Reproduction strategies of stony () in an equatorial, Indonesian . Contributions for the reef- restoration

Dissertation zur Erlangung des Doktorgrades der Naturwissenschaften – Dr. rer. nat. –

im Fachbereich 2 (Biologie/Chemie) der Universität Bremen

vorgelegt von

Sascha Bernd Carsten Romatzki

angefertigt am Zentrum für Marine Tropenökologie Center for Tropical Marine Ecology Bremen 2008

Gutachter der Dissertation:

1. Gutachter: Prof. Claudio Richter 2. Gutachter: Dr. Andreas Kunzmann

Mitglieder der Prüfungskommssion:

1. Prüfer: Prof. Kai Bischof 2. Prüfer: Dr. Uwe Krumme

1. weiteres Mitglied: Cornelia Roder 2. weiteres Mitglied: Anne Buhmann

Tag des öffentlichen Kolloquiums: 30. Januar 2009

„So eine Arbeit wird eigentlich nie fertig, man muss sie für fertig erklären, wenn man nach Zeit und Umständen das Möglichste getan hat.“

(Goethe; zur Iphigenie, aus der Italienischen Reise 1786) List of Papers

This thesis is based on the scientific publication listed below. The specific contributions of each of the authors in term of idea and concept development, data aquisition and analysis, as well as manuscript writing for the respective publication are indicated.

Publication I Title: Determination of fine-scale temporal variation in acroporid and pocilloporid settlement in North ,

Authors: S.B.C. Romatzki and A. Kunzmann Journal: Marine Ecology Progress Series (submitted) The original idea and concept of this publication was developed by S.B.C. Romatzki, who also independently conducted all of the fieldwork and sample processing. Data analyses were carried out by S.B.C. Romatzki with input by S. Schmidt-Roach. The manuscript was written by S.B.C. Romatzki, with revesions and improvement by A. Kunzmann.

Publication II Title: Gametogenesis of four scleractinian corals in the

Authors: S.B.C. Romatzki and A. Kunzmann Journal: Proceedings of the 11th International Symposium (submitted) The original idea and concept of this publication was developed by S.B.C. Romatzki, who also independently conducted all of the fieldwork and sample processing. The manuscript was written by S.B.C. Romatzki, with revesions and improvement by A. Kunzmann.

Publication III Title: Determining the influence of an electrical field on the performance of transplants: Comparison of various transplantation structures

Author: S.B.C. Romatzki Journal: Journal of Experimental (submitted) The original idea and concept of this publication was developed by S.B.C. Romatzki, who also independently conducted all of the fieldwork and sample processing. The concept of experiment 3 in this publication was developed by S.B.C. Romatzki, S.C.A. Ferse and E. Borell. Data analy- ses were carried out by S.B.C. Romatzki with input by E. Borell. The manuscript was written by S.B.C. Romatzki, with revesions and improvement by M. Schmid, K. von Juterzenka and J.-H. Steffen.

Content

Acknowledgements………………………………………………………………….……….....I Summary…………………………………………………………………………….………..III Zusammenfassung…………………………………………………………………….……...VI

Chapter 1 – Thesis Overview……………………………………………………….………....1 1. General Introduction…………………………………………………………….……….....3 1.1 Reproduction in scleractinian corals……………………………………….………..3 1.1.1 Asexual reproduction……………………………………………….……..3 1.1.2 Sexual reproduction……………………………………….…….………...4 1.1.3 Mass spawning…………………………………………………...... 4 1.1.4 Reproduction in high latitude vs equatorial reefs………………...... 5 1.2 Reef restoration and reef rehabilitation……………..…………………….……..….6 1.2.1 Pros and cons……………………………………………………..………6 1.2.2 What is “restoration” and what is “rehabilitation” then……….…….….7 1.2.3 Old and new approaches………………………………………….……....7 1.2.4 Using of coral fragments………………………………………….……....9 1.3 Aims of the thesis…………………………………………………………….……..9 2. General Material & Methods……………………………………………………….…….10 2.1 Study Area………………………………………………………………....….…...10 2.2 Settlement Tiles……………………………………………………………….…...11 2.3 In situ monitoring of reproduction status…………………………………….....…12 2.4 Histological preparation for examination of reproduction status……………….....13 2.5 Surface estimation and zooxanthella-count of coral fragments………………..….14 3. General Results & Discussion…………………………………………………………….14 3.1 Reproduction and settlement in equatorial reefs: continuous with seasonal peaks……………………………………………………………………….…..…14 3.2 Manipulation of coral longitudal growth enhancement………………………...... 16 4. Concluding remarks………………………………………………………………….…....17 5. References………………………………………………………………………….………19

Chapter 2 – Paper I………………………………………………………………………….27 Determination of fine-scale temporal variation in acroporid and pocilloporid settlement in , Indonesia

Chapter 3 – Paper II………………………………………………………………………...50 Gametogenesis of four scleractinian corals in the Celebes Sea Chapter 4 – Paper III………………………………………………………………………..64 Determining the influence of an electrical field on the performance of Acropora transplants: Comparison of various transplantation structures

Appendix……………………………………………………………………………………..92 Disclaimer…………………………………………………………………………………....94 Acknowledgements

I would like to thank Prof. Matthias Wolff for agreeing to supervise this thesis in first place, but special thanks go to Prof. Claudio Richter, who took over the supervision in the very last minute due to a very time schedule. Thanks a lot to both for making the change so uncomplicated. I further would like to thank Dr. Andreas Kunzmann, who implemented the idea to conduct a PhD thesis in Indonesia and encouraged me to step into the unknown, for his guidance and criticism throughout the thesis.

My deepest gratitude goes to some very special people in for their generous sponsoring and support. Without their help I never would have been able to conduct this thesis: Thanks to Christiane Mueller from Froggies Divers who gave me shelter from the first day of my arrival, who showed me suitable sites for my project, who gave me my first lessons in Manado Culture and lifestyle. Not to forget the uncountable tank-fillings and meals. I also would like to thank Hanne Darbol and Gaspare Davi from Gangga Resort and Spa for their generous hospitality, their always warm and heartfelt welcome, their unlimited support, interest and open minds for new project ideas. Staying at your places was always a feeling of coming home with a touch of holiday. Thanks also to all the great staff of these wonderful resorts who always were very helpful.

Special thanks to Sebastian Ferse, my roommate, dive buddy and colleague, who was often the only source of inspiration and scientific input. Despite many difficulties during all the time we spend together, I for sure wouldn’t have been able to conduct my project in Manado without him.

Sebastian Schmidt – our MSc student, who came at the “right” time – I would like to thank a lot for many delicious pancakes…well and some other things of course – but they were not that important ;)…. Thanks also to Leyla Knittweis and in particular Esther Borell, for collective suffering and many encouraging phone calls.

Further thanks are dedicated to Fontje Kaligis and family for taking care of me during the first weeks of my arrival and their help in getting my life organized in Manado. Prof. Eddy Tambajong for his generous lessons in histological technique and for letting me use his private laboratory. Tere and Mer from UNSRAT for their assistance with coral tissue preparation and histological sectioning.

Thanks also to those people, who helped me at one point or another in Manado: Dr. Ineke Rumengan, who helped me with organizational issues at UNSRAT-University, Asci and Francesco for a relaxing time on their liverboard, Angelique & Paul from Murex and Jaakko from Living Colours for help in dive logistics as also Jan & Henriette Bebe for their European humor and Danish “friendliness”.

I Following people from IPB I would like to thank: Totok Histirianato for excellent work and help with the Visa, Unggul Aktani for supervision, Neviaty Zamani and Pak Adi for the organization and use of the microscope camera.

I further would like to thank the German community in Indonesia, starting with Harry Palm and family for their hospitality and Harry’s scientific advice, Karen von Juterzenka & Michael Schmidt for several weeks of accommodation, supporting words, the organization of a microscope, numerous corrections on the papers and their friendship, Jan-Henning Steffen for being a port in stormy seas and a second home during times in Jakarta, Landry Pramoedji and Lutz Kleeberg for letting me stay in their house for such a long time during the final steps of this thesis.

I also would like to thank Eberhard Krain for being the knight in shining when help was most needed.

Andrew Baird, Sangeeta Mangubhai and Nami Okubo gave many helpful comments, corrections and clarified questions regarding coral recruitment and histology. Thanks for your very prompt answers and provision of your papers.

This study was supported by a PhD scholarship of the DAAD (German Academic Exchange Service). I especially would like to thank Frau Krüger-Rechmann and Brigitte Gerlach for their support during the time of the scholorship.

My Carol I would like to thank for accompanying me during bad and better times and many culinary experiences making the work on this thesis a whole lot easier.

And last but not least my dear parents Helga & Paul Romatzki who gave me all the support on earth to fulfill my dream. Trotz der vielen freundlichen Hilfe und Unterstützung, die ich von dritter Seite erfahren habe, wart letztendlich Ihr es, die diese Arbeit von Anfang bis Ende ermöglicht habt. Danke für Euren nie endenden Zuspruch, Eure Geduld und Eure Unterstützung.

II Summary

In the present dissertation the sexual and asexual reproduction-strategies of stony corals in an equatorial Indonesian coral reef will be examined. Special attention is applied to the seasonality of sexual reproduction through coral larvae in an important, but yet by coral literature disregarded geographical region, the so-called “”. A crucial criterion for countries to be accepted as members were coral reef regions with a coral diversity of more than 500 . This high diversity is unique worldwide and is not even surpassed by the world famous and it’s circa 360 species of stony corals. The outlines of this triangle are marked by the Philippines in the North, Malaysia and Indonesia in the South and West Guinea and Papua New Guinea in the East. Sulawesi, as one of the three biggest main islands of Indonesia, is located in the middle of the triangle. The present work examined reefs in and around the National Park in North Sulawesi in the Celebes Sea. The research area was chosen, as there is only few or almost no existing literature covering the reproduction timing and cycles of south-east Asian equatorial reefs – especially in Indonesia. Furthermore the here examined reefs are still in good condition as the implementation of the in 1991 lead to a strong decrease of dynamite fishing compared to other parts of Indonesia where it is still very common. Also the industrialization in this region is not as progressive as to be noticeably negative for the surrounding reefs. Two mainly cause anthropogenic disturbances as they are transporting garbage and wastewater from the seaside located town of Manado. The examined reefs were located outside the influencing factors due to the course of the . The first chapter of this work deals with the explanation of terms that are often used in context with the present topic. Generative sexual and vegetative asexual reproduction are specified. The asexual reproduction through the so-called “budding” plays an important role in growth and regeneration of corals and coral fragments. Coral fragments on the other hand have an important function in reef restoration and rehabilitation projects. The differences between reef restoration and rehabilitation will be examined closer as both terms are often used in the wrong context. Both are manipulating interventions that differentiate from each other as the former aims for the previous natural condition while latter often uses elements and structures that are not equivalent to the natural condition. The second chapter deals with the settlement of coral larvae on artificial settlement tiles in a quantitative study. Three settlement frames were permanently installed in each of four reef sites at a depth of 5 m with a distance of 25 to 50 m between each other. Each frame contained 12 tiles that were replaced at two-month intervals. The examination of the bleached tiles showed, that the primary settlers were larvae of and Pocilloporidae. So-called growth charts were used to determine the time of settlement of Acroporidae and Pocilloporidae recruits with weekly accuracy. Although larvae of both families were found year round, those of the Acroporidae

III showed clear seasonal patterns. An intensified increase in acroporid recruits could be noticed during the April weeks of 2006 and 2007 as also between May and June 2006 and 2007. Settlement plates were able to give information about the peak reproduction times of coral families but not of single species, as current techniques of coral recruit identification for this application are not reliable enough. Chapter 3 tried to close this information gap by focusing on the reproduction strategy and cycles of four selected coral species by histological examination of their tissues. These species were chosen after a visual assessment of the examined reef sites in terms of their abundance and the possibility for easy and multiple sampling. Therefore all chosen species were those with a branching growth form. Whereas Seriatopora hystrix is a brooder, Pocillopora verrucosa, Acropora yongei and Acropora pulchra are known to be spawners. The coeval presence of planula-larvae, ovaries and spermaries in all S. hystrix-specimens from different sampling times referred to a year-around brooding in the examined sites. Ripe eggs, which are spawned together with spermaries as reddish egg-sperm-bundles, could only be found in A. yongei in June 2006 and May 2007 and in A. pulchra in April/May 2006 and March 2007 - suggesting a single annual reproduction cycle. The presence of zooxanthellae in P. verrucosa is a indicator for a imminent spawning, as zooxanthellae appear three to four days before this event in eggs of this species. Zooxanthellae were found in eggs in January and March, as also between July and August suggesting there is a biannual gametogenic cycle for this species in North Sulawesi. The fourth and last chapter deals with the hypothesis of enhanced growth of coral fragments transplanted in and exposed to an electrical field. The hereby examined technique is based on the saltwater electrolysis that uses a low current between an anode and a cathode both submerged in saltwater. The cathode is used as an artificial reef structure. Due to a higher Ion- at the cathode, precipitation takes place on its surface in the form of lime scale analog to natural limes stone as found in the reef. With accelerating coating the structure gains size and stability. The rough surface of the accretion is a suitable settlement substrate for marine larvae. The unlimited possibilities in cathode design allow a wide range for an application in reef rehabilitation. So far the observed enhancement of coral growth by the inventors remains in question and has lead to controversy among scientists as only very few studies have been published in regards to this topic. Common consensus is that enhanced coral growth is due to the enhanced Ion-concentration and the resulting increase of the pH-value in approximate distance to the cathode. Finger sized fragments of A. yongei and A. pulchra were transplanted on three different constructions for the clearance of different aspects. In the first experiment two tunnel shaped structures were connected to two different electrical currents to test if intensity of current has an effect on the growth of corals. In a second experiment fragments were transplanted to table-like structures each with different heights. Here differences in growth performance were tested to find the ideal transplantation elevation. In a third experiment fragments transplanted in direct contact to the cathode material were compared to those only transplanted within an electrical field and without direct cathode contact. Controls

IV with similar designs to the cathodes but without electricity were used for all three experiments. Furthermore growth rates of branches within the donor colonies were measured. Results of all experiments showed no unusual increased growth for the used transplants. A direct comparison with “uncharged” controls showed no or negative significant growth differences. A comparison of growth data from the transplants and remaining branches from the donor colonies showed significant lower growth rates for the transplants. Best results were achieved with the lowest current of 7 A (approximately 1.67 A m-2) and when fragments were exposed to the electrical field but not in direct contact to the cathode or exceeding 1.67 A m-2. The results also showed that transplants performed differently between experiments. These results differ significantly from the observations of the developers, who claimed five to twenty times higher coral growth on electrical structures. Insights gained from these experiments will help to facilitate the planning of future rehabilitations projects when making decisions on transplant species, size, cathode design and operational current.

V Zusammenfassung

In der vorliegenden Dissertation werden die sexuellen als auch die asexuellen Reproduktionsstrategien von ausgewählten Steinkorallen in einem äquatornahen, indonesischem Korallenriff untersucht. Besondere Aufmerksamkeit galt der Saisonalität der sexuellen Vermehrung durch Korallenlarven in einer wichtigen, aber bisher in der Korallenliteratur vernachlässigten geographischen Region, dem so genannten Korallendreieck. Als ausschlaggebendes Kriterium für die Aufnahme von ausgewählten Gebieten der Mitgliedsstaaten galt eine Korallenvielfalt von mehr als 500 Arten. Diese Artenvielfalt ist weltweit einzigartig und kann auch von dem weltbekannten Great Barrier Reef mit seinen circa 360 Steinkorallenarten nicht überboten werden. Die heutigen Mitgliedstaaten der Philippinen im Norden, Malaysia und Indonesien im Westen und Papua Neuguinea im Osten bezeichnen die Grenzen des Dreiecks. Sulawesi als eine der drei großen Hauptinseln Indonesiens, befindet sich in der Mitte dieses Dreiecks. Die in der vorliegenden Arbeit untersuchten Riffe liegen im und um den Bunaken Nationalpark im Norden Sulawesis in der Celebes See. Das Untersuchungsgebiet wurde auf Grund der wenigen bzw. kaum existierenden Veröffentlichungen bezüglich der Vermehrungszeiten und –zyklen von äquroerialen Korallen - im Besonderen in indonesischen Riffen - gewählt. Des Weiteren befinden sich die Riffe in einem guten Zustand da vor allem durch die Implementation des Nationalparks 1991 das Dynamitfischen im Verhältnis zu anderen indonesischen Riffen stark zurückgegangen ist. Auch ist die Industrialisierung in dieser Region noch nicht soweit fortgeschritten, als das diese sich merkbar negativ auf die umliegenden Riffe auswirken. Anthropologe Störungen sind vor allem auf zwei Riffmündungen und erhöhte Mülleinträge der anliegenden Stadt Manado zurückzuführen. Die untersuchten Riffe lagen auf Grund des Strömungsverlaufs außerhalb dieser Einflussfaktoren. Das erste Kapitel dieser Arbeit beschäftigt sich mit der Erklärung von Begriffen, die häufig in Verbindung mit der vorliegenden Thematik genannt werden. Es wird auf die generative sexuelle als auch auf die vegetative asexuelle Vermehrung eingegangen. Die asexuelle Vermehrung durch das so genannte “budding” spielt eine wesentliche Rolle im Wachstum und der Regenerierung von Korallen und Korallenfragmenten. Korallenfragmente wiederum spielen eine große Rolle bei Riffrestaurierungs und –rehabilitierungsmaßnahmen. Die Unterschiede zwischen Riffrestaurierung und Riffrehabilitierung werden auf Grund häufiger Verwechslung näher erläutert. Beide sind manipulierende Eingriffe, die sich vor allem darin unterscheiden, dass ersteres eine Rückführung des alten Zustands anstrebt, während letzteres sich Elemente und Strukturen bedient, die nicht dem alten, natürlichen Zustand entsprechen. Im zweiten Kapitel wird das Ansiedeln von Korallenlarven auf ausgebrachten Kalksteinplatten quantitative untersucht. Hierzu wurden jeweils drei Besiedlungsrahmen in einem Abstand von 25 bis 50 Metern und einer Tiefe von 5 m in vier verschiedenen Stationen installiert. Jeder Rahmen fasste 12 Platten die in 2-monatigen Intervallen gegen neue ersetzt wurden. Die Untersuchung

VI der gebleichten Platten zeigte, dass vor allem Larven von Acroporidae und Pocilloporidae sich ansiedelten. Mit Hilfe von Wachstumsschlüsseln für diese Familien, konnte die Ansiedlungszeit der untersuchten Rekruten auf eine Woche genau berechnet werden. Obwohl zwar Larven beiden Familien sich ganzjährig ansiedelten, zeigten diese der Acroporidae eindeutig saisonelle Peaks. Zu einem verstärkten Auftreten von Acropora- Rekruten kam es jeweils in den April-Wochen 2006 und 2007 sowie zwischen Mai und Juni 2006 und 2007. Da die sichere Bestimmung von Korallenrekruten auf Spezies-Level auf Grund der geringen Größe zum jetzigen Wissensstand unmöglich ist, konnten keine konkreten Angaben auf Grundlage von Besiedlungsplatten bezüglich der Laichzeiten für einzelne Arten getroffen werden. Histologische Untersuchung im dritten Kapitel zeigen daher an Hand von vier ausgewählten Steinkorallen deren Fortpflanzungsstrategie und Entwicklungszylen. Bei der Auswahl der Arten wurden solche gewählt, die nach visueller Einschätzung der untersuchten Riffe am häufigsten in den untersuchten Gebieten gefunden werden konnten und eine einfache, multiple Probenentnahme erlaubten. Alle Arten gehören daher zu den Astbildenden, wobei Seriatopora hystrix als Brüter sowie Pocillopora verrucosa, Acropora yongei und Acropora pulchra als Laicher bestätigt werden konnten. Die gleichzeitige Anwesenheit von Planulalarven, Ovarien und Spermarien in allen S. hystrix-Präparaten weisen auf eine kontinuierliche Vermehrung in dem untersuchten Gebiet hin. Reife Eier, die zusammen mit Spermien als rote Ei-Sperma-Bündel gelaicht werden, wurden in Acropora yongei in den Monaten Juni 2006 und Mai 2007 und in A. pulchra im April/Mai 2006 und März 2007 gefunden, was auf ein einmaliges Ablaichen im Jahr hinweist. Die Anwesenheit von Zooxanthellen in P. verrucosa-Eier sind Anzeichen für ein kurzbevorstehendes Ablaichen, da Zooxanthellen etwa 3 bis 4 Tage vor dem Ereignis in den Eiern auftauchen. Zooxanthellen wurden zwischen Januar und März, als auch zwischen Juli und August in Eiern gefunden, was einen zweimaligen Gametenzyklus pro Jahr in dieser Art in Nord Sulawesi wahrscheinlich macht. Das vierte und letzte Kapitel befasst sich mit der Hypothese, dass Korallentransplantate, die einem elektrischen Feld ausgesetzt sind ein erhöhtes Wachstum aufweisen. Die hierbei untersuchte Technik basiert auf der Salzwasser-Elektrolyse, bei der es bei Anlegen einer niedrigen Spannung zwischen einer Anode und Kathode zu Kalkfällung durch erhöhte Ionen- Konzentrationen an letzterer kommt. Die Kathode wird hierbei als künstliches Riff-Substrat benutzt. Der mit der Zeit wachsende Kalkmantel entspricht in der Zusammensetzung dem vom natürlichen Korallenkalk und führt zu einem Wachsen der eigentlichen Kathodenstruktur, mit einer einhergehenden Stabilitätszunahme. Die Beschaffenheit dieses Kalkmantels macht ihn zu einem idealen Siedlungssubstrat für Larven mariner Organismen. Den gestalterischen Möglichkeiten der Kathodenstruktur sind in der Anwendung in der Riffrehabilitierung keine Grenzen gesetzt. Der von den Erfindern beobachtete Nebeneffekt eines erhöhten Wachstums von Korallen, die auf diesen Strukturen wachsen wurde bisher kaum wissenschaftlich untersucht und führte zu kontroversen Diskussionen unter Wissenschaftlern. Von einigen Autoren wird die erhöhte Ionen- Konzentration und dem daraus erhöhten pH-Wert in Kathoden-Nähe als mögliche Ursache für

VII erhöhtes Wachstum betrachtet. Fingerlange Fragmente der beiden Acropora-Arten A. yongei und A. pulchra wurden auf drei verschiedenen Konstruktionen zur Klärung unterschiedlicher Aspekte transplantiert. In einem ersten Experiment wurden zwei tunnelartige Strukturen mit unterschiedlichen elektrischen Spannungen betrieben, um festzustellen ob Unterschiede einen in Längenwachstum nachweisbaren Effekt haben. In einem zweiten Experiment wurden die Fragmente auf tischartige Gestelle transplantiert, die sich in der Höhe unterschieden. Hierbei sollte untersucht werden, ob es eine ideale Transplantations-Höhe gibt, die sich in Unterschieden im Wachstum bemerkbar macht. In einem dritten Experiment wurde der Unterschied zwischen Fragmenten, die in direktem Kontakt mit der Kathode stehen und solchen die sich lediglich in einem elektrischen Feld ohne direkten Kontakt zur Kathode stehen, untersucht. Für alle drei Experimente gab es Kontrollstrukturen in jeweils gleichem Design jedoch ohne Strom. Außerdem wurden Wachstumsraten von “normalen” Zweigen innerhalb der Spenderkolonien gemessen. Alle Ergebnisse zeigten in den verschiedenen Experimenten kein erhöhtes Wachstum für die verwendeten Transplantate: Ein direkter Vergleich mit den “ungeladenen” Kontroll-Strukturen zeigte keine bis negative signifikante Wachstumsunterschiede. Ein Vergleich der Wachstumsraten mit denen von verbleibenden Ästen in den Spenderkolonien zeigte hingegen signifikant niedrigere Wachstumsunterschiede. Beste Ergebnisse wurden unter geringster Spannung von 7 A (ca. 1.67 A m-2), bzw. in einem elektrischen Feld erzielt, jedoch nicht bei einem direkten Kontakt mit der Kathode oder einer Spannung über 1.67 A m-2. Allerdings zeigten die Ergebnisse auch unterschiedliche Performance zwischen den einzelnen Experimenten. Diese Ergebnisse unterscheiden sich erheblich von den Beobachtungen der Erfinder, die allgemein von einem fünf bis zwanzigfach erhöhtem Korallenwachstum sprechen. Erkenntnisse aus diesen drei Experimenten sollen dabei helfen, die Planung zukünftiger Projekte, und die Wahl der Korallenarten, Größe, Kathoden-Design und Stromstärke zu erleichtern.

