Proceedings 9th International Coral Reef Symposium, Bali, Indonesia 23-27 October 2000

Molecular control of development in the reef coral, Acropora millepora

E. E. Ball1, D. C. Hayward1, J. Catmull1,2, J. S. Reece-Hoyes1,2,

N. R. Hislop1,2, P. L. Harrison3 and D. J. Miller2 ABSTRACT A brief overview of the embryonic and larval development of Acropora, including some previously unpublished data, provides the background for this review of our rapidly expanding knowledge of the genes that control early development in corals, with particular emphasis on Hox and Hox-like genes. Since the Phylum Cnidaria is widely accepted to be an ancient group of organisms, genes, and motifs within genes, that are shared by corals and higher metazoans are presumably ancient. Thus, shared genes allow us to study how gene structure and function have changed with time, while genes specific to higher metazoans have, presumably, evolved more recently. Anatomically, corals have many fewer cell types than higher metazoans, but it is not clear that this apparent simplicity will be reflected at the molecular level. We have already found Acropora representatives of structural genes, housekeeping genes, nuclear receptors, Hox-like genes, Pax genes and components of the dpp signalling pathway. However, thus far there is no unequivocal evidence for the cluster of Hox genes, known as the zootype genes, that is otherwise widespread among the Metazoa. As more data become available, the Cnidaria are making an increasing contribution to our knowledge of the evolution of gene structure, function, and regulation. We here illustrate the evolutionary approach that we are taking to the characterisation of coral genes with a review of our work on the Acropora Hox-like gene, cnox2-Am.

Keywords Coral, Acropora, Cnox2, Hox nisms which had reached the indicated level of development. Introduction way of approaching this problem is to attempt to establish Phylogenies based on morphological data (e.g. Fig. 1) the presence, structure and probable function in are in agreement with those based on molecular data that cnidarians of genes that are important in controlling the the Cnidaria are primitive and can therefore be regarded early development of higher animals. Although the as an outgroup for the higher Metazoa (Adoutte et al. freshwater cnidarian, Hydra, has been the subject of 1999). For this reason the Cnidaria are particularly useful many molecular studies, developmental stages are not for understanding the evolution of gene structure and easily obtainable. In contrast, broadcast spawning function and the extent to which molecular mechanisms scleractinian corals present an opportunity to obtain large governing development are common to all animals. numbers of nearly syn-chronous developmental stages One (e.g. Harrison et al. 1984, Babcock et al. 1986, Harrison and Wallace 1990). We here first present a simplified description of the em-bryonic development of the branching reef coral, Acro-pora millepora, including some new anatomical ob-servations. Then we review knowledge of Hox-like genes in Acropora and other cnidarians, illustrating our ap-proach to analysis of their structure and function from our published work on the gene cnox2–Am. Finally, we discuss possible future techniques for studying gene function in Acropora and summarise conclusions from the work thus far.

Methods

Technical details of molecular techniques discussed in this paper are given in Hayward et al. (2001). For ana- tomical studies, embryos were fixed in 2.5% glutaral- dehyde in 0.1M sodium cacodylate buffer (pH 8.0) made Fig. 1 A traditional version of the family tree of o metazoan life (Adoutte et al. 1999), showing that the up in filtered (0.45 μm) sea water for 2-8 hours at 4 C Phylum Cnidaria, which contains the corals, separated (Harrison 1988). Embryos were then rinsed for an extended period in buffer, followed by post-fixation for 1 from the ancestors of the higher Metazoa quite early in o evolution. To the left of the tree are listed some of the hour at 4 C in 1% OsO4 in sodium cacodylate buffered morphological features probably associated with orga- seawater. This was followed by dehydration and embed-

1 Research School of Biological Sciences, Australian National University, P.O. Box 475, Canberra, ACT 2601, Australia e-mail: [email protected] Tel: +61-2-61254496; Fax: +61-2-61258294 2 Biochemistry and Molecular Biology, James Cook University, Townsville, Queensland 4811, Australia 3 School of Resource Science and Management, Southern Cross University, PO Box 157, Lismore, NSW 2480, Australia

