reproduce asexually by simple binary cell-division. But sexual reproduction does occasionally occur and is marked by production of haploid daughter-cells covered by holococcoliths. Following fusion of two daughter-cells, a diploid cell is produced and the coccolith type reverts to heterococcolith. This award-winning image provides rare documentation of this process, as the diploid cell with its cover of heterococcoliths (in pink) emerges from within the cover of the fused daughter-cells (in green). For a detailed figure explaining the life-cycle refer to Plate 2, Figure 7. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 69).

Fig.2: Syracosphaerapulchra Coccolithophores can produce dilferent scales to cover their surface. This cell has an inner, endothecal layer (in pink), with coccoliths bearing spine-like protrusions around the flagellar pole (top right), and an outer, exothecal cover (in green). This specimen is from the Canary Islands, North Atlantic (R/V Poseidon cruise P233b, station 3). Image courtesy of Claudia Sprengel, Alfred Wegener Institute, Bremerhaven, Germany.

Fig.3: Syracosphaera nodosa In this specimen, the wheel-like exothecal coccoliths (in green) have fallen off and are lying on the filter. Again, the inner coccoliths (in pink) show spine-like protrusions around the flagellar pole. This specimen is from the Alboran Sea, western Mediterranean (R/ V Hesperides cruise Mater II, station 61).

Fig.4: Syracosphaera anthos This is a rare example o f Syracosphaera anthos with the exothecal cover of coccoliths (in green) almost completely in place. Species with exothecal coccoliths are very fragile and often disintegrate during the filtration process. This specimen is from the North Atlantic (station FBI 6).

Fig.5: Algirosphaera robusta This species is of interest since it lives in the deep photic zone, more than 50m below the surface, and is therefore poorly known. A single cell was recently isolated from Mediterranean sea-water and successfully maintained in culture, which has allowed detailed study of its morphology and behaviour. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 69).

Fig.6: Algirosphaera robusta This very different image was taken with a TEM, and shows the internal structure of a coccolithophore. Two types of heterococcolith cover the cell - some are hood-like (in pink) and others elongated (in green) around the flagellar pole. The interior is dominated by a large chloroplast (in dark green), the light-collecting motor of the cell. Above this sits the nucleus (in pink), which contains the organism’s genetic material. In addition, a coccolith in production can be seen inside the cell. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 69). Image courtesy of Ian Probert, Université de Caen Basse- Normandie, Caen, France.

Fig.7: Emiliania huxleyi This is perhaps the best known and most intensively researched coccolithophore. With the exception of the Southern Ocean around Antarctica, this species can be found in all oceans. This single species can form massive blooms (see Plate 2, Figure 6), the lateral extent of which can be very broad and, with the ability to calcify and fix biomass, this species, and coccolithophores in general, contributes significantly to the carbon cycle. This specimen is from the Canary Islands, North Atlantic (R/V Poseidon cruise P233b, station 2).

Fig.8: Reticulofenestra sessilis This species (in green) forms unique symbiotic associations with the diatom Thalassiosira (in pink). This specimen is from the G ulf of Mexico (R/V Gyre cruise 90-G-15, station 9).

reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 196

M Geisen. I. Prohert & J. Young: Coccolithophores for exhibition, p. 3-7 Journal of Nannoplankton Research, 24. 1. 2002

m m

Syraco sphaera pulchraCalcidiscus leptoporus Syracosphaera pulchraCalcidiscus

• s

Syracosphaera nodosa

Algirosphaera robusta Algirosphaera robusta

s ''

m

Emiliania huxleyi Reticulofenestra sessHis reproduced from Journal of Nannoplankton Research. 24. 1. 2002 5 A ppendix 197

Journal o f Nannoplankton Research, 24, 1, 2002 M. Geisen, I. Probert & J. Young: Coccolithophores for exhibition, p. 3-7

Plate 2

Fig.l: Discosphaera tubifera The coccoliths produced here are trumpet-shaped. The actual cell is very small in comparison to the coccoliths. The cell diameter is less than the length of the coccolith, thus the coccolith is produced inside the cell with its base folded, which causes considerable stretching for the cell. This specimen is from the Alboran Sea, western Mediterranean (RA7 Hesperides cruise Mater II, station 69).

Fig.2: Rhabdosphaera clavigera Rhabdosphaera (in green) forms long protrusions from the centre of the coccolith, and the precise architecture of these spines can be seen. A coccosphere of Emiliania huxleyi (in pink) can be seen in the background. These specimens are from the Alboran Sea, western Mediterranean (RA^ Hesperides cruise Mater II, station 69).