VIII Chapter 1

Thesis Overview

1 2 1. General Introduction

1.1 Reproduction in scleractinian corals

Reproduction is the biological process that leads to the production of new organisms. It is a fundamental feature of all known life as every organism is the result of reproduction. Reproduction can be divided into two types: sexual or asexual. The term “sexual” can also be referred to as generative, while the term “asexual” as vegetative. The distinction between sexual and asexual reproduction is biologically insignificant in the phylum (order Scleractinia), as borders between these reproduction types are blurred: Cnidarians often can reproduce in both ways and show a great variety in either of the two (reviewed in Fautin 2002).

1.1.1 Asexual reproduction Corals display, due to the high plasticity of their tissue, a wide variation in asexual reproduction modes (Harrison and Wallace 1990). Asexual propagules can be produced e.g. via bailout (Sammarco 1982), development of anthocauli (Krupp et al. 1993), asexually produced planulae (Ayre and Resing 1986), or budding (Rosen and Taylor 1969). The latter is a main feature in the order Scleractinia and takes place either intratentacularly, when a parent polyp divides itself in two or more daughter polyps, or extratentacularly, when daughter polyps form on the side of the parent polyp. Sometimes, both forms are found (Veron 2000). The manner by which it reproduces influences the shape of the colony. The interconnection of the polyps with each other is a significant part of life history in colonial corals. In branching corals such as most Acroporidae, the term “coral growth” most often refers to “longitudal length increase”. This should not be confused with linear skeletal extension of individual axial polyps through enhanced calcification in the apical region of the branch (Goreau and Goreau 1959, Oliver 1984), as it rather is the increase in branch length through asexual reproduction of polyps and the growth of their associated skeleton. Thus, the axial polyp is the parent polyp that does both - growing in length and budding extratentacularly (Wainwright 1963, Gladfelter 1983). Although the term “reproduction” is generally only used when gametes are involved (e.g. Fadlallah 1985, Sier and Olive 1994), it is important to understand that most “coral growth”-literature does indeed describe the consequences of asexual reproduction, as it is not strictly referring to the extension of single polyps, but rather to the extension of colony dimensions caused by the increase in polyp number within a colony. Further forms of asexual reproduction are to be found in the partial displacement or fragmentation of coral colonies caused by natural physical impacts of heavy waves and moving objects (Highsmith 1982). These fragments have the ability to survive, reattach to a suitable substratum and establish new fecund clone colonies, so that the whole process contributes to the distribution of the species

3 and is seen as a further reproduction strategy. Both restoration and rehabilitation projects often take advantage of this evolutionary adaptation by employing transplantation of coral fragments as a tool, as will be discussed in later chapters.

1.1.2 Sexual reproduction The individual polyps of colonial or solitary corals can be further divided into hermaphrodite or gonochoric , which sexually reproduce either as broadcast spawners or brooders. While most coral species are hermaphrodites, gonochrorism - as found in Porites and Fungia - is more of an exception (Veron 2000). In spawning corals, male and female gametes are released simultaneously into the water column, where fertilization and further larval development takes place. In brooders, on the other hand, fertilization of eggs and growth of planula larvae takes place inside the polyps, and larvae are released when fully developed and ready to settle. Early coral reproduction literature saw the brooding mode as the most typical form of sexual reproduction in scleractinian corals (see review by Harrison and Wallace 1990), while nowadays - with increasing studies from diverse geographic regions - data reveal that many species are spawners (see review by Guest et al. 2005a). Generally, both modes are stable within a genus, species population or colony, though there are reported exceptions. Acroporidae provide one example of different reproductive behaviour within one genus: while members of the subgenus Acropora are spawners, species of the subgenus Isopora are brooders (e.g. Acropora brueggemanni, A. palifera). In Western Australia some colonies of Pocillopora damicornis were reproducing as brooders, while others in the same reef released eggs and sperm (Ward 1995). Exceptions within a species from geographically divided areas were reported for Pocillopora verrucosa, which was reported as a brooder for one part of the world (Stimson 1978), while in several others it was identified as a spawner (Fadlallah 1985, Sier and Olive 1994, Kruger and Schleyer 1998). However, as the brooding report for P. verrucosa is found in an earlier publication, some authors acknowledge the possibility of species misidentification as having lead to wrong conclusions.

1.1.3 Mass spawning The best-known but not yet completely understood phenomenon of annual cnidarian reproduction is the “mass spawning” first described from the Great Barrier Reef in Australia. Here, populations of widespread and abundant corals are spawning together on the same few lunar nights in October of each year (Harrison et al. 1984, Willis et al. 1985, Babcock et al. 1986). Since these early reports from the 1980s, additional observations followed from a wide geographical range of locations around the world (see Dai et al. 1992, Guest et al. 2002, Bastidas et al. 2005, Hatta 2005, Vize et al. 2005). However, the term “mass spawning” is often used in a context different from the original definition: First described as a “synchronous mass spawning” that involved 105 scleractinian corals resulting into extensive egg-sperm-slicks drifting at the surface, “mass spawning” nowadays is also used to describe the synchronous gamete release of only few

4 coral species or just multiple colonies at the same time (see Bastidas et al. 2005, Hatta 2005, Vize et al. 2005). The term mass spawning also appears in coral reproduction reports from the north coast of in Indonesia (see in Tomascik et al. 1997), without any further reports following.

1.1.4 Reproduction in high latitude vs equatorial reefs Variations in environmental parameters are commonly believed to influence the onset and timing of reproduction in corals. Each of these cues is acting on a different scale, such as water and solar insulation for the month or season (Shlesinger and Loya 1985, Penland et al. 2004), tidal amplitude for the day of the month and light for the time of night (Harrison et al. 1984). Towards the equator such parameters have a much narrower range than they have in high-latitude regions. The previously described mass spawning phenomenon of the Great Barrier Reef is seen to be directly linked to such fluctuations: Oliver et al. (1988) recorded the geographic extent of mass spawning from the subtropical (23.5°S) to the tropical Great Barrier Reef (9.5°S). Their further comparison of reproduction seasonality and synchrony among and within coral species, distributed from the southern Great Barrier Reef towards the low latitudes of Papua New Guinea, revealed a decrease in both, with no occurrence of any great extent in equatorial regions. It is therefore hypothesized that the breakdown of seasonal and synchronous coral reproduction in low-latitude New Guinea is the result of the low variability of environmental parameters. Further evidence for this hypothesis is seen in the absence of mass spawning events in the northern Red Sea, where environmental parameters are also relatively constant (Shlesinger and Loya 1985). However, mass spawning of corals has been observed in other regions with narrow environmental fluctuations such as the low-latitude Solomon Islands (Baird et al. 2001), and in an equatorial reef in Singapore (Guest et al. 2002). Although Guest et al. (2005b) agree that such fluctuations, while less varied, are still sufficient for the physiological processes involved in synchronised spawning, they also argue that mass spawning is more the result of a strong selective promoting reproduction success through synchronisation, than it is the result of the presence of strong fluctuations of environmental parameters. As even different species of corals need similar environmental conditions for a successful reproduction, synchronous reproduction is as likely to occur in equatorial assemblages as it is at higher latitudes. Guest et al. (2005b) further argued that no coastal environment is truly aseasonal, and therefore reproductive seasonality and some degree of multispecific spawning may occur even on equatorial reefs. Although the extended seasonal patterns of spawning recorded in equatorial Kenya (Mangubhai and Harrison 2008) could support this statement, the highly asynchronous reproductive patterns of broadcast-spawning corals in these reefs do not agree with the assertion of Guest et al. (2005b) that mass spawning is a characteristic of equatorial reefs. With only little detailed research and the sometimes contradictory findings for equatorial regions, compared to the many detailed reproduction studies from high-latitude reefs, it remains impossible to determine which of the previous study reflects a general latitudinal trend. More studies on

5 reproduction patterns from a wider range of equatorial regions are needed in order to help elucidate the extent of synchrony and seasonality in coral reproduction.

1.2 Reef restoration and reef rehabilitation

With the world’s coral reefs declining at a dramatic speed (Precht and Dodge 2002, Wilkinson 2004), the disciplines of conservation biology and restoration ecology are undergoing a remarkable growth (Young 2000). Beside the conventional passive measures of protecting, managing and conserving marine habitats, active interventions such as the restoration or rehabilitation of coral reefs are becoming more popular (Clark and Edwards 1995, Jaap 2000, Schuhmacher et al. 2000, Fox and Pet 2001, Nonaka et al. 2003, Yeemin et al. 2006).

1.2.1 Pros and cons Reef restoration and rehabilitation efforts are controversial topics that are extensively discussed. Considering the ability of coral reefs to naturally recover from disturbances, a careful consideration of all pros and cons of such manipulative human interventions is needed. The success of restoration and the time needed for total reef recover depends on the survival of coral fragments, as well as the settlement of coral larvae (e.g. Baird and Hughes 1997, Bowden- Kerby 2003). To successfully settle, coral larvae need a suitable substrate. Where reefs or reef areas are insufficiently supplied with new recruits due to a lack of suitable settlement substrate, this self-recovery fails. Examples where self-recovery was found to be unlikely to impossible are reefs damaged by blast or dynamite fishing, reef flat dredging or coral harvesting – activities resulting in highly unstable rubble fields (Lindahl 1998, Fox et al. 1999, Fox et al. 2003). In such cases, human intervention in the form of active restoration and rehabilitation can be an effective measure (Bowden-Kerby 2003). Modern approaches as proposed by van Treeck and Schuhmacher (1998) are combining conservation and rehabilitation measures with their ideas of underwater parks. To the authors, so-called DAD (Diver Aggregation Device) have two advantages: a) the rehabilitation of one area through the deployment of artificial DADs, resulting in “natural” reefs, and b) the conservation of other areas as a result of the release from pressure caused by extensive dive tourism. Although reef restoration and rehabilitation efforts have different ecological approaches, they have two things in common: They are always labor intensive and often cost intensive (Fox et al. 2005). Furthermore, they are always small-scale projects, causing critics to question the high costs

6 for the comparably small benefits produced. For example, in Indonesia alone, there are 7.5 * 10

7 ha of coral reefs (Cesar 1996). Globally, there are 2.55 * 10 ha (Spalding and Grenfell 1997), which is in stark contrast to the largest reef rehabilitation project to date with 7.1 ha in Costa Rica (Guzman 1991; reviewed in Edwards and Clark 1998).

6 The calculated costs per square meter when using coral transplantation range from $1200 - $11000 -2 -2 m in the USA (Rinkevich 2005) to $50 - $73 US m for transplantation in Thailand (Yeemin et al. 2006). Beside the economic disadvantages, other authors also see unpredictable ecological consequences: Restoration or rehabilitation methods involving species manipulation and/or coral transplantation can influence the basic ecology of a partially intact reef system. Further, fast growing alien coral species can out-compete local slow growing corals (Bowden-Kerby 2003). Most often these fast growing species are branching corals that almost directly increase the fish abundance (Ferse 2008). However, their fragile structure can also lead to an increase in rubble production when damaged, thus resulting in further destabilization of the surrounding substrate. Nonetheless, as reef restoration methods have already been established in many areas, it is noteworthy that the methods used maximize success in fulfilling their – economic – goals (Spieler et al. 2001).

1.2.2 What is “restoration” and what is “rehabilitation” then? The terms “conservation”, “restoration” and “rehabilitation” often lead to confusion and misunderstanding and are often used in the wrong context. Therefore the following explanation should give a short overview for better understanding: “Conservation” is defined as the preservation and protection of natural habitat and biodiversity. The establishment e.g. of an MPA () is one of these measures in which a management of the resources by limiting human activities takes place, rather than a direct intervention and manipulation of a habitat. “Restoration” is a human intervention to accelerate the recovery of a natural habitat, or to bring back as closely as possible to their pre-disturbance states (Yap 2000). It is often seen as a manipulative tool of conservation efforts. “Rehabilitation” on the other hand is the act of partial or full replacement or substitution of alternative qualities of structural or functional characteristic of an (Edwards 1998). Most of today’s “reef restoration projects” are using artificial reef elements and would thus, according to the previous definition, constitute “rehabilitation” efforts rather than conservation or restoration of a habitat.

1.2.3 Old and new approaches In the past, materials such as bottles, Pedi cabs (e.g. Indonesia), ships (e.g. USA) and car tires (e.g. Philippines) were immersed as artificial reefs for restoration efforts (reviewed in Waltemath and Schirm 1995, Pickering et al. 1998). These materials turned out to be unsuitable in many cases due to their instability and leaching of toxic substances and, in some cases, very limited settlement of marine sessile organisms. In many cases the dumping of such materials was nothing more than an endeavor to justify the dumping of industrial waste in coastal areas. Nowadays there are three major “techniques” dominating the attention of the media and public

7 that find application in many rehabilitation projects: 3 “Reefballs” (Barber and Barber 1994) are spherical concrete structures of approximately 1m volume made out of a pH-neutral, environment friendly cement mix. With their hollow structure and diverse holes they can provide shelter for marine organisms. Coral fragments can be attached by underwater epoxy and the rough surface of the spheres provides a suitable settlement substrate for the larvae of sessile organisms. The main disadvantages are the high and the number of units one would need to rehabilitate bigger areas. “Ecoreefs” are snowflake-shaped porcelain units 0.5 m in diameter with the main aim of providing a maximum surface area for coral settlement and of stabilizing rubble. Unfortunately the main components have to be shipped from the place of production to the area where they will be used. When the components are assembled, transplants are easy to attach on the numerous arms with cable ties (Razak 2006). While the previous conventional approaches aimed at the provision of a suitable transplant basis and the stabilization of substrate without further manipulation, the “mineral accretion technology” claims to have both a suitable substrate and an active support of transplant performance on a technical basis (Hilbertz and Goreau 1996). This approach utilizes the principle of seawater electrolyses to build up a calcium carbonate-coating around the core material. A metal structure submerged in seawater works as a cathode and is the actual substrate basis, while a Titanium mesh deposited close-by functions as the anode. When a direct current (DC) is established between the two electrodes, CaCO3 and magnesium hydroxide [Mg(OH)2] will be deposited on the cathode (Hilbertz 1992). Depending on the induced current, the ratio between the crystalline form of

CaCO3 (Aragonite) and Mg(OH)2 (Brucite) is shifted. Less current tends to result in the more stable form Aragonite (Hilbertz 1992). This substrate is similar to natural reef limestone in chemical and physical appearance (van Treeck and Schuhmacher 1997). The possibilities in cathode design are limitless. Its shape and form can be customized for special needs, as long as the material used is conductive. The transportation of core structures normally doesn’t need more than a common dive operator boat. However, the use of cable can be extensive if the power source is located on land. On the other hand, the use of other, better power supply (e.g. sea-based solar panels) is even more technical and cost intensive. Furthermore the maintenance of cables, connections and power supply can be intensive. Previous observations indicated that corals transplanted to these so-called “BioRocks” are growing faster than their natural mother colonies. The generated electric field is believed to cause this positive side effect by providing CaCO3 enriched water that enhances natural calcification in corals, and by providing extra electrons from the electrochemical processes for ATP production, thereby enhancing metabolic efficiency in terms of coral growth and fecundity (Hilbertz and Goreau 1996). Until today, only one peer-reviewed article has been published in which the authors were able to measure a significant increase in girth growth of the transplanted corals, nubbins of Porites cylindrica (Sabater and Yap 2002).

8 1.2.4 Using coral fragments When using coral transplants, the primary aim is to gain an immediate increase in coral cover, diversity (Edwards and Clark 1998), and fish abundance (Clark and Edwards 1999). Furthermore the presence of coral fragments is believed to attract coral larvae, and with that accelerate the natural colonization by corals (Gittings et al. 1988). Coral transplantation was used as a restoration tool for the first time in the early 1970s (Maragos 1974). Since then, coral transplantation has found a wide range of applications in restoration and rehabilitation projects dealing with reefs that were damaged by ship groundings (Gittings et al. 1988, Jaap 2000, Schuhmacher et al. 2000), dynamite fishing (Auberson 1982, Bowden-Kerby 2003), Acanthaster plancii-outbreaks (Harriott and Fisk 1988) and tourism (Rinkevich 1995, van Treeck and Schuhmacher 1999). When using coral fragments, caution must be taken in many ways: The harvest of coral fragments should be as non-destructive and sustainable for the donor reef and colonies as possible. Collected fragments have to be handled with great care, as the tissue is sensitive. Their size and the availability of transplantation substratum are limiting factors for a positive performance: coral fragments are able to colonize areas poorly suited for larvae colonization, such as for example sand bottoms (Highsmith 1982, Heyward and Collins 1985). Their size and the transplantation technique have to be chosen carefully. Survival rates of unsecured fragments can be increased when using larger fragments of more than 30 cm, while small sized fragments display higher mortalities (Bowden- Kerby 1997). When small sized fragments are firmly attached to e.g. concrete nails driven in the substrate, the survival rates can increase (Okubo et al. 2005). When choosing suitable corals for transplantation, the life-history strategy of species should be considered: survival rates of coral fragments with a k-mode (slow growth, as e.g. found in massive corals) strategy were observed to be superior to those of a r-mode (rapid growth) species (Yap et al. 1992). A comparison of the survival rates of Acropora species with different growth forms showed that the staghorn type had the best survival rates compared to bushy and tabular types (Smith and Hughes 1999). Those fragments with multiple tips reached twice the total length compared to single-tip fragments in the same time (Rinkevich 2000), and non-spawning fragments of Acropora formosa had a four times increased growth rate compared to those of bigger, spawning fragments (Okubo et al. 2007). However, even closely related species that are both dominant in the same reef area respond differently to transplantation as well as to their new environment (Yap and Molina 2003).

1.3 Aim of the study

The aim of this thesis is to address the lack of information on peak seasons of sexual reproduction and reproduction cycles from equatorial Indonesian reefs. These data will help to shed more light on the question whether synchronous mass spawning is also characteristically for equatorial reefs,

9 or if such observations are exceptions and cannot be generalized. To fulfil this goal, recruitment studies and histological techniques of coral tissue examination have been used. The study further attempts to test if the asexual reproduction, in terms of longitudinal length increase of coral fragments, can be positively manipulated by the use of an electrical field. Such reef rehabilitation measure would be helpful in areas where natural recruitment is lacking by increasing the coral cover – both instantly and in the long term - that then could function as an attraction for and a new source of natural recruits. Furthermore, if the corals are able to benefit from an electrical current, as manifest in increased growth rates, it is assumed that they are able to allocate more resources towards sexual reproduction. Objectives of this study are to: a) clarify if seasonal recruitment patterns in close proximity to the equator in North Sulawesi occur b) identify the coral families that are involved in a) c) find evidence if mass spawning is part of these patterns d) reveal the number of gametogenic cycles in selected coral species e) test the hypothesis that transplanted coral fragments show increased growth rates and higher survival due to stimulation in the vicinity of an electrical field f) clarify whether a positive response by fragments varies among transplanted species

2. General Material & Methods

2.1 Study Area

South-East Asia encompasses approximately 30% of the world’s coral reefs (Chou 1997). The Indo-Pacific between the Indonesian and French Polynesia is the global center of marine diversity for several major taxa such as corals, reef fish and crustaceans (Bryant et al. 1998, Bruno and Selig 2007). Its reefs had an average coral cover of approximately 42% in the early 1980s that declined region-wide to 22% by 2003, though exceptional reefs with more than 90% coral cover still can be found (Bruno and Selig 2007). In 2003 the so-called Coral Triangle (Fig. 1) was outlined as a first step in a marine ecoregional conservation assessment. The boundaries of this area are primarily based on high coral biodiversity in countries harboring more than 500 coral species, namely the five countries of Indonesia, East Timor, Philippines, Malaysia (Sabah), and Papua New Guinea (Green and Mous 2004). Indonesia is the largest archipelago in the world, with approximated 17,500 islands and 54,700 km of coastline stretching out from 6°N to 11°S. A deep-water runs between Bali and Lombok in the south and Borneo and Sulawesi in

10 the north and separates the country into two biological zones by a sharp biogeographic break known as the Wallace Line. It was first thought to separate only terrestrial forms from eastern and western Indonesia, but new genetic studies of crustaceans give evidence for a marine Wallace’s Line too (Barber et al. 2000). Sulawesi, east of the Wallace line, is located right in the center of this Coral Triangle. Its northern tip, the southern Philippines and Borneo’s north-eastern coast are framing the Celebes Sea – part of an ancient ocean basin and now hotspot in an area of high marine biodiversity. The sites studied here are all located at the far northern tip of Sulawesi in and around the Bunaken Marine National Park (Fig. 2). The Park includes reefs of the main land in close proximity of the provincial capitol Manado, and the islands Bunaken, Manado Tua, Siladen, Nain and Mantehage. The absence of a allows the coastal area of the park to drop directly down to the continental slope. The depth between the islands of the park is from 200 m to 1840 m (Mehta 1999).

2.2 Settlement tiles & settlement frames

For the recruitment study, untreated and uncoated natural limestone tiles were used. These tiles, locally called “batu alam”, were easy to find in local hardware shops. Tiles were custom-cut in

Fig. 1 Map of South-East Asia showing the boundaries of the Coral Triangle with North Sulawesi in it middle. SOURCE: Coral Geographic (Veron et al. unpublished data)

11 the dimensions 15 x 15 x 1 cm. Tiles were recycled after examination for recruitment by sanding them with a grinder. All calcareous structures, like barnacles, oysters, Canalipalpata and corals’ skeletons could be efficiently removed without leaving remnants behind. The design of the frames (see Fig. 2 in Chapter 2) allowed for an alignment with one side directly facing the water current. Furthermore, the frame design and the alignment of tiles in a 60° angle assured that one side of the tile sets always lay in the slipstream of the current. This was done to minimize any potential effect from changing current direction and intensity on recruitment rates.

2.3 In situ monitoring of reproduction status

On every sampling date, branches of Acropora colonies were broken off below an expected sterile zone to check for their reproductive state as described in Wallace (1985a). This is an easy way to conduct monitoring for branching corals with bigger polyps. Mature eggs, planulae or sperm- egg bundles are usually red, pink or orange and easy visible to the naked eye due to their size in the skeletal structure (Fig. 3). The presence of pigmented eggs is an indicator for an upcoming

Fig. 2 Map of North Sulawesi showing the study area and the sampling sites: Meras close to the city of Manado, Raymond’s Point and Fukui on Bunaken and Lihage - a small island south-west of Gangga. The location within Sulawesi is indicated on the map in the inset.

12 Fig. 3 Broken off Acropora-fragment with pink coloured egg in the upper right corner of the fracture

spawning event on or shortly after the subsequent full moon (Harrison et al. 1984, Baird et al. 2002, Guest et al. 2005a).

2.4 Histological preparation for examination of reproduction status

Pieces of coral colonies were broken off with tongs or a hammer and labelled for later identification. They were then immediately transferred into jugs filled with a 10 % Formalin-seawater and stored for at least one week for proper fixation. Skeletal parts were decalcified in 11 % formic acid for up to three days, changing the solution as necessary. In many cases a total decalcification of the skeleton of the more massive Pocillopora verrucosa branches was not necessary, as tissue could be easily peeled off from remaining skeletal fragments. The tissue was then rinsed with running tap water and stored in a 70 % Ethanol solution. For the production of histological preparates approximately 1 cm2-quadrates of tissue were cut from the stored tissue samples with scissors. They were run through a series of alcohols (80 % for 1 hour, 90 % for 1 hour, 95 % for 1 hour, 96 % for hour) for dehydration. They were then transferred to xylene-creosote (1 hour), xylene (1 hour), paraffin-xylene (1 hour) and paraffin (at least 1 hour) before finally being embedded in paraffin blocks (Tioho et al. 2001). Cutting series of 6μm thickness were made through centrally located polyps and mounted on objective slides. Slides were washed in a xylene-substitute for 10 min to solve the paraffin.

13 To evaporate the xylene, the slides were dried with a hair dryer before hydrated in a series of alcohol (95 % for 3 min, 70 % for 3 min, H2O for 3 min). After drying, they were colored in - hematoxylen for 4 min, transferred to H2O for 4 min, rinsed in clean H2O and dried. After 2 min of being immersed in eosin-phloxine and being washed in 96 %-Alcohol, they were put into xylene for 5 min before cover glasses were finally mounted with Shandon E-Z-Mount.

2.5 Surface estimation and zooxanthellae-count of coral fragments

The tissue of coral branches was removed with an airbrush gun and seawater in half-light. Tissue parts were finely ground and then seawater was added to a total volume of 40 ml. From this solution, 5 ml samples were removed and preserved with 1 ml formalin (40 %) for the determination of zooxanthellae density. Zooxanthellae from these sub samples were counted in a Neubauer counting chamber, following standard procedures. The remaining 35 ml of the solution were stored in dark plastic containers and re-frozen for later chlorophyll extraction following the description of Gardella and Edmunds (1999). Chlorophyll-a and -c were calculated according to the equation of Jeffrey and Humphrey (1975). The blank coral skeletons were kept for surface area estimations. For this, the blank skeletal fragments were coated in varnish to close all pores, weighted, and recoated in paraffin and weighted a final time, utilizing the technique of Stimson and Kinzie III (1991). A calibrating diagram was established using objects with a known surface area and treated in the same manner as the coral skeletons above. The corals’ surface area was then calculated using the diagram. From each of the preserved sample solutions, chlorophyll was extracted from sub-samples, as described in Gardella and Edmunds (1999), and measured in a spectrophotometer for chlorophyll-a and -c concentrations.