ding in Epon 812 or LR White resins using standard High magnification videography of Acropora techniques. planulae has shown that ciliary movement along the Serial sections (2-5 μm) of material embedded in planula creates strong, local currents capable of Epon 812 were cut, floated onto microscope slides, and transporting small particles towards the oral pore dried down. As we wished to correlate stages of (Harrison unpublished). This morphology and ciliary differentiation and the appearance of new cell types with movement raises the possibility that planulae may be gene ex-pression, we experimented to find a histological supplementing the energy obtained from yolk absorption stain that would maximise our recognition of newly with nutrition derived from small particulate matter or the appearing cell types. Best results, in comparison to capture of small prey, which are carried to the oral pore toluidene blue, Harris hematoxylin, basic fuchsin or by the beating cilia as the planula swims along, aboral Alsop's (1974) stain, were obtained with the polychrome end first. The particulate matter or prey may then be stain of Van Reempts and Borgers (1975). Celestine blue digested extracellularly within the oral cavity. Thus, our staining was for two hours at 60oC, followed by anatomical studies on Acropora are consistent with counterstaining with Cason’s stain for 15 min at 60oC. previous behavioural observations of particle feeding in planulae of Caryophyllia (Tranter et al. 1982), Results and Discussion Cyphastrea (Wright 1986), and Fungia (Schwarz et al. 1999). More detailed research on feeding behaviour in Morphological development coral planulae is needed.

A schematic diagram of the embryonic and larval development of Acropora millepora is presented in Figure 2, while scanning electron micrographs of selected stages are presented in Miller and Ball (2000). Recent papers on the embryology of other species of Acropora containing light micrographs or scanning electron micrographs of developmental stages include Hayashibara et al. (1997), Gilmour (1999), containing micrographs from Heyward (1987) and Wallace (1999). As described for other species (reviewed in Harrison and Wallace 1990, Hayashibara et al. 1997), observations of living A. millepora planula larvae reveal them to be highly active swimmers. The planulae are initially spheroidal and spin slowly about their axis near the water surface as cilia develop over the epidermis and begin to beat synchronously (Harrison unpublished). With subsequent development, the larvae become pear-shaped and then elongated, and are in-creasingly active, swimming throughout the water column aboral end foremost while rotating about their longitudinal axis. Elongated, motile planulae with a well developed, semi- translucent epidermis exhibit a range of swimming behaviours including rapid, directed straight swimming, spiralling and spinning. As planulae become competent to settle, they begin to swim lower in the water column and initiate demersal swimming and benthic searching and substratum testing behaviour. Preliminary analysis of our sectioned material shows that there are marked differences between the ectoderm covering most of the planula and that lining the oral cavity (Fig. 3). The former (Fig. 3A) has a diversity of Fig. 2 Schematic summary of the embryonic cell types including nematocysts (n), cnidocytes, and development of Acropora millepora. Times shown are several types of cells of a secretory appearance (gc#1, from the time of fertilization and are typical for this gc#2), similar to those described in planulae of other species in the vicinity of Townsville, Queensland, coral species (reviewed in Harrison and Wallace 1990, Australia. Colloquial terms for some of the stages are Hayashibara et al. 1997). The oral ectoderm, in contrast, also shown. 72 hr and 96 hr planulae are all in the same appears to consist of a single type of secretory cell (Fig. orientation: or=oral end, ab=aboral end. Modified from 3B). Cilia line the oral cavity, and appear to be densely Miller and Ball (2000). packed (Fig 3B), although the apparently greater density may only be due to the limited space within which they are confined.