Fig.3: Gephyrocapsa ornata This species reflects the ornamentation, with an elevated bridge spanning the central area (in pink). Whilst a number of Gephyrocapsa species have been described, and Gephyrocapsa oceanica is distributed globally, Gephyrocapsa ornata is very rare. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 44).

Fig.4: Michaelsarsia elegans Michaelsarsia belongs to a group of coccolithophores the most striking feature of which is the presence of long appendages (in pink) around the flagellar pole. These are highly modified coccoliths, the function of which, however, is unknown. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 69).

Fig.5: Periphyllophora mirabilis This species represents another example of a holococcolith. The hole surrounded by the coccoliths (in green) is the flagellar opening. Creating a good micrograph of holococcoliths is difficult, as they are rare in the plankton and tend to disintegrate rapidly. This specimen is from the Alboran Sea, western Mediterranean (R/V Hesperides cruise Mater II, station 15).

Fig.6: Emiliania huxleyi bloom off Cornwall Under certain conditions, Emiliany huxleyi can form massive blooms which can be detected by satellite remote sensing. What looks like white clouds in the water, is in fact the reflected light from billions of coccoliths floating in the water-column. Image courtesy of Steve Groom, Plymouth Marine Laboratories.

Fig.7: Coccolithophorid life-cycles Schematic representation of coccolithophorid life-cycles. The diploid stage of Calcidiscus leptoporus is covered with heterococcoliths (left, in pink) which are produced inside the cell, whereas the motile stage is covered with holococcoliths (right, in green) which are produced outside of the cell membrane.

reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 198

M. Geisen. I. Prohert & J. Young: Coccolithophores for exhibition, p. 3-7 Journal of Nannoplankton Research, 24, 1, 2002

Wi'-U Rhabdosphaera clavigera

■V'-.-'C. - : V.v Gephyrocapsa ornata Michaelsarsia eiegans

w #

Periphyllophora mirabilis Emiliania huxieyi b\oom

Coccolithophorid life cycles

Coccolithophorid life-cycles reproduced from Journal of Nannoplankton Research, 24, 1, 2002 5 Appendix 199

Bibliography Markus Geisen

Bollmann J, Brabec B, Cortés M Y, Geisen M (1999) Determination of absolute coccolith abundances in deep-sea sediments by spiking with microbeads and spraying (SMS- method). Mar Micropaleontol 38; 29-38 Geisen M (2003) Studies on the species level variation of selected coccolithophores. PhD thesis, University College London, p. 199 Geisen M, Billard C, Broerse A T C, Cros L, Probert I, Young J R (2002) Life-cycle associations involving pairs of holococcolithophorid species: intraspecific variation or cryptic spéciation? Eur J Phycol 37: 531-550 Geisen M, Bollmann J, Herrle J O, Mutterlose J, Young J R (1999) Calibration of the random settling technique for calculation of absolute abundances of calcareous nannoplankton. Micropaleontology 45: 437-442 Geisen M, Young J R, Probert I (subm.-a) Syracosphaera pulchra -a model species to study fine scale variation and spéciation in coccolithophores? J nannoplankton Res Geisen M, Young J R, Probert I, Saez A G, Baumann K-H, Bollmann J, Cros L, de Vargas C, Medlin L K, Sprengel C (subm.-b) Species level variation in coccolithophores. In: Thierstein H R, Young J R (eds) Coccolithophores - From molecular processes to global impact. Springer Saez A G, Probert I, Geisen M, Quinn P, Young J R, Medlin L K (2003) Pseudo-cryptic spéciation in coccolithophores. Proc natn Acad Sci USA 100: 7163-7168 Stoll H M, Ziveri P, Geisen M, Probert I, Young J R (2002) Potential and limitations of Sr/Ca ratios in coccolith carbonate: new perspectives from cultures and monospecific samples from sediments. Phil Trans R Soc Ser A 360: 710-747 Young J R, Geisen M (1998) Using spreadsheets to produce stacked histogram, stacked line and spindle charts. J Micropal 17: 104 Young J R, Geisen M (2002) Xenospheres - associations of coccoliths resembling coccospheres. J nannoplankton Res 24: 27-35 Young J R, Geisen M, Cros L, Kleijne A, Sprengel C, Probert I (subm.) A guide to extant calcareous nannoplankton taxonomy. J nannoplankton Res Ziveri P, Stoll H M, Probert I, Klaas C, Geisen M, Ganssen G M, Young J R (2003) Stable isotope "vital effects" in coccolith calcite. Earth planet Sci Lett 210: 137-149