3. General Results & Discussion

3.1 Reproduction and settlement in equatorial reefs: continuous with seasonal peaks

Settlement of coral spat is an important contribution for the growth and maintenance of coral reefs and the life history of corals themselves. High numbers of coral recruits were found on settlement tiles deployed in four reef sites in North Sulawesi (paper I). Recruits were found during every sampling interval in each site. Even if high numbers of coral recruits settle in a reef, their mortality rate is usually high: Grazing fish (Fitzhardinge 1988), sedimentation (Babcock and Mundy 1996), space competition with other fouling organism (Dunstan and Johnson 1998, Edmunds and Carpenter 2001, Schmidt 2007) or simple settlement places with unsuitable light

14 conditions (Birkeland and Randall 1981, Maida et al. 1994) are limiting the recruitment success. Those recruits that remain and finally grow to healthy, extensive coral colonies will assure the survival of species and reef systems. While coral settlement on the tiles in the equatorial North Sulawesi showed no clear seasonality, the settlement in high latitude tropical and subtropical reefs has been found to be seasonal with months of no, or very limited, recruitment (Banks and Harriott 1996, Glassom et al. 2004, Minton and Lundgren 2006). Changing environmental parameters, in particular the colder water temperature, cause the lack of recruitment during the winter season in those reefs. These parameters are in direct relation with the reproduction time in corals. In most subtropical reefs the coral reproduction season falls into a warmer month of the year (Wallace and Bull 1982, Wallace 1985b). The reefs of North Sulawesi, with their close distance to the equator, have a yearly average temperature of ca. 29.0 °C during the dry season and 28.0 °C during the wet season (paper I). Though did drop to 22.5 °C and rise to 30.1°C, these temperature extremes were only short episodes (between two measurement intervals of 30 min) caused by interaction of strong horizontal and vertical currents and the bathymetry of the area. As water temperature in the tropical areas can be seen as relatively stable with minor fluctuations between seasons, it is not surprising that no clear seasonal patterns could be detected in total coral recruitment on settlement tiles. There were seasonal differences, however, between the most common families of Acroporidae and Pocilloporidae. Pocilloporid recruits dominated tiles throughout the year, their total number being  higher than those of acroporid corals (paper I). However, there were site differences in species composition on tiles and distinct peaks in acroporid settlement showing a clear seasonality for this family. These peaks were at variance between subsequent years and between sites as also found on other reefs (see Wallace 1985b). Though temperature might play a triggering role in the development of reproduction tissue, it can be overlooked for the examined area due to the recorded minimal temperature variations. The more detailed histological examination of coral tissue from selected species in paper II showed a single annual reproduction cycle for the broadcasting species Acropora yongei and A. pulchra. By limiting and synchronizing the number of gamete cycles, broadcast spawning, as found in most Acroporidae, may maximize fertilization success and saturate planktonic predators, so that a high proportion of the eggs survive (Oliver and Babcock 1992). The lack of seasonality found in pocilloporid settlement in the examined reefs can be explained by the dominant reproduction strategy in this family. Most of its members reproduce as larvae-releasing brooders that are able to release larvae throughout the year. Furthermore, Pocilloporidae were one of the dominating families in the four stations (unpublished data), with their brooded larvae most likely to settle in close distance from their mother colonies on settlement tiles nearby, as demonstrated by Tioho et al. (2001). One of such brooding pocilloporid is Seriatopora hystrix (paper II), a common species found in all examined sites. Ripe ova and spermaries were present in every examined sample in this species. Often neighbouring polyps contained well-developed larvae at the same time, verifying S. hystrix as a continuous brooder (paper II) and indicating that larvae of this species

15 were part of the pocilloporid recruits found year round on the settlement tiles (paper I). The other examined pocilloporid species P. verrucosa has a reproduction mode less common for this family, as it proved to be a spawner in this study. Histological examination of specimens collected in the different sites revealed an extended time of the year where ripe, zooxanthellae-containing eggs were present. This data also showed that the North Sulawesi population has asynchronous reproduction patterns and that at least some colonies were containing ripe eggs twice a year. The relationship between the timing of coral spawning and recruit settlement has been shown in a number of studies (Wallace 1985a, Hughes et al. 1999, Glassom et al. 2004, Mangubhai et al. 2007). The peaks in acroporid settlement observed in early April and early June 2006 (paper I) in Lihage concurred with histological findings and field observations of ripe eggs in A. pulchra in April and A. yongei in May (paper II). The time from gamete release until settlement competency can range from as short as 2.5 days (Miller and Mundy 2003) to a maximum competency time of 78 days (Richmond 1988), depending on the spawning species and the water temperature (Nozawa and Harrison 2000). This time range falls well into the time between histological findings of ripe eggs and increase in recruit number on settlement tiles (paper I & II). Furthermore, the increase in recruit number coincided with small temperature increases noticed by the temperature loggers tied to the settlement frames (paper I). Experiments with free-swimming larvae showed that sudden increases of temperature have a signalling effect (Coles 1985, Nozawa and Harrison 2007). Even if the results of paper I and II were able to show a lack of clear seasonality in coral recruitment and an asynchronous reproduction in some of the examined species – they were not able to rule out the existence of mass spawning based on synchrony in this equatorial region. Some of the recruitment peaks in certain weeks of the year are in fact most likely the result of recruitment from a high number of species. However, uncertainties in coral recruit identification on the species level remain as today’s techniques do not allow such a precise classification.

3.2 Manipulation to enhance growth of coral fragments and survivorship

While sexual reproduction in single colonies takes place only during certain times of the year (paper I & II), the asexual reproduction in the form of budding - better known as longitudinal growth enhancement (from hereon called “growth”) – is a continuous process without any seasonal breaks in between. It generally slows down when corals are using more energy for the development of gametes (Okubo et al. 2005) or when exposed to physical stress, as for example fragmentation. The asexual reproduction in the form of fragmentation is a natural process that finds application in the form of coral transplantation in reef rehabilitation and restoration projects and was tested in paper III. Its results demonstrated that the mineral accretion technique supports a strong and secure hold of transplanted corals. Lose attachment is known to reduce coral growth and is often the reason for loss of fragments. On the electric reef, a firm connection between fragments and substrate developed within a few days. Those fragments tied to the control structures needed several

16 weeks before firm attachment or, alternatively, never reached the same strength of attachment. The present work demonstrated survival rates of up to 100% for treated fragments, depending on type of experiment and species. The lowest survival (19%) was measured on a control structure, where fixation of the fragments was only established by cable ties. Fragments never found the same extreme hold as they did on the cathode structures. An additional affecting factor might have been the placement of the transplantation board directly on sandy bottom, with sedimentation causing extra stress (Crabbe and Smith 2005). Survival of fragments always is a critical factor for the success of rehabilitation and restoration projects involving coral transplantation (Clark and Edwards 1995). But even if survival is high, sufficient transplant growth is not guaranteed, as the monitoring of rehabilitation projects has shown (e.g., see Yap and Gomez 1985, Custodio III and Yap 1997). The main reasons for reduced growth rates in transplants are increased stress through handling of corals, tissue damage, size of fragments, species, and use of unsuitable techniques. Acropora yongei and A. pulchra exhibited very low growth rates in all three experiments, which can be interpreted as increased transplantation stress. Fragments first must expend energy towards the healing of tissue narcosis around the fracture, which often accompanies development of new polyps before the energy gets redirected towards longitudinal growth. Higher amperage demonstrated to be a stress factor as well, as those fragments that were treated with lower amperage had better growth rates. This is likely due to increased amounts of brucite, a rather soft form of mineral accretion with high alkalinity, which has been shown to negatively affect many organisms (reviewed in Eisinger et al. 2005). Furthermore, total zooxanthellae numbers in A. yongei only amounted to half the densities found in donor colonies, but were still significantly higher than those in the controls. A. pulchra showed neither significant differences between treated and untreated fragments nor compared to the donor. A reduction of zooxanthellae densities is a good indicator of physiological stress in response to anomalous environmental conditions (Jones 1997a, Jones 1997b). However, these results are in strong contrast to the enhanced growth and higher number of zooxanthellae previously observed in the tissue of fragments transplanted onto electric substrates (Hilbertz and Goreau 1996, Goreau et al. 2004, Goreau and Hilbertz 2005).

4. Concluding remarks

When converting the highest number of recruits found during the final sampling period in Meras into recruits per m2, on average, 232 recruits m-2 were observed. Ferse (2008) conducted a concomitant study in the same sampling site but in approximately 16 m depth, with a sampling interval of three month, and counted 363 recruits m-2. Though data can not be compared one to one due to the different sampling interval, it still demonstrates that recruitment in depth might be higher than in shallower depths, where the present work was carried out. A similar increase of recruitment to a depth of 20 m has also been observed by Smith (1997), who compared settlement

17 tiles deployed at different depths. However, the number of recruits in the present study on tiles deployed for a duration of one year was much smaller than those on two- and three-month tiles, indicating that a longer deployment in the water does not automatically lead to a higher number of recruits. These findings are also supported by the results of Schmidt-Roach (2008), who compared tiles deployed for different amounts of time from Meras attached in approximately 8 m depth and found an increase during the first weeks but a decline in recruitment after two month of dispersal. Despite the differences of sampling depth and the coarser sampling interval of Ferse (2008), the concomitant works noticed analogue increases of recruitment during certain times of the year: The present study with its fine scale settlement patterns noticed an increase in the last week of June and November 2006 as well as in the last week of May 2007 in Meras, which also could be detected on the three-month tiles collected in June and December 2006 and May 2007 by Ferse (2008) from bigger depth in Meras. Similar patterns were also noticeable for the neighbouring island Lihage, with recruitments peaks in June and August 2006 (present study), and the July/ August peak on Gangga (Ferse 2008). The use of untreated limestone tiles proved to be an excellent settlement substrate for recruitment studies. In combination with growth keys gained from the fluorescence census technique for the early detection of coral recruits, it was further possible to narrow the time of settlement down to one month with a certain degree of probability. A histological work as conducted in paper II with the preparation and following examination of coral tissue is very work-intensive, so that the number of replicates has to be kept small. Furthermore the chances of missing the ideal cross-sectioning of reproductive tissue are high, with a degree of uncertainty regarding the reproductive status remaining. The in situ inspection of fracture sites from Acropora branches (Wallace 1985a) seems to be much more suitable and time saving, if not even more reliable for species with big-sized polyps when only focusing on the development status of ova. However, the predicted settlement times in combination with histological examination and in situ observations of spawning events helped to gain a detailed inside view of sexual reproduction patterns for these Indonesian reefs. Though the results of the mineral accretion technique documented in paper III were in some points disappointing regarding the previously asserted “multiple-times enhancement of coral growth”, it still fulfilled the main requirements for a successful rehabilitation project when the “right” settings were utilized. The fixation of the used fragments was excellent and their growth was only slightly different from the controls not in direct contact to the cathode, and thus only exposed to a weak electrical current. A likewise pattern was found in the similar high survival rates. It is clear that this is not a “wonder tool” that will be able to save coral reefs worldwide, as in the end, transplanted fragments are still depending on suitable environmental conditions. While the developers also promote a higher stress and bleaching resistance, the results presented in paper III were not able to support this hypothesis. Furthermore, big scale projects find their limitations, as the technique remains an energy-consuming technical solution with the need of quite intensive maintenance, as cables, connections and power supply need continuous attention. It is also limited in its use in

18 remote areas where no power supply is available. Although the developers see the possibilities to power such installations by solar panels, extra safety and vigilance would be needed in this case, thus creating extras costs. However, due to the ease of installation and the low cost, the most expensive components of the project being the cables (ca 70 US $ per 100 m), battery charger (ca 60 US $) and anode material (ca 100 US $ per m2), it is an excellent technique for small-scale projects. It proved to be especially suitable for dive resorts aiming revitalize their house reefs that were damaged in one way or another and were thus not interesting enough for diving guests (personal experience). Furthermore, with increasing environmental awareness, such resorts often try to establish underwater educational paths by connecting diverse artificial structures to increase guests’ awareness. Concerning the assumed “enhanced” growth of transplanted corals on electrical structures, it is possible that observed changes in growth and expansion of colonies are more likely to be noticed on artificial structures than in the surrounding natural reef, as such structures give geometrical reference points. The fast occupancy of free space on such structures by corals also might be a result of normal space competition and better light regimes for all branches due to fewer competitors. Overall, the electro-chemical of minerals on conductive materials seems to have a higher positive effect on stability of both the structures and the transplanted corals, than having positive physiological effects on the corals.

5. References

Auberson B (1982) Coral transplantation: an approach to the reestablishment of damaged reefs. Kalikasan, Philippines Journal of Biology 11: 158-172. Ayre DJ, Resing JM (1986) Sexual and asexual production of planulae in reef corals. Mar Biol 90: 187-190. Babcock R, Mundy C (1996) Coral recruitment: Consequences of settlement choice for early growth and survivorship in two scleractinians. J Exp Mar Biol Ecol 206: 179-201. Babcock RC, Bull GD, Harrison PL, Heyward AJ, Oliver JK, Wallace CC, Willis BL (1986) Synchronous spawnings of 105 scleractinian coral species on the Great Barrier Reef. Mar Biol 90: 379-394. Baird AH, Bellwood DR, Connell JH (2002) Coral Reef Biodiversity and Conservation. Science 296: 1026. Baird AH, Sadler C, Pitt M (2001) Sychronous spawning of Acropora in the Solomon Islands. Coral Reefs 19: 286 Baird AH, Hughes TP (1997) Spatial variation in coral recruitment around Lizard Island, Australia.

19 Proc 8th Int Coral Reef Symp 2: 1207-1210. Banks SA, Harriott VJ (1996) Patterns of coral recruitment at the Gneering , southeast Queensland, Australia. Coral Reefs 15: 225-230. Barber PH, Palumbi SR, Erdmann MV, Moosa MK (2000) A marine Wallace’s line? Nature 406: 692-693. Barber TR, Barber GL 1994. Reef Ball. http://www.uspto.gov, US Patent no. 5,564,369. Bastidas C, Croquer A, Zubillaga AL, Ramos R, Kortnik V, Weinberger C, Marquez LM (2005) Coral mass- and split-spawning at a coastal and an offshore Venezuelan reef, southern Caribbean. Hydrobilogia 541: 101-106. Birkeland C, Randall RH (1981) Facilitation of coral recruitment by Echinoid excavations. Proc 4th Int Coral Reef Symp 1: 317-322. Bowden-Kerby A (2003) Coral transplantation and restocking to accelerate the recovery of coral reef habitats and fisheries resources within no-take Marine Protected Areas: Hands-on approaches to support community-based coral reef management. International Tropical Marine Ecosystems Management Symposium 2: 1-15. Bowden-Kerby A (1997) Coral transplantation in sheltered habitats using unattached fragments and cultured colonies. Proc 8th Int Coral Reef Symp 2: 2063-2068. Bruno JF, Selig ER (2007) Regional decline of coral cover in the Indo-Pacific: Timing, extent, and subregional comparison. Plus One 2: e711. Bryant D, Burk L, McManus J, Spalding M (1998) Reefs at risk. World Resources Institute p 60. Cesar H (1996) Economic analysis of Indonesian coral reefs. In: Department E. (ed.) Work in Progress: Toward environmentally and socially sustainable development. World pp 97. Chou LM (1997) The status of Southeast Asian coral reefs. Proc 8th Int Coral Reef Symp 1: 317- 322. Clark S, Edwards AJ (1995) Coral transplantation as an aid to reef rehabilitation: evaluation of a case study in the Maldive Islands. Coral Reefs 14: 201-213. Coles SL (1985) The effects of elevated temperature on reef coral planula settlement as related to power station entrainment. Proc 5th Int Coral Reef Symp 4: 171-176. Crabbe MJC, Smith DJ (2005) impacts on growth rates of Acropora and Porites corals from fringing reefs of Sulawesi, Indonesia. Coral Reefs 24: 437-441. Custodio III HM, Yap HT (1997) Skeletal extension rates of Porites cylindrica and Porites (Synarea) rus after transplantation to two depth. Coral Reefs 1997: 267-268. Dai CF, Soog K, Fan TY (1992) Sexual reproduction of corals in Northern and Southern Taiwan. Proc 7th Int Coral Reef Symp 1: 448-455. Dunstan PK, Johnson CR (1998) Spatio-temporal variation in coral recruitment at different scales on Heron Reef, southern Great Barrier Reef. Coral Reefs 17: 71-81. Edmunds PJ, Carpenter RC (2001) Recovery of Diadema antillarum reduces macroalgae cover on

20 a Caribbean reef. PNAS 98: 5067-5071. Edwards AJ (1998) Rehabilitation of coastal ecosystems. Mar Pol Bul 37: 371-372. Edwards AJ, Clark S (1998) Coral transplantation: a useful management tool or misguided meddling? Mar Pol Bul 37: 474-487. Eisinger M, van Treeck P, Paster M, Schuhmacher H (2005) Künstliche Riffe für Korallen: Auf diese Gitter können sie bauen. Biol Unserer Zeit 35: 192-198. Fadlallah YH (1985) Reproduction in the corl Pocillopora verrucosa on the reefs adjacent to the industrial city of Yanbu (Red Sea, Saudi Arabia). Proc 5th Int Coral Reef Symp 4: 313- 318. Fautin DG (2002) Reproduction of Cnidaria. Canadian Journal Of Zoology-Revue Canadienne De Zoologie 80: 1735-1754. Ferse SCA (2008) Artificial reef structures and coral transplantation fish community responses and effects on coral recruitment in North Sulawesi/Indonesia.PhD-Thesis. University of Bremen, Bremen, pp 189. Fitzhardinge RC (1988) Coral recruitment: the importance of interspecific differences in juvenile growth and mortality. Proc 6th Int Coral Reef Symp 2: 673-678. Fox HE, Mous PJ, Pet JS, Muljadi AH, Caldwell RL (2005) Experimental assesment of coral reef rehabilitation following blast fishing. Conservation Biology 19: 98-107. Fox HE, Pet JS, Dahuri R, Caldwell RL (2003) Recovery in rubble fields: long-term impacts of blast fishing. Mar Biol Bull 46:1024-1031. Fox HE, Pet JS (2001) Pilot study suggests viable options for reef restoration in National Park. Coral Reefs 20: 219-220. Gardella DJ, Edmunds PJ (1999) The microenvironment adjacent to the tissue of the scleractinian Dichocoenia stokesii and its effects on symbiont . Mar Biol 135: 289-295. Gittings SR, Bright TJ, Choi A, Barnett RR (1988) The recovery process in a mechanically dmaged coral reef community: recruitment and growth. Proc 2nd Int Coral Reef Symp: 225-230. Gladfelter EH (1983) Skeletal developmetn in Acropora cervicornis II. Diel patterns of calcium carbonate accretion. Coral Reefs 2: 91-100. Glassom D, Zakai D, Chadwick-Furman NE (2004) Coral recruitment: a spatio-temporal analysis along the coastline of Eilat, northern Red Sea. Marine Biology 144: 641-651. Goreau TF, Goreau NI (1959) The physiology of skeletal formation in corals II. Calcium deposition by hermatypic corals under various conditions in the reef. Biological Bulletin 117: 239- 250. Goreau TJ, Cervino JM, Pollina R (2004) Increased Zooxanthellae numbers and mitotic index in electrically stimulated corals. Symbiosis 37: 107-120. Goreau TJ, Hilbertz WH (2005) Marine ecosystem restoration: Costs and benefits for coral reefs. World Res Rev 17: 375-409.

21 Green A, Mous PJ (2004) Delineating the Corl Triangle, its ecoregions and functional seascape. The Nature Conservancy, Southeast Asia Center for Marine Protected Areas, Sanur, Bali, pp 26. Guest J, Baird AH, Goh BPL, Chou LM (2005a) Reproductive seasonality in an equatorial assemblage of scleractinian corals. Coral Reefs 24: 112-116. Guest J, Baird AH, Goh BPL, Chou LM (2005b) Seasonal reproduction in equatorial reef corals. Invertebr Reprod Dev 48: 207-218. Guest JR, Baird AH, Goh B, Chou L (2002) Multispecific, synchronous coral spawning in Singapore. Coral Reefs 21: 422-423. Guzman HM (1991) Restoration of coral reefs in Pacific Costa Rica. Conservation Biology 5: 189-195. Harriott VJ, Fisk DA (1988) Accelerated regeneration of hard corals: A manual for coral reef users and managers Technical Memorandum 16. Great Barrier Reef Authority, Townsville, Australia, pp 42. Harrison PL, Wallace CC (1990) Reproduction, dispersal and recruitment of scleractinian corals. In: Dubinzky Z (ed.) Ecosystems of the world: coral reefs, Vol 25. Elsvier, Amsterdam, pp 133-207. Harrison PL, Babcock RC, Bull GD, Oliver JK, Wallace CC, Willis BL (1984) Mass spawning in tropical reef corals. Science 223: 1186-1189. Hatta M (2005) An affair in the morning after mass spawning. Coral Reefs 24:102. Heyward AJ, Collins JD (1985) Fragmentation in Montipora ramosa: the genet and ramet concept aplied to a reef coral. Coral Reefs 4: 35-40. Highsmith RC (1982) Reproduction by fragmentation in corals. Mar Ecol Prog Ser 7: 207-226. Hilbertz WH (1992) Solar-generated building material from seawater as a sink for Carbon. Ambio 21: 126-129. Hilbertz WH, Goreau TJ (1996). Method of enhancing the growth of aquatic organism, and structures created thereby. http://www.uspto.gov, US Patent no. 5,543,034. Hughes TP, Baird AH, Dinsdale EA, Moltschaniwskyj NA, Pratchett MS, Tanner JE, Willis BL (1999) Patterns of recruitment and abundance of corals along the Great Barrier Reef. Nature 397: 59-63. Jaap WC (2000) Coral reef restoration. Ecol Eng 15: 345-364. Jeffrey SW, Humphrey GF (1975) New Spectrofhotometric Equations for Determining Chlorophylls

a, b, c1 and c2 in Higher Plants, Algae and Natural Phytoplankton. Biochem. Biochem. Physiol. Pflanzen: 191-194. Jones RS (1997a) Changes in zooxanthellar densities and chlorophyll concentrations in corals during and after a bleaching event. Mar Ecol Prog Ser 158: 51-59. Jones RS (1997b) Zooxanthellae loss as a bioassay for assessing stress in corals. Mar Ecol Prog Ser 149: 163-171.

22 Kruger A, Schleyer MH (1998) Sexual reproduction in the coral Pocillopora verrucosa (Cnidaria: Scleractinia) in KwaZulu-Natal, South Africa. Mar Biol 132: 703-710. Krupp DA, Jokiel PL, Chartrand TS (1993) Asexual reproduction by Fungia scutaria on dead coralla. Proc 7th Int Coral Reef Symp 1: 527-534. Lindahl U (1998) Low-tech restoration of degraded coral reefs through transplantation of staghorn corals. Ambio 27 (8): 645-650. Maida M, Coll JC, Sammarco PW (1994) Shedding New Light On Scleractinian Coral Recruitment. Journal Of Experimental Marine Biology And Ecology 180: 189-202. Mangubhai S, Harrison PL, Obura DO (2007) Patterns of coral larval settlement on lagoon reefs in the Mombasa Marine National Park and Reserve, Kenya. Mar Ecol Prog Ser 348: 149- 159. Mangubhai S, Harrison PL (2008) Asynchronous coral spawning patterns on equatorial reefs in Kenya. Mar Ecol Prog Ser 360: 85-96. Maragos JE (1974) Coral Transplantation: A method to create, preserve, and manage coral reefs Sea Grant Deposity UNIHI-SEAGRANT AR-74-03. University of pp 30. Mehta A (1999) Bunaken - Natural History Book. USAID, Manado, pp 271. Miller K, Mundy CN (2003) Rapid settlement in broadcast spawning corals: Implications for larval dispersal. Coral Reefs 22: 99-106. Minton D, Lundgren I (2006) Coral Recruitment and Sedimentation in Asan Bay and War in the Pacific NHP, Guam. National Park Service, Guam, pp 28. Nonaka M, Baird AH, Kamiki T, Yamamoto HH (2003) Reseeding the reefs of Okinawa with the larvae of captive-bred corals. Coral Reefs 22: 34. Nozawa Y, Harrison PL (2007) Effects of elevated temperature on larval settlement and post- settlement survival in scleractinian corals, Acropora solitaryensis and Favites chinensis. Mar Biol 152: 1181-1185. Nozawa Y, Harrison PL (2000) Larval settlement patterns, dispersal potential, and the effect of temperature on settlement of larvae of the reef coral, Platygyra daedalea, from the Great Barrier Reef. Proc 9th Int Coral Reef Symp 1: 409-415. Okubo N, Isomura N, Motokawa T, Hidaka M (2007) Possible Self-Fertilization in the Brooding Coral Acropora (Isopora) brueggemanni. Zoological Science 24: 277-280. Okubo N, Tangiguchi H, Motokawa T (2005) Successful methods for transplanting fragments of Acropora formosa and Acropora hyacinthus. Coral Reefs 24: 333-342. Oliver JK (1984) Intra-colony variation in the growth of Acropora formosa: Extension rates and skeletal structure of white (zooxanthellae-free) and brown-tipped branches. Coral Reefs 3: 139-147. Oliver JK, Babcock RC, Harrison PL, Willis BL (1988) The geographic extent of mass coral spawning: clues to proximate and ultimate causal factors. Proc 6th Coral Reef Symp 2: 803-810.