provide positional inform-ation along the primary axis of metazoan body plans, could serve to define an animal (the "zootype concept"). The genes of the zootype take the names of the corres-ponding genes of the fruitfly, Drosophila melanogaster, the organism in which they were first described and in which their function was deciphered. A subset of the true Hox genes forms the core of the zootype, while other genes defining position nearer the anterior and posterior ends of the embryo (i.e. empty spiracles (ems), orthodenticle (otd) and even- skipped (eve) are also in-cluded. Several Hox-like sequences have been isolated from cnidarians, (Gauchat et al. 2000, for review see Ferrier and Holland 2001) as well as orthologs of ems, otd and eve (Mokady et al. 1998, Smith et al. 1999, Miller and Miles 1993, Miller et al. unpublished). The Hox-like sequences can be placed into five groups on the basis of sequence similarity (Ferrier and Holland 2001, and references therein), but their relationships to genes from other animals are not always clear. Even within the Cni- daria the assignment of genes as orthologous can be difficult, probably reflecting the diversity of the phylum. The cnox-2 class, however, does form an unambiguous orthologous group, and representatives have been isolated from three cnidarian classes, the Anthozoa, Hydrozoa, and Scyphozoa. The predicted amino acid sequence of the Acropora cnox-2 protein, cnox2-Am, is shown in Fig. 4, aligned with two orthologs, one from the anthozoan Nematostella vectensis (sea anemone; Finnerty and Martindale 1999) Fig. 3 The ectoderm of the planula outside the oral pore and one from Hydra vulgaris (Shenk et al. 1993). As is (A) differs considerably from that inside (B). As shown often the case in comparisons of related homeodomain in (A), the ectoderm outside the pore contains a typical proteins, the homeodomain itself is highly conserved as-sortment of cnidarian cell types, including nematocysts (100% identity between the Acropora and Nematostella (n) and several types of gland cells (gc#1, gc#2), with sequences; 92% identity between the anthozoan and cilia (c) arising from the surface of the ectoderm. The Hydra sequences), while other parts of the protein are less mesogloea (m) is also apparent at the base of the similar. ectoderm. The transition between the outer ectoderm and Phylogenetic analyses reveal that cnox2-Am is most the pore ectoderm is clearly apparent and is marked by closely related to cnox-2 proteins from other cnidarians, arrows in (B). The ectoderm lining the pore is much then to the placozoan protein, Trox-2. Next most closely more uniform than that outside, apparently consisting related are the Gsx proteins of "higher" metazoans, only of ciliated gland cells. The cilia (c) are thickly followed by proteins of the Hox-3 and Hox-4 groups. packed within the oral pore. Fig. 5 shows a neighbour-joining tree illustrating these rela-tionships. Thus, it now appears that the cnidarian cnox2 genes, once thought to be Hox genes (Schummer et Hox-like genes in Acropora al. 1992, Shenk et al. 1993) are in fact more closely The Hox genes are part of a highly conserved genetic related to the more-recently recognised Gsx class of system that functions to tell cells their position along the genes. anterior/posterior axis, and therefore how they should In situ hybridization with a cnox2-Am probe reveals differentiate, in metazoan embryos (reviewed in that the mRNA is localised to two types of cells (Fig. 6). McGinnis and Krumlauf 1992). These genes have in One of these (type 1, Fig. 6C) is bipolar or multipolar, common a 180 nucleotide homeobox which codes for a with the bulk of its cytoplasm near the base of the structural feature known as the homeodomain, a 60 amino ectoderm, and basal cytoplasmic extensions that lie along acid DNA-binding portion of the protein. In triploblastic the mesogloea. The other (type 2, Fig. 6C) spans the metazoans, the Hox genes are clustered in the genome ectoderm, and has its nucleus located midway between and the order of the genes within the Hox cluster is the mesogloea and the surface of the ectoderm. These evolutionarily conserved, as is the order of their two cell types correspond in morphology to neurons and expression along the anterior/posterior axis of the sensory cells, respectively. cnox2-Am expression is spa- developing organism. The evolutionary conservation of tially restricted, as it is missing from the aboral end of all the Hox genes among animals led Slack et al. (1993) to embryos, although the zone free of expressing cells is suggest that possession of a subset of them, acting to reduced with age (Fig. 6A, B).

Fig. 4 The sequences in protein databases which are most similar to cnox2-Am are cnox2 proteins from other cnidiarians. Amino acids which are identical to those in A. millepora are shown as white on black, while similar amino acids are shown in white on gray. The homeodomain, which is the DNA-binding portion of the protein, is typically highly conserved. Abbreviations: A. millepora=Acropora millepora, N. vectensis=Nematostella vectensis, H. vul- garis=Hydra vulgaris