23 Oliver JK, Babcock RC (1992) Aspects of the fertilization ecology of broadcast spawning corals: Sperm dilution effects and in situ measurements of fertilization. Biol Bull 183: 409-417. Penland L, Kloulechad D, Idip D, van Woesik R (2004) Coral spawning in the western Pacific Oean is related to solar insolation: evidence of multiple spawning evetns in Palau. Coral Reefs 23: 133-140. Precht WF, Dodge RE (2002) Coral reef restoration in the next Millennium Implications for Coral Reef Management and Policy: Relevant Findings from the 9th Int Coral Reef Symp. USAID, Washington D.C., pp 57-58. Razak TB (2006) Hard coral & reef fish community on the EcoReefs rehabilitation site, Manado Tua Island, Bunaken National Park, North Sulawesi, Indonesia pp 35. Richmond RH (1988) Competency and dispersal of spawned versus brooded coral planula larvae. Proc 6th Int Coral Reef Symp 2: 827-831. Rinkevich B (2005) Conservation of coral reefs through active restoration measures: Recent approaches and last decade progress. Environ Sci Technol 39: 4333-4342. Rinkevich B (1995) Restoration strategies for coral reefs damaged by recreational activities: the use of sexual and asexual recruits. Restor Ecol 3: 241-251. Rinkevich B (2000) Steps towards the evaluation of coral reef restoration by using small branch fragments. Mar Biol 136: 807-812. Rosen BR, Taylor JD (1969) Reef coral from Aldabra: new mode of reproduction. Science 166: 119-121. Sabater MG, Yap HT (2002) Growth and survival of coral transplants with and without

electrochemical deposition of CaCO3. J Exp Mar Biol Ecol 272: 131-146. Sammarco PW (1982) Polyp bail-out: An escape response to environmental stress and a new means of reproduction in corals. Mar Ecol Prog Ser 10: 57-65. Shlesinger Y, Loya Y (1985) Coral community reproductive patterns: Red Sea versus the Great Barrier Reef. Science 228: 1333-1335. Schmidt S (2007) In-Situ-Observation von Korallenrecruitment via Fluoreszenzfotografie in Indonesien. MSc-Thesis. Carl von Ossietzky Universität, Oldenburg, pp 68. Schmidt-Roach S, Kunzmann A, Arbizu PM (2008) In situ observation of coral recruitment using fluorescence census techniques. J Exp Mar Biol Ecol 367: 37-40. Schuhmacher, H, van Treeck P, Eisinger M, Paster M (2000) Transplantation of coral fragments from ship groundings on elctrochemically formed reef structures. Proc 9th Int Coral Reef Symp 2: 983-990. Sier CJS, Olive PJW (1994) Reproduction and reproductive variability In the coral Pocillopora verrucosa from the Republic of Maldives. Marine Biology 118: 713-722. Smith LD, Hughes TP (1999) An experimental assessment of survival, re-attachment and fecundity of coral fragments. J Exp Mar Biol Ecol 235: 147-164. Smith SR (1997) Patterns of coral settlement, recruitment and juvenile mortality with depth at

24 Conch Reef, Florida. Proc 8th Int Coral Reef Symp: 1197-1202. Spalding MD, Grenfell AM (1997) New estimates of global and regional reef areas. Coral Reefs 16: 225-230. Spieler RE, Gilliam DS, Sherman RL (2001) Artificial substrate and coral reef restoration: what do we need to know to know what we need. Bull Mar Sci 69: 1013-1030. Stimson JS (1978) Mode and Timing of Reproduction in Some Common Hermatypic Corals of Hawaii and Enewetak. Mar Biol 48: 173-184. Stimson JS, Kinzie III RA (1991) The temporal pattern and rate of release of Zooxanthellae from the reef coral Pocillopora damicornis (Linnaeus) under nitrogen-enrichment and control conditions. J Exp Mar Biol Ecol 153: 63-74. Tioho H, Tokeshi M, Nojima S (2001) Experimental analysis of recruitment in a scleractinian coral at high latitude. Mar Ecol Prog Ser 213: 79-86. Tomascik T, Mah, A, Nontij A, Moosa M (1997) The ecology of the Indonesian Seas (vol 1). In: Periplus (ed), Hong Kong, pp 642. van Treeck P, Schuhmacher H (1999) Artificial reefs created by electrolysis and coral transplantation - an approach ensuring the compatibility of environmental protection and diving tourism. Estuar Coast Shelf Sci 49: 75-81. van Treeck P, Schuhmacher H (1997) Initial survival of coral nubbins transplanted by a new coral transplantation technology. Mar Ecol Prog Ser 150: 287-292. van Treeck P, Schuhmacher H (1998) Mass diving tourism - a new dimension calls for new management approaches. Mar Poll Bull 37: 499-504. Veron JEN (2000) Corals of the world. Australian institute of Marine Science, Townsville, pp 398. Vize PD, Embresi JA, Nickell M, Brown DP, Hagman DK (2005) Tight temporal consistency of coral mass spawning at the Flower Garden Banks, Gulf of Mexico, from 1997–2003. Gulf of Mexico Science: 107-114. Wainwright SA (1963) Skeletal organization in the coral, Pocillopora damicornis. Quart J Micr Sci 104: 169-183. Wallace CC (1985a) Reproduction, recruitment and fragementation in nine sympatric species of the coral genus Acropora. Mar Biol 88: 217-233. Wallace CC (1985b) Seasonal peaks and annual fluctuations in recruitment of juvenile scleractinian corals. Mar Ecol Prog Ser 21: 289-298. Wallace CC, Bull GD (1982) Patterns of juvenile coral recruitment on a reef front during a summer spawning period. Proc 4th Int Coral Reef Symp 2: 345-350. Ward S (1995) Two patterns of energy allocation for growth, reproduction and storage in the scleractinian coral Pocillopora damicornis. Coral Reefs 14: 87-90. Wilkinson C (ed.) (2004) Status of coral reefs of the world: 2004. Australian Institute of Marine Science, Townsville, pp 264.

25 Willis BL, Babcock RC, Harrison PL, Oliver JK (1985) Patterns in the mass spawning of corals on the Great Barrier Reef from 1981 to 1984. Proc 5th Int Coral Reef Congr 4: 343-348. Yap HT (2000) The case for restoration of tropical coastal ecosystems. Ocean Coast Manag 43: 841-851. Yap HT, Alino PM, Gomez ED (1992) Trends in growth and mortality of three coral species ( Scleractinia), including effects of transplantation. Mar Ecol Prog Ser 83: 91- 101. Yap HT, Gomez ED (1985) Growth of Acropora pulchra: III. Preliminary observations on the effects of transplantation and sediment on the growth and survival of transplants. Mar Biol 87: 203-209. Yap HT, Molina RA (2003) Comparison of coral growth and survival under enclosed, semi-natural conditions and in the field. Mar Pol Bul 46: 858-864. Yeemin T, Sutthacheep M, Pettongma R (2006) Coral reef restoration projects in Thailand. Ocean Coast Manag 49: 562-575. Young TP (2000) Restoration ecology and conservation biology. Biol Cons 92: 73-83.

26 Chapter 2

Determination of fine-scale temporal variation in acroporid and pocilloporid settlement in North Sulawesi, Indonesia

27 First row from left to right: development Pocillopora early stage recruit to colony Second row from left to right: development Acropora early stage recruit to colony Scale bars represent 1 mm

28 Marine Ecology Progress Series (submitted)

Determination of fine-scale temporal variation in acroporid and pocilloporid settlement in North Sulawesi, Indonesia

Sascha B.C. Romatzki and Andreas Kunzmann

Center for Tropical Marine Ecology (ZMT), Fahrenheitstraße 6, 28359 Bremen

Abstract

Spats of either brooding or broadcast spawning scleractinian corals play an important role in the recovery and maintenance of coral reef systems. In this study variation in settlement of corals in a low-latitude Indonesian reef was observed. Limestone tiles were used as settlement substrate at four sites in the Bunaken National Park and nearby reefs over a 2-year period. Tiles were replaced every two months. The time of settlement of Acroporidae and Pocilloporidae recruits was determined with a weekly accuracy using size-age-keys from a concomitant fluorescence study for the early detection of coral settlement conducted in the same sites. A total of 4280 Pocillopora and 3150 Acropora recruits were recorded on 1440 tiles. Sizes of pocilloporid recruits ranged from 0.5 to 8.4 mm and for Acroporidae from 0.5 to 4.9 mm. Recruits were found during each sampling period and at each site on tiles throughout the year. Abundance of pocilloporid recruits showed no clear seasonality in settlement in contrast to those of Acroporidae, which peaked in April of each year and between May and June in 2006 and 2007. Settlement rates appear to be correlated with temperature changes. The highest average coral density per sampling period was 5.8 recruits tile-1 ± 0.24 (mean ± SD). A threefold higher Acroporidae density (7.71 recruits tile-1 ± 1.15, mean ± SD) compared to other acroporid peaks was calculated for May 2007. These data suggest that there is a steady settlement of pocilloporid recruits in the monitored reefs while settlement for Acroporidae showed both clear seasonal peaks and occasional settlement year around. Based on the observation of several distinct peaks in settlement per year in successive years, it is believed that there must be several mass spawning events.

Key words: Indonesia; North Sulawesi; settlement; fluorescence technique; seasonality; temperature; coral; reproduction

29 Introduction

The recovery and maintenance of coral reefs depends largely on the settlement of planktonic coral larvae (e.g. Baird and Hughes 1997). Sources of these larvae are either hermaphroditic or gonochoritic reef corals which can reproduce sexually or asexually (Szmant 1985). Accordingly they can act as broadcast spawners or planulae-releasing brooders. Broadcast spawners release their gametes for external fertilization of eggs followed by an extended planktonic period that can lead to a wide-scale dispersal before settlement (e.g. Harrison and Wallace 1990, Wilson and Harrison 1998). However, more recent studies also show the ability for a rapid settlement of larvae (Miller and Mundy 2003) as it is commonly described for brooders (e.g. Harii et al. 2001, Tioho et al. 2001). Coral reproduction and settlement of coral larvae show spatial and temporal variations that differ between regions (Fadlallah 1983). Corals in the Gulf of Mexico (Vize et al. 2005), Caribbean (Szmant 1985), Western Australia (Simpson et al. 1993) and Great Barrier Reef (Harrison et al. 1984) follow more or less seasonal reproduction patterns, while those from the northern Red Sea (Shlesinger and Loya 1985) or North and South Taiwan (Dai et al. 1992) follow no clear seasonal patterns. Oliver et al. (1988) recorded a decrease in reproductive seasonality and spawning synchrony of coral species from high latitude sides in the Great Barrier Reef to low latitude reefs of Papua New Guinea. In contrast, recruits of brooding corals are the dominant species on settlement tiles at high latitude reefs with a corresponding and declining rate of broadcasting spawning coral recruits (Banks and Harriott 1996, Fautin 2002, Hughes et al. 2002, Glassom et al. 2006). Many of the equatorial and subequatorial brooding corals, particularly in the family Pocilloporidae, follow hereby a lunar periodicity (Fadlallah 1983). The best-known but not yet totally understood phenomenon of annual cnidarian reproduction is probably the synchronous mass spawning as first described from the Great Barrier Reef (Harrison et al. 1984, Willis et al. 1985, Babcock et al. 1986), which is correlated with a temporary extensive increase of larvae settlement on settlement tiles shortly after (Wallace 1985, Willis and Oliver 1988). Earlier observations of mass spawning came explicit from regions of higher latitudes. Their characteristically changes in environmental parameters as water temperature and light intensity gave arguments for a geographical limitation of this phenomenon (Shlesinger and Loya 1985, Harrison and Wallace 1990): Although species develop independently from each other, their needs for suitable environmental parameters for a successful reproduction are more or less similar (Guest et al. 2005a). More recent studies reported mass spawning corals from the low latitude reefs in the Solomon Islands, which lack major differences in temperature or (Baird et al. 2001), and from an equatorial tropical reef in Singapore (Guest et al. 2002, Guest et al. 2005b). In the present work the results of a two-year settlement study from the Celebes Sea of North

30 Sulawesi are presented. The Celebes Sea as part of the coral Triangle is one of the world richest areas in coral diversity (Green and Mous 2004), though least studied and most endangered. Though 445 species of scleractinian corals from more than 58 genera and subgenera have been counted alone in the Bunaken Nationalpark area (Mehta 1999, Donelly et al. 2002), the spatial and temporal variability in coral settlement from this South-East Asian region is still relatively poor described with little data available. The objective of this study was to investigate, a) if seasonal recruitment patterns in approximate distance to the equator in North Sulawesi occur, b) which coral families are involved in it and c) if mass spawning is part of these patterns.

Material and methods

Location

This study was conducted in North Sulawesi, Indonesia in and close to the Bunaken National Park. A total of four sites have been chosen for settlement studies (Figure 1). The north tip of Sulawesi and its islands are affected by turbid water with strong currents, rips and up and down with water depths down to 2000 m starting a few hundred meters off . All sites were within a radius of 20 km.

Fig. 1 Map of North Sulawesi showing the 4 sampling sites: Meras close to the city of Manado; Raymond’s Point on Bunaken; Lakehe on the island Gangga and the island Lihage south-west of Gangga.

31 The site Meras (1˚31’47.50” N; 124˚49’54.51’’E) is located in the bay of Manado. It has an extended reef flat of up to 100 m width, turning into a sloping parallel to the coastline. and current are light, while visibility is often low. Even though the City of Manado is 1.5 km south of the reef, influence must be low due to a mainly north-south current. Acropora colonies are found in several patches. Raymond’s Point (1˚37’54.04”N; 124˚44’11.49”E) is a drop off on the wind exposed side of Bunaken Island in the Bunaken National Park. It is located on the entrance of a channel between Bunaken and Manado Tua. The current is moderate to strong with shifting directions. Coral colonies are compact, and Pocilloporidae are the dominating family. Lihage (1˚45’36.57”N; 125˚02’10.03”E), is a small island with an expansive reef flat and moderate current. The reef flat starts at a depth of 1 m and descends to 12 m. The edge of the flat goes over into a rubble slope. Acropora palifera and various Pocillopora dominate the reef. Lakehe (1˚46’29.14”N; 125˚03’28.90”E), on the east side of Gangga island, is a small fringing reef with a thin belt of coral cover, that quickly turns into a rubble slope. The visibility is often poor and murky, while the current is moderate.

Settlement tiles

Three settlement frames per site were deployed in the reefs of Meras and Raymond’s Point from March 2005 to June 2007. Additional six frames were installed in the reefs of Lihage and Lakehe from June 2005 until May 2007.

Fig. 2 Settlement frame holding 3 untreated limestone tiles (15x15x1 cm) per side. Tiles were fixated with cables ties and hanging in a 60˚ angle.

32 Each frame covered four directions (Figure 2), with every side housing three settlement tiles (12 tiles frame-1, 36 tiles site-1). Tiles (untreated, uncoated limestone, 15 x 15 x 1 cm) were hung into the frame with cable ties at a 60˚ angle. The space between tiles and reef bottom was approximately 30 cm. The frames were arranged in a line at 6-7 m depth with a distance of 50 m from frame to frame. Tiles were replaced in two-month intervals and labeled during collection. Labels gave information about site, frame, and position in frame. A fourth frame of identical design in each site was used to examine long-term settlement effects. Therefore tiles were left at the site for 12 months before being collected. After collection tiles were directly transferred into bleach-solution and left for 24-48 hours, they were rinsed afterwards with running water to remove sand and soft organisms and then dried. The front- and backside of the tiles, as well as all four edges were examined with a dissecting microscope. Each coral-recruit was photographed with a macro lens mounted on an Olympus C5050 for later identification and measurement. Outer and inner diameters of the recruits were either measured directly with a binocular micrometer or in scaled photos analysed with the ImageJ photo-processing program (National Institutes of Health, USA). Recruits were categorized into the families Pocilloporidae, Acroporidae and Poritidae (Figure 3) following Babcock et al. (2003). Those not belonging to the three families were recognized as “other families”. A more detailed categorization of recruits from this study did not seem reliable, even though there are significant differences of the inner diameter and skeletal morphology of the primary polyp of Pocilloporidae (Baird and Babcock 2000, Babcock et al. 2003). Skeletons of recruits in this study were not of high enough quality to be able to rely on the measurements taken for the inner diameter or to identify clear differences. A concurrent study focused on the early detection of coral recruits by using fluorescence technique (Piniak et al. 2005, Baird et al. 2006). Schmidt-Roach et al. (2008) were hereby able to establish a general growth key for Pocilloporidae (0.24 mm week-1) and Acroporidae (0.17 mm week-1) on settlements tiles of same material, collected from an identical depth in Meras. He measured a

a) b) c)

Fig. 3 Recruits of a) Pocilloporidae, b) Acroporidae and c) Poritidae. Notice the columella in the center of the pocilloporid recruit that is missing in the acroporid recruit. Distance between the black lines on the left of each photo is 1 mm.

33 mean diameter of 1.14 mm for Acroporidae and 1.09 mm for Pocilloporidae at first appearance. By using this key, a calculation of theoretical settlement time down to one week was possible with following formulas: a) Age in weeks = (outer recruit diameter – diameter at first appearance)/growth key +1 b) Settlement during week of year = week of tile recovery – age in weeks

If the calculated settlement time resulted in a negative number or 0, the age was taken as not older than one week. If age resulted in a number bigger than the time of tiles in the water, the time of tiles in the water was taken as maximum age. If the week number felt into two month, the higher number of days decided about month affiliation.

Environmental Parameters

Visibility was measured with a Secchi-disc and salinity from water samples in 5 m depth with a standard refractometer (English et al. 1994). Both parameters were measured in weekly intervals between December 2006 and June 2007 in Meras and Raymond’s Point. Submergible temperature loggers (HOBO TidbiT, USA) were deployed in 5 m depth in every site and set to measure in 30- minute intervals. In the end only incomplete data from Meras and Raymond’s could be recovered, because of malfunctions, and lost and stolen loggers.

Statistics

Non-parametric Kruskal-Wallis tests (Sokal and Rohlf 1985) were used for all settlement data as: a) data were unbalanced due to stolen settlement frames occurring once in every site b) normality could not be obtained by transformation of data. Comparison between sites was followed by a Mann-Whitney-U-test for pair wise comparison of the sites. All data were analyzed with JMP 7.0 (SAS-Institute, USA).

Results

Environmental Parameters

Dry season is generally from May to September (south-east monsoon), while the wet season

34 is from November to March (north-west monsoon). Salinity ranged from 30 to 32‰ in 5 m independent from season. In Meras visibility ranged from as low as 6.5 m in January, 2007 to as good as 28 m in May, 2007, and in Raymond’s Point from 27.5 m in December, 2006, to 14.5 m in January, 2007. Average temperature in Meras between December 2005 and January 2007 was 28.5˚C ± 0.6 (mean ± SD) with a minimum temperature of 22.4˚C and a maximum of 30.5˚C. On March 3, 2006, the device logged a distinct with a temperature increase of almost 5 ˚C within 30 minutes. Late June/ early July (29.0˚C ± 0.3, mean ± SD) and November/ December 2006 (29.1˚C ± 0.5, mean ± SD) were the warmest times of the year in contrast to middle March/ April (27.8˚C ± 0.3, mean ± SD) and September (28.1˚C ± 0.8, mean ± SD). Raymond’s Point temperatures were logged between December 2005 and July 2006. Mean temperature was 28.9˚C ± 0.5 (mean ± SD) with a minimum of 25.8˚C and maximum of 30.5˚C. Temperatures were on average 0.45˚C higher than mean temperatures in Meras (28.5˚C ± 0.6, mean ± SD) with a minimum of 22.8˚C and maximum of 30.1˚C during this time.

Taxonomic pattern and recruit density in the two-month sampling period

A total of 4170 (55%) Pocilloporidae, 3107 (41%) Acroporidae and 188 (2.5%) Poritidae recruits were found on 1440 tiles from March 2005 to May 2007 at the four sites. An additional 113 recruits (1.5%) were classified as “others” without a further differentiation in family level. Pocilloporid recruits were the dominant family in all sites except Meras, which was dominated by Acroporidae (Table 1). Lakehe showed the highest proportion of Pocilloporidae with 79 % while the Acroporidae proportion was here the lowest (19%). The number of recruits per tile ranged from 0 to a maximum of 28 (= 550 recruits m-2). The site Lihage had the highest average number with 5.8 ± 0.24 (mean ± SD; = 114 recruits m-2) recruits per tile, while the lowest was found at Raymond’s Point with 4.2 ± 0.19 recruits (mean ± SD; = 82

Table 1. Number of recruits on all examined tiles (n) per station from March 2005 until May 2007. Number of tiles per station is indicated by n.

Lakehe Lihage Meras Raymond’s Point (n = 348) (n = 350) (n = 373) (n = 370) no (%) no (%) no (%) no (%) Pocilloporidae 1297 (79) 1032 (50) 820 (39) 874 (56)

Acroporidae 317 (19) 948 (46) 1196 (39) 584 (37) Poritidae 12 ( 1) 42 ( 2) 39 ( 2) 94 ( 6) Others 16 ( 1) 23 ( 1) 53 ( 3) 17 ( 1)

35 a) Raymond'sPoint

100% other families 90% Poritidae 80% Acroporidae Pocilloporidae 70% 60% 50% 40% 30% 20% 10% 0%

-07 ch-06 ch-07 y May-05 July-05 May-06 July-06 uary-07 Ma July-07 Mar Mar January-06 Jan September-05November-05 September-06November-06

b) Molas

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

-07 ry-07 May-05 July-05 May-06 July-06 May July-07 March-06 March-07 January-06 ptember-06 Janua September-05November-05 Se November-06

c) Lihage

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 6 -05 -06 May-05 July May-06 July-06 May-07 July-07 March-0 March-07 ptember-05 January January-07 Se November-05 September-06November-06

d) Lakehe

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 5 -0 05 -07 er er- May-05 July-05 b May-06 July-06 May-07 July-07 mber-06 March-06 e March-07 January-06 January Septemb Novem September-06Nov

Fig. 4 Percentage average family distribution per 2-month tile and sampling period in each site: a) Raymond’s Point, b) Meras, c) Lihage and d) Lakehe. Tiles were first deployed for Raymond’s Point and Meras in March 2005 and in June 2005 for Lihage and Lakehe.

36

temperature [˚C] temperature temperature [˚C] temperature 30.0 29.5 29.0 28.5 28.0 27.5 27.0 26.5 26.0 25.5 30.0 29.5 29.0 28.5 28.0 27.5 27.0 26.5 26.0 25.5 May -07 May -07 Apr -07 Apr -07 temperature Mar -07 Mar -07 Feb -07 Feb -07 Jan -07 Jan -07 Acroporidae Dec -06 Dec -06 Nov -06 Nov -06 Oct -06 Oct -06 Pocilloporidae Sep -06 Sep -06 Aug -06 Aug -06 Jul -06 Jul -06 Jun -06 Jun -06 May -06 May -06 Apr -06 Apr -06 Mar -06 Mar -06 Feb -06 Feb -06 Jan -06 Jan -06 Dec -05 Dec -05 Nov -05 Nov -05 Oct -05 Oct -05 Sep -05 Sep -05 Aug -05 Aug -05 Jul -05 Jul -05 Jun -05 Jun -05 May -05 May -05 b) Meras a) Raymond's Point Apr -05 Apr -05

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

recruits tile tile recruits recruits tile tile recruits

-1 -1

Fig. 5 a&b Average settlement number (mean ± SD) per week and site on 2-month tiles: a) Raymond’s Point, b) Meras. Month on x-axis are divided into weeks. Notice that the last data point outside b) is 7.71 recruits tile-1. Average temperature data per week are plotted on the secondary y-axis for a) Meras from December 2005 until January 2007 and b) Raymond’s Point from December 2005 until Juli 2006.

37 May -07 May -07 Apr -07 Apr -07 Mar -07 Mar -07 Feb -07 Feb -07 Jan -07 Jan -07 Dec -06 Dec -06 Nov -06 Nov -06 Oct -06 Oct -06 Sep -06 Sep -06 Aug -06 Aug -06 Jul -06 Jul -06 Jun -06 Jun -06 May -06 May -06 Apr -06 Apr -06 Mar -06 Mar -06 Feb -06 Feb -06 Jan -06 Jan -06 Dec -05 Dec -05 Nov -05 Nov -05 Oct -05 Oct -05 Sep -05 Sep -05 Aug -05 Aug -05 Jul -05 Jul -05 Jun -05 Jun -05 May -05 May -05 d) Lakehe c) Lihage Apr -05 Apr -05

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

recruits tile tile recruits recruits tile tile recruits

-1 -1

Fig. 5 c&d Average settlement number (mean ± SD) per week and site on 2-month tiles: c) Lihage and d) Lakehe. Month on x-axis are divided into weeks.

38 recruits m-2). Recruits on tiles were found during each sampling period and at each site throughout the year. Family composition differed between sampling periods and sites (Figures 4). Pocilloporid coral recruits dominated tiles in three stations during almost every sampling period, beside the May 2005, July 2006 and May 2007 periods, which were dominated by Acroporids. Meras was the only site, where Acroporids mainly dominated tiles throughout the time of sampling with an extreme peak of 90% in the July 2007 period. Exceptional periods were only in January and March 2006 when pocilloporid recruits dominated there.

Size of recruits

The largest pocilloporid recruit observed was 8.0 mm after 90 days immersion (34 polyps) and the smallest 0.5 mm (1 polyp). Settlement size of Acroporidae ranged from 0.5 to 4.9 mm (1 and 11 polyps; 90 days immersion) and for those of Poritidae from 0.5 to 3.0 mm (1 and 11 polyps; 84 days).

Comparison of locations

Number of recruits of both families between locations was significantly different for year 1 (June 2005 – May 2006) and year 2 (June 2006 – April 2007) (Table 2). In year 1 most Acroporidae recruits were found in Lihage, while in year 2 more Acroporidae were found in Meras (Table 3). In both years Lakehe had the lowest number of Acroporidae. In year 1 the highest number of Pocilloporidae was found in Lakehe, which is more than double the numbers in Meras. In the second year increased numbers of Pocilloporidae could be found in Lihage with the lowest number in Raymond’s Point. There was no significant difference between the two years (all sites) for number of Pocilloporidae but a significant difference for those of Acroporidae (Kruskal-Wallis: 2 = 33.098, p < 0.001, df = 1), which was higher in the second year. The significant increase of Acroporidae could also be seen within all sites, except for Lakehe where it did not differ. Consequently Pocilloporidae

Table 2. Number of recruits between locations was compared with Krusakal-Wallis test with  = 0.05 for significance for pooled years: year 1 = June 2005 until May and year 2 = June 2006 until May 2007 (for Lakehe and Lihage April 2007).

year 1 year 2 df p df p Acroporidae 3 187.303 < 0.001 3 187.303 < 0.001 Pocilloporidae 3 56.284 < 0.001 3 21.94 < 0.001

39 Table 3. Average number (± SD) of recruits per tile based on monthly data: June 2005 until May 2006 was pooled as year 1 and June 2006 until May 2007 (for Lakehe and Lihage April 2007) as year 2. Numbers in {} are converted recruit number per m2. Years were compared with a Mann- Whitney U test with  = 0.05 for significance.

year 1 year 2 df 2 p Lakehe Acroporidae 0.19 (±0.03) 0.61 (±0.08) 1 26.986 < 0.001 Pocilloporidae 2.09 (±0.13) 1.17 (±0.07) 1 19.171 < 0.001 Lihage Acroporidae 1.09 (±0.08) 1.28 (±0.09) 1 0.342 0.559 Pocilloporidae 1.22 (±0.07) 1.37 (±0.07) 1 1.528 0.217 Meras Acroporidae 1.07 (±0.06) 1.79 (±0.15) 1 23.055 < 0.001 Pocilloporidae 0.97 (±0.07) 0.94 (±0.06) 1 2.779 0.096 Raymond’s Acroporidae 0.57 (±0.04) 0.77 (±0.08) 1 6.844 0.009

Point Pocilloporidae 1.15 (±0.06) 0.90 (±0.09) 1 0.803 0.370

showed a significant decrease in Lakehe for the second year (Mann-Whitney comparison of pairs: p < 0.001). Composition of both families in terms of settlement number within locations was significantly different: Pocilloporidae were dominant in every site for both years (Mann-Whitney pair wise comparison: p < 0.05) with the exception of Meras in year 1 when families did not differ significantly, but were dominated by Acroporidae in year 2 (Mann-Whitney: p < 0.001).

Fine-scale temporal settlement patterns

A comparison of Acroporidae and Pocilloporidae recruits in one-month intervals was possible after applying the established growth key (Schmidt-Roach et al. 2008). The difference between months for Acroporidae as well as Pocilloporidae was significant in every site and every year (Kruskal- Wallis; p < 0.005). Seasonal peaks of Acroporidae were reached in May 2005 in Raymond’s Point (Figure 5a) and Meras (Figure 5b) and in April and June 2006 in all sites. Additional peaks were found in April 2007 for Lihage (Figure 5c) and Lakehe (Figure 5d), while settlement started to increase in Meras and Raymond’s Point a month later in May 2007. In Meras during the final sampling period in May 2007, the theoretical acroporid settlement (7.71 ± 1.15 recruits tile-1, mean ± SD) was three times higher than any of the previous acroporid peaks (max 2.75 ± 0.61 recruits tile-1, mean ± SD) observed. At the same time settlement in Raymond’s Point was more than doubled (1.97 ± 0.79 recruits tile-1, mean ± SD) to previously observed peaks (max 0.78 ± 0.2 recruits tile-1, mean ± SD). The highest number of pocilloporid recruits was counted on the August 2005 tiles in Lakehe with 3.13 ± 0.09 (mean ± SD). Distinct seasonal settlement patterns for Pocilloporidae could only be seen for Lakehe (August and October 2005 and 2006, with an

40 additional peak in May 2006), while the other sites showed no clear seasonal patterns for this family. Pocillopora settlement numbers seemed to decrease from 2005 to 2006, while Acropora numbers increased during that period.

Orientation of recruits on tiles

Although the number of recruits on the backside of the tiles (surface area = 225 cm2) was higher than on the edges (surface area = 15 cm2), the theoretical number after transformation to equal surface areas (1 m2) showed a different trend: 79.57 % of all recruits would have been on the lower edge of the tile, followed by the backside with 9.35 % and the lateral sides (4.35 %). The tile frontside (2.92 %) and the upper edge (3.81 %) were the places with the lowest recruitment. Settlement preferences (Figure 6) of Acroporidae and Pocilloporidae recruits were identical for the front- and backside of tiles with only minor differences. Therefore Acroporidae preferred the more shaded upper and lower edges, while Pocilloporidae were more frequently found on the lateral edges. Poritidae never settled on the upper edge and were rare on the tile front side (0.6 %). 89.3 % of all “other families” were found on the lower edge but never on the lateral edges.

100%

90%

80%

70%

60%

50% sides upper edge 40% lower edge frontside backside coral family distribution coral family distribution 30%

20%

10%

0% Pocilloporidae Acroporidae Poritidae other families

Fig. 6 Percentage distribution of coral recruit families on 2-month tile surfaces displayed as pooled data for all sites.

41 3.5 3 a) Raymond's Point b) Meras

3 2.5

2.5 2 -1 2 1.5

recruits tile 1.5

1 1

0.5 0.5

0 0 0-1 1-2 2-3 3-5 5-10 10-15 15-20 >20 0-1 1-2 2-3 3-5 5-10 10-15 15-20 >20

3 3 c) Lihage d) Lakehe other families Poritidae 2.5 2.5 Acroporidae Pocilloporidae

2 2

recruits tile 1.5 1.5

1 1

0.5 0.5

0 0 0-1 1-2 2-3 3-5 5-10 10-15 15-20 >20 0-1 1-2 2-3 3-5 5-10 10-15 15-20 >20 size category in [mm] size category in [mm]

Fig. 7 Distribution of recruits in size classes and families on 1-year tiles for a) Raymond’s Point, b) Meras, c) Lihage and d) Lakehe.

Settlement on 1-year-tiles

An average of 9.8 ± 2.8 (mean ± SD; 190 recruits m-2) were found on 1-year tiles in Raymond’s Point (Figure 7a), 3.4 ± 0.8 (mean ± SD; 67 recruits m-2) in Meras (Figure 7b), 5.7 ± 0.6 (mean ± SD; 112 recruits m-2) in Lihage (Figure 7c) and 1.9 ± 0.3 recruits (mean ± SD; 37 recruits m-2) in Lakehe (Figure 7d). Newly and very small recruits were rare on all tiles. Percentage distribution of Pocilloporidae on all 1-year tiles was 58.7%, for Acroporidae 25%, 0.5% for Poritidae and 15% for all unidentified corals. Pocilloporidae coral recruits dominated at all sites, while Poritidae was only recorded in Raymond’s Point. Raymond’s Point counted also the higher number of big sized recruits. Encrusting Bryozoa were the dominant space competitors on every tile and at every

42 site, covering most of the tiles surfaces, followed by Bivalvia, Balinidae and Canalipalpata in descending order of importance.

Discussion

Retrospect on methods

Laboratory experiments showed that species identification due to distinctive morphological characteristics of newly settled pocilloporid recruits of known age and source of origin is possible (Babcock, 2003). Crucial identification character hereby was the size of intact skeletons of the primary polyps. Recruits in the present work were not identified to such a level, as the quality of high portions of examined skeletons did not allow a reliable identification. Therefore the diagrams of weekly settlement (Fig 5), that assumes an average start size rather than considering the latter mentioned size differences between species, can be a source for errors. By interpreting the peaks on a monthly basis, these errors should be small. While counting recruit skeleton it was furthermore expected, that they were from recruits still alive when tiles were harvested. Consequently the age determination of recruit skeletons has to be interpreted with caution as these “skeletal fingerprints” remain on substrate although the animals are already dead for days or weeks (Harrison and Wallace 1990) distorting the actual time of settlement.

Pocilloporidae vs. Acroporidae

The majority of pocilloporid corals are known to release brooded larvae while acroporid corals are generally characterized as broadcast spawners, however there are exceptions: Pocillopora verrucosa is a spawner (Sier and Olive 1994) while species like Acropora brueggemanni from the acroporid subgenus Isopora are brooders (Okubo et al. 2007). The larvae of broadcast spawning corals are reported to dominate the coral spats composition in tropical reefs (Baird and Babcock 2000). The present study is in contrast to previous findings, as spats of the genus Pocilloporidae dominated here the majority of all tiles. Furthermore pocilloporid spats were found year around in all sites, so that a lack of recruits was more the exception than the rule, this also concurs with an equatorial settlement studies from Kenya (Mangubhai et al. 2007). The pocilloporid percentage of spats on the 2-month and 1-year tiles were also higher than those of pocilloporid spats found on 6-month tiles in Komodo (Fox 2004), south of the Equator. However, the difference could be explained due to uncertainties in species identification to some degree. There fore the acroporid percentage on 1-year tiles in the present study is comparable with the Komodo findings. A twofold higher number of pocilloporid recruits with distinguishable temporal patterns as found

43 in Lakehe is in direct contrast to the other examined sites. As Lakehe is also a site with low coral cover and diversity it is possible that most larvae came rather from a specific pocilloporid species close by than from multiple species (Tioho et al 2001). Settlement number peaked in 10 to 12 weeks intervals and this intervals fit well into up to 6 annual reproductive cycles for some brooding pocilloporid corals (Stoddart and Black 1985). Therefore histological examination will be necessary to elucidate the reproductive mechanism and number of annual gametogenic cycles of the dominating pocilloporids. The abundance of Acroporidae showed in contrast to Pocilloporidae a clear seasonality in settlement in all sites with peaks in the months of April, between May and June as also in November of 2006. In the present study from north of the Equator the abundance patterns in Pocilloporidae are identical with those in Komodo south of the equator (Fox 2004). Therefore Acroporidae in present study were spawning during both monsoon seasons, with higher settlement rates during the SE- monsoon, while acroporid abundance increased in Komodo during the NW-monsoon. These contradictory findings might be a direct result of high spat mortality and the cruder sampling interval in Komodo to the present weekly settlement data. Those Acroporidae which were found sporadically in other months than the peak season, were found in sites where isoporan corals (e.g. A. brueggemanni and A. palifera) were identified as part of the local coral community. Furthermore planulae in different development stages were found in fractures from freshly broken off branches of various A. brueggemanni colonies during weekly checks (Romatzki, pers. observation). Both findings give strong evidence that isoporan Acropora are a possible larvae source.

Fine-scale settlement and field observations

The calculated times for settlement with the help of family growth keys are rather approximate values, not exactly reflecting the exact time of larvae and gamete release. Nevertheless these times correspond to direct spawning observations of Acropora yongei and A. pulchra in the first week of May 2007 in Lihage (Romatzki, pers. observation). Both species coexisted in a single stand with the dimensions 15 by 30 m. Spawning occurred during two subsequent nights, four and five days after full moon. Patchy spawn slicks floating on the surface, as also drifting eggs in the water column were witnessed in front of the neighbouring island Gangga. It is unlikely that all observed eggs originated from the prior described single coral stand. Therefore it is more probable that various Acropora colonies and species in diverse locations were spawning synchronous during that event and in subsequent nights. Three weeks later an extreme increase in acroporid settlement was noticed on tiles in Meras and Raymond’s Point. As adult colonies of A. yongei occurred in these sites just sporadically, while A. pulchra was not present at all, it suggests that also here several other acroporid species were involved, contributing to additional evidence of mass spawning.

44 Settlement competency and physical factors

It is highly likely that these recruits were coming from the particular spawning night as several days pass between spawning, egg fertilization, planula development and settlement competency. The time from gamete release until settlement competency can range from as short as 2.5 days (Miller and Mundy 2003) to a maximum competency time of 78 days (Richmond 1988) depending on the broadcast spawning species. For brooding species the time of settlement competency can even be as long as 103 days (Richmond 1987), though larvae of brooding Pocillopora damicornis were found in close proximity to the parent colonies, indicating settlement immediately after release (Stoddart and Black 1985, Tioho et al. 2001). Nevertheless, the mass of larvae from specific species follows a distinct rather than a continuous settlement pulse (Miller and Mundy 2003). Maximum settlement rates between 10 and 32 days after spawning (Wilson and Harrison 1998) fit well into the above observations of spawning event and sudden increase in settlement. Furthermore the simulations demonstrated that positive buoyant particles could stay well within a reef area for up to 20 days (Black 1993), although the ultimate settlement site for free drifting larvae depends on the bathymetry, weather, interaction of reef and current. Water that flows along a curved or uneven coastline or object may separate from it and generate an eddy likewise retaining recruits (Largier 2003). The reduced rates of coral settlement in Raymond’s Point might be explained due to its exposure to strong currents. Located in the channel between Bunaken und Manado Tua, currents are strong during intermediate periods but calm during tides. In combination with the reef topography (see description of the site) shelter, allowing settlement of planulae might be limited. Moreover, water flow does negatively influence the settlement rates for some brooding species (Harii and Kayanne 2002), although settlement of other species is not affected by it. However swimming speed of planula is too slow to headway against more rapid horizontal currents, though swimming speed would allow substantial vertical migrations of up to 11 m h-1 (reviewed in Hodgson 1985). Also if currents can reduce the actual settlement success on a short temporal range, they also can be an advantage for those corals that manage to settle under such conditions. Currents can support the removal of sediment both from the coral surface and the reef, encounter rate of zooplankton and dissolved nutrients uptake, modulating physiological processes as enhancing the photosynthesis by symbiontic algae and increasing the rates of coral tissue (reviewed in Sebens et al. 2003). This would explain the twofold higher settlement number on 1-year tiles in Raymond’s Point in comparison to the 2-month tiles and 1-year tiles of all other sites. These sites were either more current protected or more exposed to anthropogenic influences than Raymond’s Point. It is well known that temperature affects physiological processes during all life-history stages. In case of planktonic larvae their respiration and metabolism rates are increasing with higher temperatures (Maldonado and Young 1996): Larvae of Porites lutea increased their respiration significantly when exposed to a small temperature change in the range from 26.8˚C to 29.7˚C

45 (Edmunds 2005). Settlement competency as discussed above is seen to be directly related to the energy reserves a larvae has (Richmond 1987), so that high metabolism rates can lead to a shortage of energy and with that to a shortened competency time (Edmunds et al. 2001). Furthermore increasing temperatures may modify the selective value for larvae to settle in specific locations. However, larvae settlement is greater at higher temperatures whereas larvae of different species are adapted to species specific and narrow temperatures ranges. Post-settlement of larvae can be negatively effected if these temperature ranges are exceeded for a prolonged time (Coles 1985, Nozawa and Harrison 2007). Settlement rates in Meras of both, acroporid and pocilloporid spats followed temperature changes in either direction: dropping temperatures were followed by regressive settlement rates and increasing settlement followed raising temperatures. This effect was most apparent in April until June 2006 when steadily increasing mean temperatures from 27.8˚C to 29.4˚C were overlapping with increasing settlement rates of Acroporidae. As earlier and later parts of the logger data are lost, it will remain subjective if settlement rates consequently followed a temperature increase, making further assessments necessary. Nevertheless, with the hourly high-frequency temperature variability as measured in present sites, rapid temperature increases in certain ranges are suggested to have a trigger effect on larvae settlement. Finally, by summarizing the previous results it can be said that the examined sites were served with coral recruits from both, seasonal as well as continuous monthly events. Spats from broadcast spawning species were more seasonally present in specific months of the year, while brooders were abundant year around. Overlapping of monthly and seasonal settlement patterns lead to intensive settlement peaks that could be backtracked with some accuracy to specific weeks of the year. Time and intensity of settlement in combination with spawning observation indicate mass spawning and a high potential of self-seeding in the examined reefs. Furthermore there is presumptive evidence that intensity of settlement followed seasonal temperature changes and is triggered by impulsive changes to higher temperatures.

Acknowledgments

We would like to thank C. Muller and G. Davi for their hospitality, all staff of Froggies Divers, Bunaken and Gangga Spa and Resort for all support in the dive logistics. We further more would like to thank S. Ferse and S. Schmidt-Roach for help in the field and S. Mangubhai for helpful comments on the manuscript. This research was supported by a DAAD (German Academic Exchange Service) scholarship for doctoral candidates.

46 References

Babcock RC, Baird AH, Piromvaragorn S, Thomson DP, Willis BL (2003) Identification of scleractinian coral recruits from Indo-Pacific reefs. Zoological Studies 42: 211-226. Babcock RC, Bull GD, Harrison PL, Heyward AJ, Oliver JK, Wallace CC, Willis BL (1986) Synchronous spawnings of 105 scleractinian coral species on the Great Barrier Reef. Mar Biol 90: 379-394. Baird AH, Hughes TP (1997) Spatial variation in coral recruitment around Lizard Island, Australia. Proc 8th Int Coral Reef Sym 2: 1207-1210. Baird AH, Babcock RC (2000) Morphological differences among three species of newly settled pocilloporid coral recruits. Coral Reefs 19: 179-183. Baird AH, Saddler C, Pitt M (2001) Synchronous spawning of Acropora in the Solomon Islands. Coral Reefs 19: 286. Baird AH, Salih A, Trevor-Jones A (2006) Fluorescence census techniques for the early detection of coral recruits. Coral Reefs 25: 73-76. Banks SA, Harriott VJ (1996) Patterns of coral recruitment at the Gneering Shoals, southeast Queensland, Australia. Coral Reefs 15: 225-230. Black KP (1993) The relative importance of local retention and inter-reef dispersal of neutrally buoyant material on coral reefs. Coral Reefs 12: 43-45. Coles SL (1985) The effects of elevated temperature on reef coral planula settlement as related to power station entrainment. Proc 5th Int Coral Reef Symp 4: 171-176 Dai CF, Soog K, Fan TY (1992) Sexual reproduction of corals in Northern and Southern Taiwan. Proc 7th Int Coral Reef Symp 1: 448-455. Donelly R, Neville D, Mous PJ (2002) Report on a rapid ecological assessment of the , Papua, Eastern Indonesia, held October 30 - November 22, 2002. The Nature Conservancy and WWF, Sanur, Bali, p 250. Edmunds PJ (2005) Effect of elevated temperature on aerobic respiration of coral recruits. Mar Biol 146: 655-663. Edmunds PJ, Gates RD, Gleason DF (2001) The biology of larvae from the reef coral Porites astreoides, and their response to temperature disturbances. Mar Biol 139: 981-989. English S, Wilkinson C, Baker V (1994) Surevey manual for tropical marine resources. Australian Institute of Marine Science, Townsville, p 117. Fadlallah HY (1983) Sexual reproduction, development and larval biology in scleractinian corals. Coral Reefs 2: 129-150. Fautin DG (2002) Reproduction of Cnidaria. Can J Zool 80: 1735-1754. Fox HE (2004) Coral recruitment in blasted and unblasted sites in Indonesia: assessing rehabilitation potential. Mar Ecol Prog Ser 269: 131-139.

47 Glassom D, Celliers L, Schleyer MH (2006) Coral recruitment patterns at Sodwana Bay, South Africa. Coral Reefs 25: 485-492. Green A, Mous PJ (2004) Delineating the Corl Triangle, its ecoregions and functional seascape. The Nature Conservancy, Southeast Asia Center for Marine Protected Areas, Sanur, Bali, p 26. Guest JR, Baird AH, Goh B, Chou L (2002) Multispecific, synchronous coral spawning in Singapore. Coral Reefs 21: 422-423. Guest J, Baird AH, Goh BPL, Chou LM (2005b) Reproductive seasonality in an equatorial assemblage of scleractinian corals. Coral Reefs 24: 112-116. Guest JR, Baird A, Goh BPL, Chou LM (2005a) Seasonal reproduction in equatorial reef corals. Invert. Reprod. & Develop 48: 207-218. Harii S, Kayanne H (2002) Larval settlement of corals in flowing water using a racetrack flume. Marine Technology Society Journal 36: 76-79. Harii S, Omori M, Yamakawa H, Koike Y (2001) Sexual reproduction and larvae settlement of the zooxanthallate coral Alveopora japonica Eguchi at high latitudes. Coral Reefs 20: 19-23 Harrison PL, Wallace CC (1990) Reproduction, dispersal and recruitment of scleractinian corals. In: Dubinzky Z (ed) Ecosystems of the world. Elsvier, Amsterdam, pp 133-207. Hodgson G (1985) Abundance and distribution of planktonic coral larvae in Kaneohe Bay, Oahu, Hawaii. Mar Ecol Prog Ser 26: 61-71. Harrison PL, Babcock RC, Bull GD, Oliver JK, Wallace CC, Willis BL (1984) Mass spawning in tropical reef corals. Science 223: 1186-1189. Hughes TP, Baird AH, Dinsdale EA, Harriott VJ, Moltschaniwskyj NA, Pratchett MS, Tanner JE, Willis BL.(2002) Detecting regional variation using meta-analysis and large- scale sampling: Latitudinal patterns in recruitment. Ecology 83: 436-451. Largier JL (2003) Considerations in estimating larvaal dispersal distances from oceanographic data. Ecol Appl 13: 71-89. Maldonado M, Young CM (1996) Effects of physical factors on larval behavior, settlement and recruitment of four tropical demosponges. Mar Ecol Prog Ser 138: 169-180. Mangubhai S, Harrison PL, Obura DO (2007) Patterns of coral larval settlement on lagoon reefs in the Mombasa Marine National Park and Reserve, Kenya. Mar Ecol Prog Ser 348: 149- 159. Mehta A (1999) Bunaken - Natural History Book. USAID, Manado, p 271. Miller K, Mundy CN (2003) Rapid settlement in broadcast spawning corals: Implications for larval dispersal. Coral Reefs 22: 99-106. Nozawa Y, Harrison PL (2007) Effects of elevated temperature on larval settlement and post- settlement survival in scleractinian corals, Acropora solitaryensis and Favites chinensis. Mar Biol 152: 1181-1185. Okubo N, Isomura N, Motokawa T, Hidaka M (2007) Possible Self-Fertilization in the Brooding

48 Coral Acropora (Isopora) brueggemanni. Zoological Science 24: 277-280. Oliver JK, Babcock RC, Harrison PL, Willis BL (1988) Geographic extent of mass coral spawning: clues to ultimate causal factors. Proc 6th Int Coral Reef Symp 2: 803-810. Piniak GA, Fogarty ND, Addison CM, Kenworthy WJ (2005) Fluorescence census techniques for coral recruits. Coral Reefs 24: 496-500. Richmond RH (1987) Energetics, competency, and long-distance dispersal of planula larvae of the coral Pocillopora damicornis. Mar Biol 93: 527-533. Richmond RH (1988) Competency and dispersal of spawned versus brooded coral planula larvae. Proc 6th Int Coral Reef Symp 2: 827-831. Schmidt-Roach S, Kunzmann A, Arbizu PM (2008) In situ observation of coral recruitment using fluorescence census techniques. J Exp Mar Biol Ecol 367: 37-40. Sebens KP, Helmuth B, Carrington E, Agius B (2003) Effects of water flow on growth and energetics of the scleractinian coral Agaricia tenuifolia in Belize. Coral Reefs 22: 35-47. Shlesinger Y, Loya Y (1985) Coral community reproductive patterns: Red Sea versus the Great Barrier Reef. Science 28: 1333-1335. Sier CJS, Olive PJW (1994) Reproduction and reproductive variability In the coral Pocillopora verrucosa from the Republic of Maldives. Marine Biology 118: 713-722. Simpson CJ, Cary JL, Masini RJ (1993) Destruction of corals and other reef animals by coral spawn slicks on Ningaloo Reef, Western Australia. Coral Reefs 12: 185-191. Sokal RR, Rohlf FJ (1985) Biometry. Freeman and Co., San Francisco, CA, p 859. Stoddart JA, Black R (1985) Cycles of gametogenesis and planulation in the coral Pocillopora damicornis Mar Ecol Prog Ser 23: 153-164. Szmant AM (1985) Reproductive ecology of Caribbean reef corals. Coral Reefs 5: 43-54. Tioho H, Tokeshi M, Nojima S (2001) Experimental analysis of recruitment in a scleractinian coral at high latitude. Mar Ecol Prog Ser 213: 79-86. Vize PD, Embresi JA, Nickell M, Brown DP, Hagman DK (2005) Tight temporal consistency of coral mass spawning at the Flower Garden Banks, Gulf of Mexico, from 1997–2003. Gulf of Mexico Science: 107-114. Wallace CC (1985) Seasonal peaks and annual fluctuations in recruitment of juvenile scleractinian corals.Mar Ecol Prog Ser 21: 289-298. Willis BL, Babcock RC, Harrison PL, Oliver JK (1985) Patterns in the mass spawning of corals on the Great Barrier Reef from 1981 to 1984. Proc 5th Int Coral Reef Congr 4: 343-348. Willis BL, Oliver JK (1988) Inter-reef dispersal of coral larvae following the annual mass spawning on the Great Barrier Reef. Proc 6th Int Coral Reef Symp 2: 853-859. Wilson JR, Harrison PL (1998) Settlement-competency periods of larvae of three species of scleractinian corals. Mar Biol 131: 339-345.