interactions appear to be evolutionarily conserved, since in vertebrates related genes are expressed in a similar fashion along the dorso-ventral axis of the developing neural tube (Cornell and Von Ohlen 2000). The existence in cnidarians of clear orthologs of the vnd and msx genes (Schummer et al. 1992, Grens et al. 1996, Miller un- published), together with the observed spatially restricted expression of cnox2-Am in neural cells, leads us to speculate that the origin of this neural patterning system predates the diploblast/triploblast split (Valentine et al. 1999). However, the spatial restriction of the Gsx ortho- logs in the nervous systems of higher metazoans is in the dorsal-ventral axis, whereas cnox2-Am expression is restricted in the oral/aboral axis of Acropora embryos. It will, therefore, be of particular interest to see whether other genes with axially restricted zones of expression in higher metazoans are consistent with the assumed correspondence of the cnidarian oral/aboral axis to the anterior/posterior axis of higher organisms. What can we conclude about the occurrence of a Hox Fig. 5 Phylogenetic trees examine the relationships gene cluster in Acropora? There is currently no between related genes. This neighbour-joining tree of published evidence for the existence of a cnidarian Hox homeodomain sequences, constructed using PAUP cluster, although a Hox-like and an eve-like gene are (Swof-ford, 2000) indicates that cnox2-Am is most physically linked on the chromosomes of both Acropora closely related firstly to cnox-2 proteins from other (Miller and Miles 1993) and the sea anemone, cnidarians, then to Gsx genes from a diversity of higher Nematostella (Finnerty and Martindale 1999), as is also metazoans, and then, more distantly, to the true Hox the case in vertebrates (e.g. Faiella et al. 1991, Scott genes. Numbers against branches indicate the percentage 1992). In Amphioxus, the Gsx gene is part of the of 10,000 bootstrap replicates supporting the topology ParaHox cluster, a presumed sister group to the true Hox shown. cluster. Both the ParaHox and Hox clusters are presumed to have arisen by duplication of an ancestral ProtoHox In higher metazoans, Gsx orthologs are expressed in cluster (Brooke et al. 1998). The presence of clear the nervous system. In Drosophila, interactions between orthologs of Hox and ParaHox genes in Cnidaria would the genes ind (the Drosophila Gsx ortholog), msx and vnd lend support to the idea that the duplication event (two other homeobox genes), result in spatially restricted occurred before the Cnidaria diverged from the rest of the domains of expression which specify three adjacent co- Metazoa (Finnerty and Martindale 1999, Ferrier and lumns of neuroblasts along the dorso-ventral axis. These Holland 2001). However, although the assignment of

cnox-2 genes as Gsx orthologs appears unambiguous, 2, a sensory cell, spans the ectoderm and its nucleus (n) determining the relationships between cni-darian Hox- lies midway between the mesogloea and the outer surface like genes and the true Hox genes of higher organisms on of the ectoderm. (D) further examples of staining cells in the basis of sequence is not straightforward, and it the ectoderm. remains a distinct possibility that the cluster had not In the future we envisage that more direct tests of arisen at the time that the cnidarians separated from the function will become possible. One approach that has precursors of higher metazoans. Clear orthologs of ems, already been successful is to test whether a coral gene can otd and eve, genes which specify anterior and posterior substitute for its ortholog in Drosophila (Hayward et al. position during the development of higher animals, do in prep, Plaza et al. in prep). Another approach, which appear to be present in Cnidaria, however (Mokady et al. has yet to be tested, is whether double stranded RNA 1998, Smith et al. 1999, Miller et al. unpublished; Miles inter-ference (RNAi) can be used to selectively inhibit and Miller 1992, Miller and Miles 1993, Finnerty and gene expression during coral development. An initiative Martindale 1999). currently under way is the development of normalised cDNA libraries for use with microarray technology. This approach will enable us to determine changes in gene expression which occur at key developmental stages, and to investigate the molecular bases of larval settlement, calcification and responses to stress. Although work on coral genes is still in its infancy, a few generalisations are already clear. First, many gene families are older than we anticipated. For example, an ortholog of the decapentaplegic gene, which plays a role in determining the dorsal/ventral axis in arthropods and vertebrates, is present in Acropora (Hayward et al. in prep), as are genes encoding its receptor and smad pro- teins, intracellular mediators of the dpp signal (Samuel et al. 2001). Another example is provided by the nuclear receptor superfamily. We have found ten different nuclear receptor genes in Acropora, (Grasso et al. in press) most, if not all of which, fall into one of six phylogenetically distinct classes (Nuclear Receptor Nomenclature Com- mittee, 1999), indicating that this class is ancestral, and had undergone considerable divergence prior to the diploblast/triploblast split. Similarly, the presence of Acropora Pax genes which are clearly related to those from higher organisms indicates that this family too had undergone a similar divergence (Catmull et al. 1998, Miller et al. 2000). Thus, it appears that many develop- mentally important genes had already diversified before the ancestral cnidarian separated from the evolutionary line leading to the higher Metazoa. Given the hundreds of millions of years since this separation one might expect that certain gene families could have diversified more widely in cnidarians than in the rest of the metazoans. We have yet to find such a gene family, although the characterisation of the Acropora genome is still in its early stages. There is, however, evidence that individual genes, such as nanos (Mochizuki et al. 2000), are present in duplicate in Hydra and corals but not, for example, in Fig. 6 In situ hybridization reveals the cell types in Drosophila. Whether this is due to selective duplication which the cnox2-Am mRNA is localised. In both (A), or selective loss is unclear. Our failure to find a more which shows an embryo as it begins to elongate following diverse cnidarian gene family could be due to our search closing of the gastral pore, and (B), which shows a strategy, which thus far has started from conserved, relatively mature planula, expression is localised to a developmentally important vertebrate genes. The mole- subset of cells in the ectoderm except at the aboral end cular biology of corals is just beginning, but it is already (zones between the arrowheads). (C-D) Higher magni- apparent that by using molecular techniques and by fication reveals cells of two distinct types, as seen clearly visualising gene expression patterns it is possible for us to in (C). Type 1, a neuron, is localised toward the base of learn a great deal about the changes that have occurred in the ectoderm. It is either bipolar or multipolar and has gene regulation and function in the course of animal extensions that grow laterally along the mesogloea. Type evolution.