49 Chapter 3

Gametogenesis of four scleractinian corals in the Celebes Sea

50 51 Proceedings of the 11th International Coral Reef Symposium (submitted)

Gametogensis of four scleractinian corals in the Celebes Sea Sascha B.C. Romatzki and Andreas Kunzmann

Center for Tropical Marine Ecology (ZMT), Fahrenheitstraße 6, 28359 Bremen

Abstract

Spatial and temporal variability in reproduction of equatorial coral species is still relatively undescribed. This study examined the mode and timing of reproduction in common species of North Sulawesi, Indonesia. Samples from Pocillopora verrucosa, Seriatopora hystrix, Acropora pulchra and A. yongei were collected in 2-weekly intervals between June 2005 and May 2007. All collected samples were examined for their gametic cycle by using standard histological examination to quantify the gonad development stage. The presence of zooxanthella in P. verrucosa eggs indicated a biannual reproduction cycle with spawning periods between January until March and June until August. S. hystrix produced larvae all year round with planulae present in every month. Egg-sperm-bundles were found annually in A. pulchra in April 2006 and March 2007, while A. yongei contained such bundles in May 2006 and 2007 when spawning was also witnessed.

Key words: Indonesia; scleractinia; gamaetogenesis; equatorial

Introduction

Investigations on coral reproduction have been conducted throughout the world, revealing a variety of reproduction modes, patterns and timing between, but also within, geographical regions (Sammarco 1985, Richmond and Hunter 1990): While the majority of coral species found in the Indo-Pacific reproduce as broadcast spawners, those in the Atlantic are predominantly brooders (Harrison and Wallace 1990). The majority of corals from high latitude reefs tend to have a seasonal reproduction schedule, whereas with decreasing distance to the equator clear schedules are blurring: Oliver (1988) measured a breakdown in reproductive seasonality and spawning synchrony of coral species from high to low latitudes so that the breeding season is hypothesized to

52 be protracted near the equator (Harrison and Wallace 1990). Latitudinal variation in reproduction time has also been recorded on the species level for Acropora (Isopora), changing from seasonal in higher to year-around in lower latitudes (Kojis 1986). Species also show variability in the number of reproduction cycles: brooders often release planulae multiple times per annum, while the most common pattern in broadcast spawners is single annual spawning of colonies (Fadlallah 1983). However, biannual and multiple spawning of some broadcasting species was also reported (Stobart et al. 1992, Mangubhai and Harrison 2008). The common cues for timing and synchronisation of coral reproduction are seen in environmental parameters. Each of them is suggested to work on a different scale, including water temperature and solar insulation for the season or month (Shlesinger and Loya 1985, Penland et al. 2004), tides for the day of the month (Harriott 1983) and light intensity for the time of the day (Harrison et al. 1984). Although species develop independently from each other, their needs for suitable environmental parameters for a successful reproduction are more or less similar. Their simultaneous response – with the extreme found in the annual mass spawning in the Great Barrier Reef - is seen as a result of strong selective pressure promoting fertilization success (Harrison and Wallace

Figure 1 Map of North Sulawesi showing the four sampling sites: Meras close to the city of Manado, Raymond’s Point and Fukui on Bunaken and Lihage - a small island south-west of Gangga.

53 1990). Further evidence for this hypothesis is seen in the absence of mass spawning events in the northern Red Sea, with its relatively constant environmental conditions (Shlesinger and Loya 1985). However, mass spawning corals have been observed in the low latitude Solomon Islands (Baird et al. 2001) and in a Singaporean equatorial reef (Guest et al. 2002), which both have only minor fluctuations in environmental parameters. Based on these findings Guest et al. (2005a) argues that mass spawning is more the result of reproduction success through synchronisation, than it is of environmental parameters; and with that it is as likely to occur in equatorial assemblages, as it is at higher latitudes. Data on the gametogenesis of numerous coral species and precisely predictable dates of their spawning are available for many areas in the Caribbean, Gulf of Mexico (e.g. McGuire 1998, Beaver et al. 2004, Bastidas et al. 2005, Vize 2005, Vize et al. 2005, Schloeder 2007) and the Great Barrier Reef (e.g. Wallace 1985, Willis et al. 1985), but not much data is available from South-East Asian reefs, even though this region encompasses almost 1/3 of the world’s coral reefs (reviewed in Guest et al. 2005b) and more than 75% of all known scleractinian corals. Hereby the reefs of North Sulawesi, Indonesia lay in the centre of this coral diversity hotspot (Green and Mous 2004). Despite the importance of the area, in terms of its worldwide outstanding coral diversity, very little is known about its coral reproduction and timing (e.g. Tomascik et al. 1997, Bachtiar 2000, Crabbe and Smith 2002, Zakaria 2004). In this paper we examine the gametogenesis of four selected scleractinian species for their timing and a possible seasonality in an equatorial reef. Histological examinations of common brooders and spawners from these reefs were used to reveal their gametic-cycles.

Methods

The study has been conducted in North Sulawesi, Indonesia (Fig. 1). Samples of four common scleractinian coral species were collected from four sites in approximate distance to the Equator (Table 1). Fukui was sampled in a weekly, Meras and Raymond’s Point in a two-weekly and Lihage in a five-week interval. Three branches were taken repeatedly from the same tagged colonies on each sampling date in a depth of 6 m (S. hystrix was collected in 30 m). Samples were immediately fixated in 10% Formalin with seawater for at least a week, decalcified in 11% formic acid for up to three days with the solution changed when necessary, rinsed with running tap water and then stored in 70% Ethanol. The tissue preparation for histological examination followed standard procedures as described in Tioho (2001). Tissues were stained with Gill-Hematoxylen and Eosin-Phloxin. Tissue slides were cut 6 m thick and examined by microscopy. Development condition of gonads was determined following Szmant-Froehlich et al. (1985) with slightly modified criteria for each species: Four development stages were classified by criteria such as the geometric mean diameter [square root (greatest x least diameter)] and

54 Figure 2 Cross section of a Pocillopora verrucosa polyp. ms = mesenteria, mf = mesenterial filament, nc = nucleus, vc = vacuole, sp = spermary, oo = oocyte

morphology of oocytes and testes. Further criteria for stage IV were based on the presence of zooxanthellae in P. verrucosa oocytes (Hirose et al. 2001), when eggs were packed together into bundles in A. yongei and A. pulchra or planulae presence in S. hystrix.

Results and Discussion

In North Sulawesi Pocillopora verrucosa was found to be a hermaphrodite spawner with ovaries and testes in the same polyp (Fig. 2). This concurs with most literature findings, with the exception of one observation from Enewak, where P. verrucosa has been reported as a brooder (Stimson 1978). Samples of each month contained stage II, III and IV oocytes, while stage I was rarely found (Fig. 3). The lack of stage I oocytes might be a result of a very short-lived stage, as was formerly hypothesized by Kruger and Schleyer (1998). Comparative data of mature egg size showed a diameter range from 53.5m ± 15.7 in the Maldives (Sier and Olive 1994) to approximately 150 m estimated in the Red Sea (Fadlallah 1985, Shlesinger and Loya 1985), South Africa (Kruger and Schleyer 1998) and Japan (Hirose et al. 2001). However, oocyte size in this study ranged from a minimum of 12 m to a maximum of 158 m, with a pre-spawning average size (stage IV) of 110 m ± 15 (n=58). Although spawning could never be witnessed, a presence of zooxanthellae in

55 Table 1 Overview of collected specimen at the four sampled sites. For each sample three polyps were examined by using standard histological methods.

Location Coordinates Species collected No samples examined Meras 1˚31’47.50” N P. verrucosa n = 177 (Manado Bay) 124˚49’54.51”E A. yongei n = 49

Raymond’s Point 1˚37’54.04”N P. verrucosa n = 284 (Bunaken Island) 124˚44’11.49”E A. yongei n = 87

Fukui 1˚45’36.57”N P. verrucosa n = 165 (Bunaken Island) 124˚44’23.39”E S. hystrix n = 129

Lihage Island 1˚45’36.57”N P. verrucosa n = 94 125˚02’10.03”E A. yongei n = 78 A. pulchra n = 41

P. verrucosa oocytes indicated an imminent spawning. Oocytes showed to take up zooxanthellae three to four days prior to the spawning date (Hirose et al. 2001). Despite the fact that stage IV oocytes were found in numerous months, those containing zooxanthellae were limited (Table 2). Samples containing zooxanthellae were found one to three days prior to a New Moon, but never before a Full Moon. These findings basically fit into the reports that were identifying P. verrucosa as a spawner around New Moon (Fadlallah 1985, Shlesinger and Loya 1985, Sier and Olive 1994, Kruger and Schleyer 1998, Hirose et al. 2001). The presence of zooxanthellae in Table 2 also show that P. verrucosa in North Sulawesi spawns more than once per annum with two cycles most likely, as also reported in several other species (Richmond and Jokiel 1984, Stobart et al. 1992, Mangubhai and Harrison 2008). However, in this study it could not be clarified if it is a feature of single colonies or reflects asynchrony between multiple colonies within a population, as the data of tagged colonies was limited. Asynchrony of spawning periodicity between the populations of the different sites, while just separated by less than 30 km, was shown during some months. Spawning seemed to occur in four-month intervals in three sites, while in Lihage a tendency to 5 months was more likely. Inconsistent reproduction patterns between closely located sites, or different morphs within the same location are known from closely related, but brooding Pocillopora (Fan et al. 2002, Lin 2005). Patches of Acropora yongei and A. pulchra were forming together a dense single coral stand of approximately 15 x 30 m in Lihage. While all A. pulchra samples were collected in Lihage, samples from A. yongei were also collected from tagged colonies in Raymond’s Point and Meras. Oocyte diameter of both species had an approximate range from 20 m to 290 m (Fig. 4). Ripe

56 oocytes in A. pulchra were packed in egg-sperm-bundles found on 19th of March and one week before full moon on 6th April 2006. As the March sampling occurred one week after full moon and as the consecutive April sampling contained a bigger portion of stage IV eggs, April is suggested to be the spawning month for 2006. Ripe egg bundles were also found on the 15th May 2006 in A. yongei samples two days after the full moon. A visual census of freshly broken off branches on the 5th May 2007 in Lihage revealed egg-sperm bundles in A. yongei colony patches. Synchronous spawning in these multiple patches was observed four and five days after full moon on the 6th and 7th May 2007. The bigger event took place in the first night and encompassed the entire coral stand.

Figure 3 Pocillopora verrucosa: Diameter and development stage of oocytes (top) and spermaries (bottom) for each sampling month. Notice that April has not been sampled.

57 Table 2 Overview of Pocillopora verrucosa sampling dates and locations for the most likely or close to spawning times in accordance to zooxanthellae in oocytes, oocyte diameter, lunar condition and development stage. NM = new moon, FM = full moon, + = number of days after lunar phase, - = number of days before lunar phase, y = yes, n = no

Sample date Location zooxanthellae Lunar condition oocyte stage mean size 13 Feb 2006 Lihage n FM-1 3 76±4.18 26 Feb 2006 Raymond’s Point y NM-1 4 113±12.3 26 Mar 2006 Meras, Fukui y NM-3 4 118±7.5 28 Jun 2006 Raymond’s Point y NM+3 4 107±11.5 21 Aug 2006 Fukui y NM-2 4 111±2.5 21 Aug 2006 Meras y NM-2 4 106±7.4 15 Aug 2006 Lihage n NM-8 3 95±4.7 09 Jan 2007 Lihage n NM-8 3 76±13.8

Such a split spawning in corals over consecutive nights is suggested to be the result of a delay in oocyte maturation caused by variation in trophic conditions of individual polyps, due to the lower light intensity in shaded colony areas (Shimoike et al. 1992). As maturation status must have changed in present observation within a day from “not ready” to “ready to spawn”, a variation in concentration of sex steroid or pheromones in the water as potential spawning triggers (Tarrant 2005) seems to be more likely. However, no further mature eggs were found in tissue samples of other months, pointing into an annual reproduction of the examined colonies of both species in North Sulawesi. Therefore a conclusion on the population level needs more quantitative data, as regional and annual differences in within-population reproductive synchrony or split spawn events in consecutive months are common in Acropora (reviewed in Baird and Guest 2008). Seriatropora hystrix polyps contained often one oocyte surrounded by one pair of spermaries. Occasionally two equal sized oocytes were located in one polyp. Oocyte size ranged from 25 to 233 m in diameter, while spermary diameter ranged from 10 to 182 m. The single planulae findings per polyp with a maximum diameter of 408 m in present study, staying in contrast to regularly findings of two planulae per polyp with a size of 633 ± 12 m in the southern Great Barrier Reef (Tanner 1996), Tanner (1996) defined planulae as mature when mesenteries were visible. In present study mesenteries were visible despite the smaller larvae size. Planulae were present year around (Fig. 5) in examined colonies with no sign of seasonality in contrast to Hawaii and Enewetak (Stimson 1978) and in the Great Barrier Reef (Tanner 1996) where planulation showed seasonality. However sexual products of different development stages in S. hystrix, from stage I oocytes to planulae, were found simultaneously in almost every examined colony during every sampling month. The overlapping of gametic cycles in combination with findings of empty polyps during different lunar phases gives evidence for planuale release without any lunar periodicity as

58 A. yongei

A. pulchra

Figure 4 Diameter and development stage of oocytes for each sampling month for Acro- pora yongei (top) and A. pulchra (bottom). Notice that April 2007 was not sampled, but spawn- ing in A. yongei was observed in beginning of March 2007.

59 M M AAAASSADDM JJJJJ OONN D J J F F M M

2006 2007

Figure 5 Size of Seriatropora hystrix planulae and when the were found per sampling for each sampling month. Letters stay for month starting with March 2006 and ending with March 2007. Vertical lines indicate Full Moon.

also found by Stimson (1978) and Tanner (1996). However, their studies were conducted in subtropical waters, where water temperature changes between seasons, while specimen in the present study were collected in 30 m depth in equatorial waters, where average water temperature is relatively stable. Temperatures measured in a nearby station over a one-year period from December 2005 until January 2007 in 5m depth were ranging between 27˚C and 29.5˚C. Hereby March and September were the coldest months (average 28˚C) with June and December being the warmest (average 29˚C). Though studies with spawning species proved an important role for seasonal temperature changes to trigger the gametic cycle (Jokiel et al. 1985), such changes are not essential for this year-around brooder, that continues reproduction even when maintained at constant temperatures as shown in aquaria (Petersen et al. 2007). The presence of zooxanthellae in P. verrucosa eggs and egg-sperm-bundles found in A. pulchra and A. yongei are both indicators for an imminent spawning. Based on these indicators all three spawning species showed some degree of seasonality. Planulae of the brooding species S. hystrix found during all lunar phases and in all month of an entire sampled year gives evidence for a year-around brooding period in North Sulawesi.

60 Acknowledgments

We like to thank Froggies Divers and Gangga Island Resort & Spa for their continuous support in dive logistics, E. Tambajong for letting us use his private lab and for help in the histological preparation, K. v. Juterzenka and A. Baird for useful comments on the interpretation of the tissue slides.

References

Alcala AC, Dy-Liacco MS, Alcala LC Benthic lifeform composition of two coral reef sites in the central Visayas, Philippines. Regional Symposium on Living Resources in Coastal Areas: 69-74. Bachtiar I (2000) Reproduction of three scleractinian corals (Acropora cytherea, A. nobilis, Hydnophora rigida) in eastern Lombok Strait, Indonesia. Ilmu Kelautan (Journal of Indonesian Marine Sciences) 21: 18-27. Baird AH, Guest JR (2008) Spawning synchrony in scleractinian corals: Comment on Mangubhai & Harrison (2208). Mar Ecol Prog Ser (in press). Baird AH, Saddler C, Pitt M (2001) Synchronous spawning of Acropora in the Solomon Islands. Coral Reefs 19: 286. Bastidas, C., Croquer, A., Zubillaga, A.L., Ramos, R., Kortnik, V., Weinberger C, Marquez LM (2005) Coral mass- and split-spawning at a coastal and an offshore Venezuelan reef, southern Caribbean Hydrobilogia 541: 101-106. Beaver CR, Evans EF, de la Cerda AV (2004) Mass spawning of reef corals within the Veracruz Reef System, Veracruz, Mexico. Coral Reefs 23: 324. Crabbe MJC, Smith DJ (2002) Comparison of two reef sites in the Wakatobi Marine National Park (SE-Sulawesi, Indonesia) using digital image analysis. Coral Reefs 21: 242-244. Fadlallah HY (1983) Sexual reproduction, development and larval biology in scleractinian corals. Coral Reefs 2: 129-150. Fadlallah YH (1985) Reproduction in the corl Pocillopora verrucosa on the reefs adjacent to the industrial city of Yanbu (Red Sea, Saudi Arabia). Proc 5th Int Coral Reef Symp 4: 313- 318. Fan TY, Li JJ, Le SX, Fang LS (2002) Lunar periodicity of larval release by pocilloporid corals in southern Taiwan. Zoological Studies 41: 288-294. Green A, Mous PJ (2004) Delineating the Coral Triangle, its ecoregions and functional seascape. The Nature Conservancy, Southeast Asia Center for Marine Protected Areas, Sanur, Bali, p 26. Guest J, Baird AH, Goh BPL, Chou LM (2005a) Seasonal reproduction in equatorial reef corals.

61 Invert Reprod Develop 48: 207-218. Guest J, Baird AH, Goh BPL, Chou LM (2005b) Reproductive seasonality in an equatorial assemblage of scleractinian corals. Coral Reefs 24: 112-116. Guest JR, Baird AH, Goh B, Chou L (2002) Multispecific, synchronous coral spawning in Singapore. Coral Reefs 21: 422-423. Harriott VJ (1983) Reproductive ecology of four Scleractinian Species at Lizzard Island, Great Barrier Reef. Coral Reefs 2: 9-18. Harrison PL, Babcock RC, Bull GD, Oliver JK, Wallace CC, Willis, BL (1984) Mass spawning in tropical reef corals. Science 223: 1186-1189. Harrison PL, Wallace CC (1990) Reproduction, dispersal and recruitment of scleractinian corals. In Dubinzky Z. (ed.) Ecosystems of the world. Elsvier, Amsterdam, pp 133-207. Hirose M, Kinzie III RA, Hidaka M (2001) Timing and process of entry of zooxanthellae into oocytes of hermatypic corals. Coral Reefs 20: 273-280. Jokiel PL, Ito RY, Liu, PM (1985) Night irradiance and synchronization of lunar release of planula larvae in the reef coral Pocillopora damicornis. Mar Biol 88: 167-174. Kojis BL (1986) Sexual reproduction in Acropora (Isopora) (Coelenterata: Scleractinia) II. Latitudal variation in A. palifera from the Great Barrier Reef and Papua New Guinea. Mar Biol 91: 311:18. Kruger A, Schleyer MH (1998) Sexual reproduction in the coral Pocillopora verrucosa (Cnidaria: Scleractinia) in KwaZulu-Natal, South Africa. Mar Biol 132: 703-710. Lin K (2005) Timing of larval release by five coral species in Southern Taiwan: Seasonality, lunar and diurnal periodicity Marine Biology. M.Sc. thesis, National Sun Yat-Sen University, Kaohsiung p 69. Mangubhai S, Harrison PL (2008) Gametogenesis, spawning and fecundity of Platygyra daedalea (Scleractinia) on equatorial reefs in Kenya. Coral Reefs 27: 117-122. McGuire M (1998) Timing of larval release by Porites astreoides. Coral Reefs 17: 369-375. Oliver JK, Babcock RC, Harrison PL, Willis BL (1988) Geographic extent of mass coral spawning: clue to ultimate causal factors. Proc 6th Int Coral Reef Symp 2: 803-810. Penland L, Kloulechad D, Idip D, van Woesik R (2004) Coral spawning in the western Pacific Ocean is related to solar insolation: evidence of multiple spawning events in Palau. Coral Reefs 23: 133-140. Petersen D, Falcato J, Gilles P, Jones R (2007) Sexual reproduction of scleractinian corals in public aquariums current status and future perspectives. Int Zoo Yb 41: 122-137. Richmond RH, Hunter CL (1990) Reproduction and recruitment of the corals: comparisons among the Caribbean, the Tropical Pacific, and the Red Sea. Mar Ecol Prog Ser 60: 185-203. Richmond RH, Jokiel PL (1984) Lunar periodicity in larval release in the reef coral Pocillopora damicornis at Enewetak and Hawaii. Bull Mar Science 34: 280-287. Sammarco PW (1985) The Great Barrier Reef vs. the Caribbean: comparisons of grazers, coral

62 recruitment patterns and reef recovery. 5th International Coral Reef Congress 4: 391-397. Schloeder C (2007) Reproductive Ecology of the Caribbean Reef Coral Porites furcata (Anthozoa, Scleractinea, Poritidae) in Bocas del Toro. Panama. Ph.D. thesis, University of Bremen, p 166. Shimoike K, Hayashibara T, Kimura T, Omori M (1992) Observations of Split Spawning in Acropora spp. at Akajima Island, Okinawa. Proc &th Int Coral Reef Symp 1: 484-488. Shlesinger Y, Loya Y (1985) Coral community reproductive patterns: Red Sea versus the Great Barrier Reef. Science 28: 1333-1335. Sier CJS, Olive PJW (1994) Reproduction and reproductive variability In the coral Pocillopora verrucosa from the Republic of Maldives. Marine Biology 118: 713-722. Stimson JS (1978) Mode and timing of reproduction in some common hermatypic Corals of Hawaii and Enewetak. Mar Biol 48: 173-184. Stobart B, Babcock RC, Willis BL (1992) Biannual spawning of three species of scleractinian coral from the Great Barrier Reef. Proc 7th Int Coral Reef Symp 1: 494-499. Szmant-Froelich A, Reutter AM, Riggs L (1985) Sexual reproduction of Favia fragum (Esper): lunar patterns of gametogenesis, embryogenesis and planulation in Puerto Rico. Bull Mar Sci 37: 880-892. Tarrant AM (2005) Endocrine-like signaling in cnidarians: Current understanding and implications for ecophysiology. Integr Comp Biol 45:201-214. Tioho H, Tokeshi M, Nojima S (2001) Experimental analysis of recruitment in a scleractinian coral at high latitude. Mar Ecol Prog Ser 213: 79-86. Tomascik T, Mah A, Nontij A, Moosa M (1997) The ecology of the Indonesian Seas, vol 1. Periplus, Hong Kong, p 642. Vize PD (2005) Deepwater broadcast spawning by Montastraea cavernosa, Montastraea franksi, and Diploria strigosa at the Flower Garden Banks, Gulf of Mexico. Coral Reefs 25: 169- 171. Vize PD, Embresi JA, Nickell M, Brown DP, Hagman DK (2005) Tight temporal consistency of coral mass spawning at the Flower Garden Banks, Gulf of Mexico, from 1997–2003. Gulf of Mexico Science: 107-114. Wallace CC (1985) Reproduction, recruitment and fragementation in nine sympatric species of the coral genus Acropora. Mar Biol 88: 217-233. Willis BL, Babcock RC, Harrison PL, Oliver JK (1985) Patterns in the mass spawning of corals on the Great Barrier Reef from 1981 to 1984. Proc 5th Int Coral Reef Congr 4: 343-348. Zakaria IJ (2004) On the growth of newly settled corals on concrete substrates in coral reefs of Pandan and Setan Islands, West Sumatera, Indonesia Institute for Polar Ecology. Ph.D. thesis, Christian-Albrechts-Universität zu Kiel, p 168.

63 Chapter 4

Determining the influence of an electrical field on the performance of Acropora transplants: Comparison of various transplantation structures

64 65 Journal of Experimental Marine Biology (submitted)

Determining the influence of an electrical field on the performance of Acropora transplants: Comparison of various transplantation structures

Sascha B.C. Romatzki

Center for Tropical Marine Ecology (ZMT), Fahrenheitstraße 6, 28359 Bremen

Abstract

The mineral accretion method is a reef rehabilitation tool that uses the principle of electrolysis to precipitate minerals from seawater in the form of limestone as settlement- and transplant substrate on conductive metal racks. Previous observations indicated that the resulting electrical field had positive side effects on corals transplanted onto the cathode-structures. This study aimed to examine the impact of the direct contact with the cathode and the electrical field on growth, survival, chlorophyll concentration and number of zooxanthellae on two transplanted Acropora species. Acropora yongei transplants showed significantly lower growth rates when the cathodic surface exceeded 1.67 A m-2 or when in direct contact with the charged metal in comparison to a lower amperage treatment or untreated control fragments. Acropora pulchra transplants demonstrated contradictional results between the experiments with a tendency to lower growth rates when electrically treated. The number of zooxanthellae and chlorophyll c concentration in treated A. yongei fragments was significantly higher than in controls. In A. pulchra zooxanthellae density showed no difference between treated and untreated fragments but higher chlorophyll c concentration in treated fragments. Survival rates of transplants for A. yongei and A. pulchra were in all experiments high but showed significantly lower survival when exposed to a high amperage. A. pulchra demonstrated to be the more sensitive species for transplantation work. Previous reports on significantly increased growth rates due to electrical stimulation could not be supported for the here examined species.

Keywords: coral transplantation; coral growth; mineral accretion technique; zooxanthellae; chlorophyll

66 Introduction

The fields of conservation biology and restoration ecology are undergoing dramatic growth (Young 2000) as the world’s coral reefs are decreasing at an accelerating rate (Precht and Dodge 2002, Wilkinson 2004). In addition to the traditional measures of protecting, managing and conserving marine habitats, such engagements as reef-restoration and rehabilitation which manipulate reef- topography directly (Jaap 2000, Fox and Pet 2001, Nonaka et al. 2003, Yeemin et al. 2006) are becoming more popular. Hereby “restoration” is an effort aiming to bring back a pre-disturbance state (Precht 1999), while “rehabilitation” is the act of partial or full replacement, or substitution of alternative qualities or characteristics, of an ecosystem (Edwards 1998). In the past, various materials such as bottles, ships and car tires have been deployed in benthic zones in an attempt to rehabilitate reefs (reviewed in Waltemath and Schirm 1995, Pickering et al. 1998). These materials turned out to be, in many cases, unsuitable due to their mobility and leaching of toxic substances. In turn, very limited settlement of marine sessile organisms as well as low survival rates of transplants was recorded. A newer approach in coral reef rehabilitation and transplantation is the mineral accretion method (Hilbertz et al. 1977, Hilbertz and Goreau 1996, van Treeck and Schuhmacher 1999) utilizing the principle of seawater electrolysis. Sea water electrolysis is characterized by a direct current (DC) between two electrodes immersed in saltwater, resulting in CaCO3 and magnesium hydroxide

Mg(OH)2 deposition on the cathode. Depending on the induced current, the ratio between the crystalline form of CaCO3 (Aragonite) and Mg(OH)2 (Brucite) will be shifted. Less current tends to result in the more stable form Aragonite (Hilbertz 1992). This substrate shows similarities to naturally occurring reef limestone, and is therefore a good settlement substrate (van Treeck and Schuhmacher 1997). Transplanted corals are cemented with the ongoing accretion into this material, making additional fixation unnecessary (Eisinger 2005). The chemical reaction of the electrolysis leads to an increase in pH-value and with that to an enhancement of mineral precipitation around the cathode from supersaturated medium (Hilbertz and Goreau 1996). Both are believed to support and facilitate access for the physiological processes of calcification and coral growth (reviewed in Sabater and Yap 2004). Goreau et al. (2004) also measured higher densities of zooxanthellae in some species, but lower chlorophyll concentrations per symbiont cell in coral transplants. Both coincided with higher coral skeletal growth rates and better developed branching morphology, which are seen as indicators of health in coral. In contrast, Sabater and Yap (2002) measured a significant increase in girth growth in the basal region of Porites cylindrica nubbins, but no increased linear growth. The mineral accretion method still has not been tested adequately as most of the previous studies are lacking real controls with the exception of Sabater and Yap (2002), which is the only paper known to the author using an experimental set up with a control. Many factors are still undetermined with regards to this method such as applicable species, ideal fragment size, structure design,

67 location for deployment and amount of electrical current. Also the observations of enhanced coral growth remain subjective as no comparison with growth data from donor colonies were made in previous studies. It also remains unclear from the inventor’s side, if the attached corals have to be in direct contact with the cathode, or if exposure to the electrical field already stimulates a positive transplant performance. The present work addresses these open questions as it sets out to test the hypothesis that increased growth and higher survival rates are due to stimulation from an electrical field in direct comparison to those without electrical stimulation and branches within the donor colonies. Further it will clarify whether a positive response by fragments varies among the transplanted species and whether the design of the electrified structure has a significant impact. It also addresses differences in survival, density of zooxanthellae and their chlorophyll concentration in the presence and absence of the applied electrical current.

Methods

Transplant response of Acropora yongei and A. pulchra was tested in three different experiments. Fragments were taken from a shallow reef terrace in 4 m depth off Lihage Island (1º45’36”N, 125º02’16”E; Fig. 1), in July 2005 for the 1st experiment and in March 2006 for the 2nd and 3rd

N

Gangga Lihage

Manado 10km

Fig. 1 Map of North Sulawesi, showing the sampling site on Lihage and the experimental site on Gangga Island.

68 experiments. All coral fragments of both species originated from an extensive, mixed coral stand approximately 30 m long and 15 m wide with 95 % live coral cover. Finger-sized fragments (6-8 cm), generally with single tips – except for A. pulchra, which sometimes had two axial polyps - were broken off using a long-handled pliers to minimize tissue necroses. They were then transported in buckets with seawater within 10 minutes to Gangga Island (1º45’35”N, 125º03’15”E; Fig. 1) and placed in baskets next to the transplantation structures. The attachment of transplants was conducted within 24 hours with plastic cable ties or epoxy, depending on the experiment. Each experiment focused on a different approach:

Experiment 1: testing the effects of different electrical currents on the growth of coral transplants

Three tunnel shaped structures (6 m length, 1.5 m height, 2 m width) were welded from construction rebar (ø 12 mm) and uncoated, standard chicken wire to form the transplantation base (Fig. 2 a). Each tunnel structure had a metallic surface area of 4.2 m2. They were anchored at a depth of 5-6 m on the sea floor 4 m apart from each other. An anode (1.5 m2 titanium mesh; 39.4 mesh m-1; approximately 0.025 m2 of surface area m-2 of mesh (data from A. Spenhoff personal communication) was deployed under the transplantation tables of each of the electrical loaded structures (from here on “cathode”). The four electrodes, two cathodes and two anodes, were connected with separate 130 m-long copper cables (each with three inner wires of cross-sectional area 2.5 mm2 each) to two land based common automobile 12 V-battery chargers (Delta-brand 60 Ah). Separate cables for each pole were used in order to minimize electrical resistance due to small diameter and length of conductor material. The connection points between the cables and the anodes were isolated with epoxy, to avoid any corrosion. Epoxy coverings are critical due to the acidic pH-environment generated around the anodes. A low-current electrical field was then induced between the cathode and the anode. Battery chargers were set to 7 A (theoretical calculated current density of ~1.67 A m-2 cathodic surface) for one structure and 10 A (~2.38 A m-2 cathodic surface) for the second to test for possible differences of fragment growth rates caused by the strength of the electrical field. Chargers were running continuously during the entire experiment time, with a daily 2-hour-break to avoid overheating. The third structure was used as a control. A measurement of the electrical output and ph-value around the electrodes was not possible and therefore a visual inspection of mineral accretion and output as displayed by the battery chargers was used as an indicator. One hundred fifty fragments of each of the two Acropora species were lodged into the chicken mesh of the structures 1.4 m above the reef substrate. In the case of the control structure, fragments were secured with plastic cable ties.

69 Fig. 2 From top to bottom a) experiment 1: Tunnel structure with chicken wire transplantation table. An anode was placed on the ground under the transplantation table b) experiment 2: Transplantation tables with different height levels, the anode is located in the middle of the triangle, note: photo shows structure before start of actual fragment transplantation, c) experiment 3: Matrix frame with alternating metal (cathode) and bamboo boards (electrical field). Notice that fragments were glued into concrete cups, which were then strapped on the bamboo boards. Each of the boards is holding 25 Acropora yongei and 25 A. pulchra fragments. Horizontally placed anodes were deployed on each side of the matrix as seen in the front of the photo.

70 Experiment 2: testing the effects of transplantation height on fragment performance with three elevation levels

Two sets of three transplantation tables, with different levels of elevation (one slightly above the sediment, one raised 0.8 m and one 1.6 m above the sediment), were welded from rebar and connected to each other in a triangular form (Fig. 2b). Different elevations were chosen in order to test the effects height has on performance. Each table set had a surface area of 3.5 m2. A 0.36 m2 Titanium mesh anode was wrapped around a PVC-pipe and vertically fixed in the triangle center formed by the three tables. A 130 m long copper cable was connected to the anode while a second was split 3 m before the table-set and connected to each table at identical point. Cables were connected to a Delta-brand 60 Ah battery charger as described in experiment 1. The second, unconnected set was used as a control. These arrangements were deployed at a depth of 7 m at a distance of 6 meters from each other. A total of 54 fragments from each of the two Acropora species was attached to both the cathode charged with 12 A (~3.43 A m-2 cathodic surface) and to the control. Equal fragment numbers were used for each of the three height levels. The daily operational time of the chargers was the same as in the previous experiment.

Experiment 3: A comparison of different attachment methods

A matrix with the dimensions of 5 m x 2 m was welded from ø 0.87 cm rebar. It consisted of five 1 m2 grid fields alternated with five open fields (Fig. 2c). Small steel tubes (length 4 cm, ø 2 cm) were welded onto each grid stack as transplantation sockets. The matrix (“cathode” from here on), with a metal surface area of 4.2 m2, was deployed at a depth of 4 m on a protected sandy bottom. Four titanium mesh pieces (2 x 0.30 m2 and 2 x 0.45 m2, “anode” from here on) were wrapped around PVC pipes and installed horizontally, on each side, at a distance of 50 cm from the cathode. One 120 m copper cable connected two sides of the matrix with a 12 V-battery charger while a second 120 m copper cable connected the four anodes to the charger. The multiple connection points guaranteed an even current distribution. The cables were isolated as described in the previous experiments and the battery charger set to 12 A (~2.86 A m-2 cathodic surface) with a daily 2-hour operational break to avoid overheating. Bamboo grid boards were laid into each of the five open fields (from here on “electrical field”). Four additional bamboo boards, for control fragments, were laid out at a distance of 4 m from the matrix (from here on “control”). A total of 700 coral fragments were assigned to the three treatments. The first consisted of coral fragments lodged into the metal tubes with direct contact to the charged metal (= cathode, each species n = 125). The second treatment consisted of coral fragments cemented into concrete cups, strapped onto the bamboo boards and exposed to the

71 electrical field (= electrical field, each species n = 125). In the third treatment the coral fragments were placed in concrete cups, on bamboo boards, outside the electric field (= control, each species n = 100).

Growth measurement

Linear growth of fragments was measured with calipers at six-week intervals over a period of eight months. Either the lock of an attached cable tie or the Epoxy base was used as the measurement reference point. The number of surviving fragments was counted on every frame in each six-week interval. If fragments with broken tips were recognized during measurement, these fragments were not counted for this measurement period. Simultaneous growth data from branches within the donor colonies were collected to compare natural growth with growth of manipulated fragments: different colored cable ties were used as markers and reference points for the measurement. The experiments were examined during a period of up to 10 months.

Treatment for zooxanthellae count and chlorophyll concentration

After three months, coral samples were collected from each table of experiment 2 and the donor colonies to measure the density of zooxanthellae and chlorophyll-a and -c concentrations. Samples were packed into dark plastic bags, immediately frozen after collection and stored in an icebox until further procedure. Tissue of each sample was removed with an airbrush gun and seawater in half-light. Tissue parts were finely ground and the blend was filled up to 40 ml with seawater. Samples of 5 ml were removed from this solution, and preserved with 1 ml Formalin (40%) for the zooxanthellae count. The remaining solution was re-frozen for later chlorophyll analysis. Zooxanthellae from the sub samples were counted in a Neubauer Counting chamber, following standard procedures. For surface estimations, the blank skeletal fragments were coated in Varnish, to close all pores, weighted, recoated in paraffin and then weighted a final time, utilizing the technique of Stimson and Kinzie (1991). This technique was repeated on objects with a known surface area to establish a calibrating diagram with weight and surface area as diagram axis. The coral’s surface area was then calculated using the resulting regression line of the diagram. From each of the preserved sample solutions, chlorophyll was extracted from sub-samples, as described in Gardella and Edmunds (1999), and measured in a Spectrophotometer for chlorophyll a and c concentrations.

72 Data analysis

All of the data were checked with the Grubbs Outlier test (Grubbs 1969) and were excluded when necessary. The effect of treatment on average growth was analyzed for each species, using a one-way-ANOVA for Experiment 1 and 3, and a two-way-ANOVA for Experiment 2 (factor 1: treatment, with two levels; factor 2: transplantation height, with three levels) at 5% significance level. A post-hoc Tukey-Kramer HSD test was conducted at 5% significance level, to reveal the location of significance difference. All growth data was transformed with log (ln) transformation, the data was then tested with the Levene Test for variance homogeneity (Glass 1966). Although, in some cases a Levene test showed a small deviation from homoscedasticity, an ANOVA is usually robust to small departures (Underwood 1997). This was also supported by a Welch-ANOVA test, proving equality of group means and allowing the further use of the ANOVA.

Table 1 Overall survival and loss of fragments at the end of the experiment time, given in percent. Survival was calculated after the observation of lost fragments.

A. yongei A. pulchra

treatment n survival loss n survival loss Exp1 10A 106 98% 1% 22 95% 0% 7A 105 98% 0% 23 100% 0% control 50 96% 12% 10 100% 0%

Exp 2 cathode high (12 A) 35 86% 0% 21 91% 0% cathode middle (12 A) 25 76% 0% 25 92% 0% cathode low (12 A) 26 54% 0% 26 42% 0% control high 20 80% 0% 20 90% 0% control middle 20 95% 0% 20 90% 0% control low 21 19% 0% 22 91% 0%

Exp 3 cathode (12 A) 250 57% 0% 250 52% 0% electrical field 250 93% 0% 250 82% 0% control 100 78% 0% 100 47% 0%

73 The effects of treatment on transplant survival were tested using a likelihood (G) test with William’s correction. All data was analysed using JMP (SAS Institute Inc., USA) and KaleidaGraph (Synergy Software, USA). Growth data from the donor colonies (n = 10) were excluded from statistical analysis, as numbers of replicates in the experiments were a 5 to 12 fold higher. They were, however, displayed in the graphs.

Results

Observation on the anodes and of accretion rates on the cathode structures

It is noted that small anodes charged with comparable high ampere values can result in gas evolution and an acidic plume around the anode, but such observations could not be made for any of the experiments. The possible effect of acidic plumes on dissolving arogonite deposition or on coral growth remains unaccounted. Accretion in experiment 1 after 9 months ranged from 4 to 24.5 mm on the 10 A treatment and from 2 to 18 mm on the 7A treatment. The smaller values were measured towards the sides of the structures. Accretion around the centered chcken wire transplantation platforms ranged from 20 to 24.5 mm. In experiment 2 the accretion ranged from 2 to 18 mm on the low table as a result of contact with the sandy bottom, 22 to 37 mm on the middle table and 16 to 34 mm on the high table, after 9 months treated with 12 A. In experiment 3 the mineral accretion on the matrix ranged from 12 to 24 mm after 7 months treated with 12 A. Due to the slow accretion rate and its hard

Table 2 Two-way ANOVA of average monthly length increase of Acropora yongei and A. pulchra in Experiment 2 after 8 month of exposure to 2 treatments (factor 1: treatment, with two levels): 1) direct contact with the cathode, 2) control (mounted on uncharged metal frames of same design) on 3 different elevation levels (factor 2: transplantation height, with three levels): 1) high, 2) middle, 3) low.

A. yongei A. pulchra

Source df MS F p df MS F p /

Total 473 545 Treatment 1 2.904 35.326 < 0.001 1 210.22 33.469 < 0.001 Height 2 0.088 1.071 0.344 2 61.399 9.776 < 0.001 Interaction 2 0.271 3.297 0.038 2 37.509 5.972 0.027

74 appearance it is most likely that the mineralization that occurred produced hard Aragonite rather than soft Brucite.

Experiment 1: the influences of different electrical currents on coral growth

Transplanted fragments of A. yongei steadily increased their monthly growth rates over time in all treatments (Fig. 3.a), while A. pulchra showed only an increase within the first three months of transplantation, before reaching a steady state of average growth (Fig. 3b). The average monthly

a)

12 11 10 9 8 7 6 5 4 3

average growth [mm/ month] 2 1 0 Apr-2006 Oct-2005 Jun-2006 Feb-2006 Aug-2005 Dec-2005

b) 10A 7A 12 control 11 10 9 8 7 6 5 4 3

average growth [mm/ month] 2 1 0 Apr-2006 Oct-2005 Jun-2006 Feb-2006 Aug-2005 Dec-2005

Fig. 3 Average monthly growth changes (± SD) over time for a) Acropora yongei (n = 266) and b) A. pulchra (n = 55) on structures powered with different electrical currents. Notice that controls for A. pulchra have been measured only during the last three periods.

75 a)

9

8 B B 7

6 A

5

4

average growth [mm/ month] 3

2

1

0 10A 7A control donor colony

b)

9

8

7 B 6

5

4 A

average growth [mm/ month] 3

2

1

0 10A 7A control donor colony

Fig. 4 Comparison of average monthly growth (± SD) for a) Acropora yongei and b) A. pulchra fragments treated with different electrical currents in Experiment 1 and growth rates in donor colonies (n = 10). Notice that the control in b) only displays data for three month. Letters in a) indicate results of a post hoc Tukey-Kramer HSD test. Letters in b) indicate significantly differences after a ANOVA testing only 10A and 7A. Different letters indicate significantly differences.

76 length increase of electrical treated transplants was the highest on the 7 A treatment with 6.5 mm ± 0.6 (mean ± SD) for A. yongei (Fig. 4a) and 5 mm ± 0.7 (mean ± SD) for A. pulchra (Fig. 4b). There was a correlation between growth and strength of an electrical current for transplanted A. yongei (ANOVA, p < 0.001, df = 2, F = 11.54). A. yongei showed significantly lower growth rates in treatment 10 A than in treatment 7 A or the control (Fig. 4a). Similarly, the growth rates of A. pulchra in the 10 A treatment were significantly lower than in the 7 A (T-test, p < 0.001, df = 1, F = 25.05; Fig. 4b). Survival of fragments in all treatments was high (Table 1). In both electric treatments, A. yongei showed a mortality of only 2 %, while transplants on the control had a mortality of 4 %. A. pulchra showed 0 % mortality on the 7 A and control structure, but a 5% mortality on the 10 A structure. Loss of fragments due to physical impacts could be noticed only on the control structure for A. pulchra (12%).

Experiment 2: the effects of transplantation at different height levels on growth, zooxanthellae number and chlorophyll

The correlation between height levels and treatments for both species was significant (Table 2). The average monthly growth rates for A. yongei transplants during the 8-month period reached a maximum of 3.3 mm month-1 ± 0.3 (mean ± SD) on the high control level (Fig. 5.a), although there was no significantly difference between heights for any treatment. A significantly higher growth was measured for control A. pulchra compared to the treated (Table 2; Fig. 5b). Growth also significantly differed between heights for A. pulchra transplants with the highest average of 4.7 mm month-1 ± 0.2 (mean ± SD) on the middle control level (Table 2; Fig. 5b). The zooxanthellae count, recorded three months after fragmentation, showed no difference in zooxanthellae abundance when comparing the different height treatments for the two species. However, A. yongei did show a significant higher cell density on the cathode compared to the controls (Table 3a; Fig. 6a). Therefore cell number in A. yongei branches in the donor colony (6.447*105 cells cm-2 ± 3.924*105; mean ± SD) was almost twice as large as in fragments on the cathode (cathode middle; 3.362*105 cells cm-2 ± 1.525*105; mean ± SD). There was no significant difference in cell density in A. pulchra (Table 3a; Fig. 7a). The two-way-ANOVA testing the height and the interaction between height and treatment showed no significant effect on chlorophyll-a concentrations in zooxanthellae for both species (Table 3b, Fig. 6b & 7b). However, zooxanthellae had significant higher concentrations of chlorophyll-a in fragments of both species on the cathodes (Table 3b, Fig. 6b & 7b). Height and treatment showed a significantly effect on chlorophyll-c in A. yongei (Table 3c; Fig. 6c) but not in A. pulchra (Table 3c; Fig. 7c). Mortality of fragments (Table 1) was the lowest on the middle position of the control structure

77 a) 9

8

7

6

5

4

average growth [mm/ month] 3

2

1

0 high middle low donor colony

b) cathode control donor colony 9

8

7

6 a 5 a B 4 A b

average growth [mm/ month] 3 A

2

1

0 high middle low donor colony

Fig. 5 Comparison of average monthly growth rates (± SD) for a) Acropora yongei and b) A. pulchra fragments on different heights in Experiment 2 and growth rates in donor colonies (n = 10). Letters in b) indicate results of a post hoc Tukey-Kramer HSD test. Different letters indicate significantly differences.

(5 %) for A. yongei. Fragments transplanted in the low position on the control structure had the highest mortality rate (81 %). A. pulchra transplanted on the low cathode had a mortality rate of 58 % but did not show a significant difference in survival rates when compared to the other positions. A comparison between fragments transplanted on the cathode and the control structure

78 a) 5 b) c) 7 10 22 22 cathode control 20 20 donor colony 6 105 18 18

5 105 16 16 A

14 14 A 4 105 a 12 12

10 10 3 105 8 8 b B zooxanthellae [cells/cm] 5 2 10 6 6 chlorophyll a [μg/10+e6 cells] chlorophyll c [μg/ 10+e6 cells] 4 4 b 1 105 2 2

0 0 0 high middle low donator colony high middle low donor colony high middle low donor colony

Fig. 6 Comparison of a) zooxanthellae densities (cells cm-1), b) chlorophyll-a (g 106 cells-1) and c) chlorophyll-c (g 106 cells-1) for Acropora yongei (n = 6) in different transplantation heights after 4 months in experiment 2. Error bars indicate standard deviation. Letters indicate results of a post hoc Tukey-Kramer HSD test. Different letters indicate significantly differences.

showed a significantly higher survival rate for A. pulchra on the control structure (Likelihood test with William’s correction: p < 0.001), while A. yongei showed no difference. A significant difference occurred in mortality for the low positions between the two treatments (Likelihood test: p < 0.01), with higher survival rates for A. yongei on the cathode and A. pulchra on the control. Experiment 3: a comparison of the effects of different attachment methods on growth

The treatment had an effect on both species after 10 months: A. yongei grew significantly faster inside the electrical field (4.8 mm month-1 ± 0.3; mean ± SD), than on the control (3.3 mm month-1

a) b) c) 8 105 22 22 cathode 20 20 control 7 105 donor colony 18 18 6 105 16 16

5 105 14 14

12 12 4 105 10 10

5

zooxanthellae [cells/cm] 3 10 8 8 chlorophyll a in [μg/10+e6 cells]

6 chlorophyll c [μg/10+e6 cells] 6 2 105 4 4 1 105 2 2

0 0 0 high middle low donator colony high middle low donor colony high middle low donor colony

Fig. 7 Comparison of a) zooxanthellae densities (cells cm-1), b) chlorophyll-a (g 106 cells-1) and c) chlorophyll-c (g 106 cells-1) for Acropora pulchra (n = 6) in different transplantation heights after 4 months in experiment 2. Error bars indicate standard deviation.

79 Table 3 Two-way ANOVA for a) zooxanthellae densities, b) chlorophyll a and c) chlorophyll c concentrations of zooxanthellae in experiment 2 after 4 months of exposure in Acropora yongei and A. pulchra to 2 treatments (factor 1: treatment, with two levels): 1) cathode, 2) control and on 3 different elevation levels (factor 2: transplantation height, with three levels): 1) high, 2) middle, 3) low.

A. yongei A. pulchra

Source df MS F p MS F p a) zooxanthellae density (cells cm-2)

total 35 35 treatment 1 1.997*1010 7.089 0.012 1 1.813*1010 2.12 0.156 height 2 3.324*109 1.196 0.316 2 7.149*109 0.84 0.443 interaction 2 6.458*1010 2.325 0.115 2 1.980*1010 2.31 0.116 error 30 2.778*1010 30 8.560*109

b) chlorophyll-a (g 106 cells-1)

total 35 35 treatment 1 303.805 9.82 0.004 1 149.572 10.75 0.003 height 2 67.556 1.18 0.130 2 14.796 1.06 0.358 interaction 2 49.324 1.59 0.220 2 14.421 1.04 0.367 error 30 30.928 30 13.908

c) chlorophyll-c (g 106 cells-1)

total 35 35 treatment 1 95.583 9.86 0.004 1 31.323 3.13 0.087 height 2 149.502 15.42 < 0.001 2 18.865 1.89 0.169 interaction 2 28.350 2.926 0.069 2 7.598 0.76 0.476 error 30 6.689 30 9.993

± 0.3; mean ± SD), or in direct contact with the cathode (2.5 mm month-1 ± 0.3; mean ± SD) (Table 4; Fig. 8a). Further, A.yongei showed a decreasing trend of growth over time in the electrical field and control boards, while fragments on the cathode showed a improving trend of growth (Fig. 9a).

A. pulchra transplants in this experiment grow significantly better in the electrical field treatment

80 (2.3 mm month-1 ± 0.2; mean ± SD) and on the cathode (2.1 mm month-1 ± 0.2; mean ± SD) than on the control (0.9 mm month-1 ± 0.2; mean ± SD) (Table 4; Fig. 8b). The average growth rates decreased in all treatments over time for this species (Fig. 9b).

Branches in the donor colonies of both species grew faster than those of the transplanted fragments in the experiment. Difference in A. pulchra was most obvious with branches in the donor growing four times faster than transplants.

Towards the end of the experiment 3, A. yongei fragments showed a significant higher mortality of 43 % on the cathode compared to the transplants of the control and in the elctrical field where

a) 9

8

7

6 B

5 C 4 A

average growth [mm/ month] 3

2

1

0 cathode electrical field control donor colony b) 9

8

7

6

5

4

average growth [mm/ month] A 3 A

2 B

1

0 cathode electrical field control donor colony

Fig. 8 Average monthly growth changes (± SE) over time for a) Acropora yongei and b) A. pulchra in direct contact to the cathode, inside an electrical field and on control boards in experiment 3. Letters indicate results of a post hoc Tukey-Kramer HSD test. Different letters indicate significantly differences.

81 Table 4 ANOVA of average monthly growth increase for Acropora yongei and A. pulchra in experiment 3 after 10 months of exposure to 3 treatments: 1) direct contact with the cathode, 2) inside the electrical field, 3) control.

A. yongei A. pulchra Source of variation df MS F p df MS F p

total 1151 743 treatment 2 5.594 61.289 < 0.001 2 2.284 38.132 < 0.001 error 1149 0.091 741 0.06c

mortality was less than 25 % (Table 1; Table 5). Mortality of A. pulchra was significantly higher on both, the cathode (48 %) and the control boards (53 %) than in the electrical field (Table 1; Table 5).

Observations on Acropora yongei and A. pulchra branches in the donor colonies

The average monthly length increase of branches within the donor colonies was 5.6 mm ± 0.2 (mean ± SD) for A. yongei (n = 10) (Fig. 5a) and 8 mm ± 0.5 (mean ± SD) for A. pulchra (n = 10) (Fig 5b). Zooxanthellae density in A. yongei was 6.45*105 cells cm-2 ± 9.612*104 (mean ± SD)

Table 5 Likelihood (G) tests with William’s correction comparing mortality rates of Acropora yongei and A. pulchra after 10 months grown on a cathode, inside an electrical field and control.

A. yongei A. pulchra

Comparison Gadj df p Gadj df p

cathode x electrical field 32.12 1 < 0.001 22.05 1 < 0.001 cathode x control 17.53 1 < 0.001 0.18 1 0.67 electrical field x control 3.37 1 0.07 20.81 1 < 0.001

82 a) 7

6

5

4

3

2 average growth [mm/ month} 1

0 Jul-2006 Oct-2006 Jun-2006 Feb-2007 Aug-2006 Nov-2006 Dec-2006 Sep-2006 May-2006

b)

7 cathode 6 electrical field control 5

4

3

2 average growth [mm/ month}

1

0 Jul-2006 Oct-2006 Jun-2006 Feb-2007 Aug-2006 Nov-2006 Dec-2006 Sep-2006 May-2006

Fig. 9 Comparison of average monthly growth (± SE) for treated a) Acropora yongei and b) A. pulchra fragments of experiment 3 with growth rates in donor colonies. Letters indicate results of a post hoc Tukey- Kramer HSD test. Different letters indicate significantly differences.

with a chlorophyll-a concentration of 10.435 g 106 cells-1 ± 0.49 (mean ± SD) and chlorophyll-c concentration of 5.548 g 106 cells-1 ± 0.646 (mean ± SD) (Fig. 6). Zooxanthellae density in A. pulchra was 4.34*105 cells cm-2 ± 6.523*105 (mean ± SD) with a chlorophyll-a concentrations of 12.265 g 106 cells-1 ± 1.207 (mean ± SD) but chlorophyll-c of 4.67 g 106 cells-1 ± 0.302 (mean ± SD) (Fig. 7).

83 Discussion

Two species of Acropora fragments, testing the effects of electrical current on growth rates and survival, demonstrated significant differences in three coral transplantation experiments. Coral transplantation is a preferred method in rehabilitation and restoration projects, but the quality of such a technique is not always assured (Edwards and Clark 1998). Although it leads to an immediate coral cover and attraction of reef inhabitants (see Ferse 2008), this technique can only be considered positive if the long term performance of fragments is positive (Clark and Edwards 1995). Consequently the first concern applies to the survival of coral transplants. The critical point, therefore, is the size in accordance to the attachment method. The present study focused on the use of branching Acropora fragments, as members of this family are known to exhibit good growth and survival rates if conditions are favorable (reviewed in Knowlton and Jackson 2001). The identically sized Acropora fragments of approximately 8 cm length collected from the same donor allowed a direct comparison with each other. In the first step these fragments were loosely attached to the electrical structures (cathodes), but gained a firm attachment through the ongoing mineral accretion within a matter of days. Loss of less than 1% of the fragments on the cathodes clearly proved a secure attachment with the mineral accretion method. Acorpora yongei and A. pulchra survival rates of more than 75 % for both species coincided with findings of Schuhmacher et al (2000) and Sabater and Yap (2002) who tested the mineral accretion technique for other coral species. Bowden-Kerby (1997), who seeded unfixed branching coral fragments on sandy bottom, counted 100 % mortality for finger-sized fragments and was just able to increase fragment survival when using a relatively big initial size of more than 30 cm. Also Smith and Hughes (1999) recorded a low overall survival of seeded coral fragments, although they monitored different reef substrate environments. In both cases sedimentation was one of the major factors for low survival rates. On the other hand, finger-sized fragments of Okubo et al. (2005), firmly attached to concrete nails and driven into the substrate, reached survival rates of up to 100 %. Firmly attached fragments on the controls showed similar high survival patterns in the present work, but such firm attachment was more time and material consuming than fast and easy attachment to cathode structures. Other factors than the attachment method were more likely to be responsible for disappointingly low survival in some experimental treatments, as discussed below. While high survival rates have been demonstrated by multiple other transplantation techniques, they often also showed a reduced growth rate amongst fragments (see e.g. Yap and Gomez 1985, Custodio III and Yap 1997). This may be partly due to lasting effects of stresses from handling, transport and transplantation technique or unsuitable transplantation sites. The present study was able to demonstrate growth differences between the experiments, but these experiments were not able to demonstrate the enhanced coral growth rates several times greater than normal as reported by the inventors of the method (see Hilbertz and Goreau 1996, Goreau et

84 al. 2004, Goreau and Hilbertz 2005). The fragments that were exposed to the weakest currents on cathodes (< 1.7 A m-2), those in the electrical field and those in the control group showed the best growth rates - comparable with those of the natural colonies. The electro-chemical process of mineral accretion is a result of a pH-increase from a shift of concentration gradients of dissolved minerals, as for example DIC (dissolved inorganic carbon), near the cathode. Coral growth on the other hand involves the uptake of DIC to build the CaCO3 skeletons. Controlled studies in aquaria proved, that corals kept in DIC-enriched seawater enhance their growth (Marubini and Thake 1999). Hence it is believed that this extra concentration of DIC is available for corals transplanted on the cathodes and so enhances transplant growth. Sabater and Yap (2002) suggested the idea that these mineral ions are concentrated a few millimeters above the cathode so that laterally growing species should be more affected than vertical growing ones which were used in present study. This hypothesis is based on their findings of extended girth growth and supported by the results of Schuhmacher et al (2000), who measured increased foothold in vertically growing species. However, the present study showed decreased growth rates with electric current outputs set to 10 and 12 A (respectively more than 1.7 A m-2) or when coral tissue was in direct contact with the cathode. As higher amperage yields a higher alkaline environment around the cathode, indicated by precipitation of settlement-inhibiting Brucite (Eisinger et al. 2005), it might outweight the positive effects of a higher DIC concentration. Accretion thickness often varied greatly within the structures. Thin accretion areas were often localised to areas where the rebars were welded together or where the material was deformed. Changes in the conductivity, resulting in an uneven electrical field and varying accretion thickness can be caused by poor quality welding points and material deformation. Hence, variation of average growth rates of fragments may have been affected by the local thickness of the accretion coatings, which was an unmonitored aspect of the electro-chemical processes. A direct correlation between variations of accretion and coral growth remains unclear. While Sabater and Yap (2002, 2004) and Eisinger (2005) noticed enhanced growth rates only during the early stages of mineral accretion, present results demonstrated a variation in growth rates linked to the amperage: coral growth increased over time when amperage was low (Fig. 3) but demonstrated a decreadsing rate, when amperage was high (Fig. 9) . However, the total average growth rates remained in almost every experiment lower than those of the donors (Fig. 4b, Fig. 5 & Fig. 8). Transplantation stress cannot be ruled out and could have caused the lower growth rates in the beginning of the experiments (Clark and Edwards 1995). An additional explanation is the morph of fragments used, as they were all comparably short with single growth tips. With ongoing time and extension of the fragments, the number of growth tips increases. Rinkevich (2000) demonstrated that fragments with multiple growth tips reached twice the total length of the single-tip fragments in the same amount of time, which might be due to nutrient circulation and energy transfer within the underlying tissue which supports Acropora apical growth (Veron 2000).

85 Yap and Molina (2003) found that transplantation of fragments of species that were both closely related and dominant in the same reef area had different responses to their new transplanted sites, despite very similar environmental conditions. Transplants of A. pulchra and A. yongei in present study demonstrated a similar behavior. The strong discrepancy of growth between transplants regardless of treatment and donor colony as seen in Fig. 4, 5 and 9, demonstrates that A. pulchra is the species more sensitive to transplantation. The number of zooxanthellae in A. pulchra showed no difference between the treatments and the donor. A. yongei transplants demonstrated significantly higher cell numbers than those in the control, although they were only half the cell number as counted in the donor colony. Chlorophyll- a concentrations in transplats on the cathodes were significantly higher for both species, than those of the controls and donors. These measurements coincided with lower growth rates for the cathodes. Goreau (2004), who conducted alike measurements on different Acropora species, found no significantly difference in cell number, but lower concentrations of chlorophyll-a coinciding with higher growth rates in stimulated fragments. Three explanations for higher chlorophyll-a concentration versus lower zooxanthellae number were reviewed in Jones (1997a): a) an increased nutrient availability, through decreased competition of zooxanthellae, b) breakdown products of chlorophyll interfere with absorption peaks during measurement, c) greater loss of zooxanthellae from the apical tissue, leaving dark-adapted behind. It remains unclear whether the higher chlorophyll concentration is due to one or a combination of the above factors, or an interaction between these factors and the electrical field. If a reduction in zooxanthellae density is an indicator of physiological stress (Jones 1997b) the above results would imply that transplants are more stressed than donors and treated fragments are more stressed than untreated. In retrospect of the 2nd and 3rd experiments with their low survival and growth rates, the following things should be mentioned: Fragments on the cathodes in the 2nd experiment were exposed in the beginning of the experiment to increased predation by coralivorous fishes and numerous dislodged pieces of fragments were found spread out around the structure (personal observation). Although it was only noticed during the first two months of the experiment - and damaged fragments were exchanged and not counted during that measure period – a subsequent but unnoticed fish predation leading do a decreased growth could not be ruled out, as growth measurements were conducted in 6-weeks-intervalls, giving tissue enough time to recover. The transplantation technique used for the control boards in the 3rd experiment demonstrated in a neighboring transplantation project 100 m apart from the electrical structure, impressively high growth rates for the same species located in a identical depth (personal observation). Therefore the placement of boards between two extensive Euphyllia ancora colonies (10 x 10 x 3 m) must be considered as a possible explanation for the poor performance. E. ancora is described as a very aggressive coral, protecting its space with sweeper tentacles (Lewbart and Lewbart 2006). However, as there was no direct contact between controls and E. ancora, a secretion of some unknown repellents analogue to terpenoid compounds,

86 as used by soft corals (Maida et al. 1995), could have added to the poor performance. Fragments in the 2nd experiment on the low ground level tables were more exposed to sedimentation and sedimentation is known as a stress factor causing decreased coral growth rates (Cruz-Pinon et al. 2003, Crabbe and Smith 2005).

Conclusion The mineral accretion technique has obvious advantages. The flexibility in design allows a customized approach for many purposes. Its ease of installation tolerates limited manpower. Coral transplants are easy to secure and attach with hold constantly improving with the increase of mineral accretion. The extremely high survival rates of coral transplants as demonstrated especially in experiment 1 are convincing for a successful transplantation project. On the other hand it also has some critical points that have to be considered when planning the use of this technique: a) caution has to be taken when choosing the target species, as not all species are receptive to electrical stimulation, b) an increased electrical current over ~1.67 A m-2 cathodic surface does not seem suitable for maximum growth performance for the examined species, c) direct contact to the cathode diminishes, rather than supports performance, d) bigger sized fragments of branching corals with multiple growth tips seem to be better suited for transplantation than fragments of branching corals with only a single growth tip as used here.

Acknowledgement

I would like to thank G. Davi and his staff from the Gangga Island Resort for all logistical support and the generous hospitality, C. Shwaiko, W. Hilbertz and team for technical advice and provided anode material, S. Ferse and S. Schmidt-Roach for their help in the field, J.-H. Steffen, M. Schmidt, and A. Kunzmann for advices on the text and three anonymous reviewers for very helpful comments on the manuscript. This work was supported by a DAAD PhD scholarship.

References

Bowden-Kerby A (1997) Coral transplantation in sheltered habitats using unattached fragments and cultured colonies. Proc 8th Int Coral Reef Symp 2: 2063-2068. Clark S, Edwards AJ (1995) Coral transplantation as an aid to reef rehabilitation: evaluation of a case study in the Maldive Islands. Coral Reefs 14: 201-213. Crabbe MJC, Smith DJ (2005) Sediment impacts on growth rates of Acropora and Porites corals from fringing reefs of Sulawesi, Indonesia. Coral Reefs 24: 437-441. Cruz-Pinon G, Carricart-Ganivert JP, Espinoza-Avalos J (2003) Monthly skeletal extension rates of the hermatypic corals Montastrea annularis and Montastrea faveolata: biological and

87 environmental controls. Mar Biol 143: 491-500. Custodio III HM, Yap HT (1997) Skeletal extension rates of Porites cylindrica and Porites (Synarea) rus after transplantation to two depth. Coral Reefs 1997: 267-268. Edwards AJ (1998) Rehabilitation of coastal ecosystems. Mar Pol Bull 37: 371-372. Edwards AJ, Clark S (1998) Coral transplantation: a useful management tool or misguided meddling? Mar Pol Bull 37: 474-487. Eisinger M (2005) Beiträge zu ökologischen und ökonomischen Aspekten der Korallentransplantation auf elektrochemisch erzeugte Substrate als Methode zur Rehabilitation degradierter Korallenriffe. PhD-Thesis. University of Duisburg-Essen, Essen, pp 166. http:// duepublico.uni-duisburg-essen.de/servlets/DocumentServlet?id=1 1829 Eisinger M, van Treeck P, Paster M, Schuhmacher H (2005) Künstliche Riffe für Korallen: Auf diese Gitter können sie bauen. Biol Unserer Zeit 35: 192-198. Ferse SCA (2008) Artificial reef structures and coral transplantation fish community responses and effects on coral recruitment in North Sulawesi/ Indonesia. PhD-Thesis. University of Bremen, Bremen, pp 189. Fox HE, Pet JS (2001) Pilot study suggests viable options for reef restoration in . Coral Reefs 20: 219-220. Gardella DJ, Edmunds PJ (1999) The oxygen microenvironment adjacent to the tissue of the scleractinian Dichocoenia stokesii and its effects on symbiont metabolism. Mar Biol 135: 289-295. Glass GV (1966) Testing Homogeneity of Variances. American Edu Res J 3: 187-190. Goreau TJ, Cervino JM, Pollina R (2004) Increased Zooxanthellae numbers and mitotic index in electrically stimulated corals. Symbiosis 37: 107-120. Goreau TJ, Hilbertz WH (2005) Marine ecosystem restoration: Costs and benefits for coral reefs. World Res Rev 17: 375-409. Grubbs FE (1969) Procedure for detecting outlying observations in samples. Technometrics 11: 1-21. Hilbertz W, Fletcher D, Krausse C (1977) Mineral accretion technology: applications for architecture and aquaculture. Industr Forum 8: 75-84. Hilbertz WH (1992) Solar-generated building material from seawater as a sink for Carbon. Ambio 21: 126-129. Hilbertz WH, Goreau TJ 1996. Method of enhancing the growth of aquatic organism, and structures created thereby. http://www.uspto.gov, US Patent no. 5,543,034. Jaap WC (2000) Coral reef restoration. Ecol Eng 15: 345-364. Jones RJ (1997a) Changes in zooxanthellar densities and chlorophyll concentrations in corals during and after a bleaching event. Mar Ecol Prog Ser 158: 51-59. Jones RJ (1997b) Zooxanthellae loss as a bioassay for assessing stress in corals. Mar Ecol Prog

88 Ser 149: 163-171. Knowlton N, Jackson JBC (2001) The ecology of coral reefs. M Comm Ecology. Chapter 15: 395-422. Lewbart GA, Lewbart G (2006) Invertebrate medicine. Blackwell Publishing, Ames, Iowa pp 327. Maida M, Sammarco PW, Coll JC (1995) Effects of soft corals on scleractinian coral recruitment. I: Directional allelopathy and inhibition of settlement. Mar Ecol Prog Ser 121: 191-202. Marubini F, Thake B (1999) Bicarbonate addition promotes coral growth. Limnol Oceanogr 44: 716-720. Nonaka M, Baird AH, Kamiki T, Yamamoto HH (2003) Reseeding the reefs of Okinawa with the larvae of captive-bred corals. Coral Reefs 22: 34. Okubo N, Tangiguchi H, Motokawa T (2005) Successful methods for transplanting fragments of Acropora formosa and Acropora hyacinthus. Coral Reefs 24: 333-342. Pickering H, Whitmarsh D, Jensen A (1998) Artificial reefs as a tool to aid rehabilitation of coastal ecosystems: Investigating the potential. Mar Pol Bull 37: 505-514. Precht WF (1999) Coral reef restoration: The status of the science. Reef Encounter 25: 48-49. Precht WF, Dodge RE (2002) Coral reef restoration in the next Millennium Implications for Coral Reef Management and Policy: Relevant Findings from the 9th Int Coral Reef Symp. USAID, Washington D.C. 57-58. Rinkevich B (2000) Steps towards the evaluation of coral reef restoration by using small branch fragments. Mar Biol 136: 807-812. Sabater MG, Yap HT (2002) Growth and survival of coral transplants with and without

electrochemical deposition of CaCO3. J Exp Mar Biol Ecol 272: 131-146. Sabater MG, Yap HT (2004) Long-term effects of induced mineral accretion on growth, survival and corallite properties of Porites cylindrica Dana. J Exp Mar Biol Ecol 311: 355-374. Schuhmacher H, van Treeck P, Eisinger M, Paster M (2000) Transplantation of coral fragments from ship groundings on elctrochemically formed reef structures. Proc 9th Int Coral Reef Symp 2: 983-990. Smith LD, Hughes TP (1999) An experimental assessment of survival, re-attachment and fecundity of coral fragments. J Exp Mar Biol Ecol 235: 147-164. Stimson JS, Kinzie III RA (1991) The temporal pattern and rate of release of Zooxanthellae from the reef coral Pocillopora damicornis (Linnaeus) under nitrogen-enrichment and control conditions. J Exp Mar Biol Ecol 153: 63-74. Underwood AJ (1997) Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge University Press, Cambridge pp 522. van Treeck P, Schuhmacher H (1999) Artificial reefs created by electrolysis and coral transplantation - an approach ensuring the compatibility of environmental protection and diving tourism. Estuar Coast Shelf Sci 49: 75-81.

89 van Treeck P, Schuhmacher H (1997) Initial survival of coral nubbins transplanted by a new coral transplantation technology. Mar Ecol Prog Ser 150: 287-292. Veron JEN (2000) Corals of the world. Australian institute of Marine Science, Townsville pp 398. Waltemath M, Schirm B (1995) Effects and management of artificial reefs, including experiences outside the Philippines. Artificial Reefs in the Philippines 49: 25-30. Wilkinson C (ed.) (2004) Status of coral reefs of the world: 2004. Australian Institute of Marine Science, Townsville pp 264. Yap HT, Gomez ED (1985) Growth of Acropora pulchra: III. Preliminary observations on the effects of transplantation and sediment on the growth and survival of transplants. Mar Biol 87: 203-209. Yap HT, Molina RA (2003) Comparison of coral growth and survival under enclosed, semi-natural conditions and in the field. Mar Pol Bull 46: 858-864. Yeemin T, Sutthacheep M, Pettongma R (2006) Coral reef restoration projects in Thailand. Ocean Coast Manag 49: 562-575. Young TP (2000) Restoration ecology and conservation biology. Biol Cons 92: 73-83.

90 91 Appendix

2

1

Cross section thru a Pocillopora verrucosa 3 polyp. 1 = mesentery, 2 = ova with vacuuls and pink coloured nucleus, 3 = spermary.

3

2

1

Cross section thru Seriatropora hystrix polyps. 1 = polyp outline, 2 = egg, 3 = planula larvae.

1

2

Cross section thru an Acropora yongei branch showing radial arranged egg (1) - sperm (2) - bundles.

92 93 Gemäß §6 der Promotionsordnung der Universität Bremen für die mathematischen, natur- und ingenieurwissenschaftlichen Fachbereiche vom 14. März 2007 versichere ich, dass:

1. die Arbeit ohne unerlaubte fremde Hilfe angefertigt wurde

2. keine anderen als die angegebenen Quellen und Hilfsmittel benutzt wurden

3. die den benutzten Werken wörtlich oder inhaltlich entnommenen Stellen als solche kenntlich gemacht wurden

Bremen, 24. September 2008

Sascha Romatzki

94