References four gamete-spawning Acropora (Anthozoa, Sclerac- tinia). Proc. 8th Int Coral Reef Sym 2: 1231-1236 Adoutte A, Balavoine G, Lartillot N, de Rosa R (1999) Hayward DC, Catmull J, Reece-Hoyes JS, Berghammer Animal evolution. The end of the intermediate taxa. H, Dodd H, Hann SJ, Miller DJ, Ball EE (2001) Gene Trends Genet 15: 104-108 structure and larval expression of cnox2-Am from the Alsop DW (1974) Rapid single-solution polychrome coral Acropora millepora. Dev Genes Evol 211: 10- staining of semithin epoxy sections using 19, DOI 10.1007/s004270000112 polyethylene glycol 200 (PEG 200) as a stain solvent. Heyward AJ (1987) Genetic systems and hereditary Stain Technol 49: 265-272 structures of reef corals. PhD thesis, James Cook Babcock RC, Bull GD, Harrison PL, Heyward AJ, Oliver University of North Queensland, Townsville: 116 pp JK, Wallace CC, Willis BL (1986) Synchronous McGinnis W, Krumlauf R (1992) Homeobox genes and spawnings of 105 scleractinian coral species on the axial patterning. Cell 68: 283-302. Great Barrier Reef. Mar Biol 90: 379-394 Miles A, Miller DJ (1992) Genomes of diploblastic orga- Brooke NM, Garcia-Fernandez J, Holland PWH (1998) nisms contain homeoboxes: sequence of eveC, an The ParaHox gene cluster is an evolutionary sister of even-skipped homolog from the cnidarian Acropora the Hox gene cluster. Nature 392: 920-922 formosa. Proc Roy Soc London B 248: 159-161 Catmull J, Hayward DC, McIntyre NE, Reece-Hoyes JS, Miller DJ, Ball EE (2000) The coral Acropora: What it Mastro R, Callaerts P, Ball EE, Miller DJ (1998) Pax- can contribute to our knowledge of metazoan evolu- 6 origins--implications from the structure of two coral tion and the evolution of developmental processes. Pax genes. Dev Genes Evol 208: 352-356 BioEssays 22: 291-296 Cornell RA, Von Ohlen T (2000) Vnd/nkx, ind/gsh, and Miller DJ, Hayward DC, Reece-Hoyes JS, Scholten I, msh/msx: conserved regulators of dorsoventral neural Catmull J, Gehring WJ, Callaerts P, Larsen JE, Ball patterning? Current Opinion in Neurobiology 10: 63- EE (2000) Pax gene diversity in the basal cnidarian 71 Acropora millepora (Cnidaria; Anthozoa): Implicati Faiella A, D'Esposito M, Rambaldi M, Acampora D, ons for the evolution of the Pax gene family. Proc Balsofiore S, Stornaiuolo A, Mallamaci A, Natl Acad Sci USA 97: 4475-4480 Migliaccio, Gulisano M, Simeone A (1991) Isolation Miller DJ, Miles A (1993) Homeobox genes and the and mapping of EVX1, a human homeobox gene zootype. Nature 365: 215-216 homo-logous to even-skipped, localized at the 5' end Mochizuki K, Sano H, Kobayashi S, Nishimiya-Fujisawa of HOX1 locus on chromosome 7. Nucleic Acids Res C, Fujisawa T (2000) Expression and evolutionary 19: 6541-6545 conservation of nanos-related genes in Hydra. Dev Ferrier DEK, Holland PWH (2001) Ancient origin of the Genes Evol. 210: 591-602 Hox gene cluster. Nature Reviews Genetics 2: 33-38 Mokady O, Dick MH, Lackschewitz, D, Schierwater B, Finnerty JR, Martindale MQ (1999) Ancient origins of Buss LW (1998) Over one-half billion years of head axial patterning genes: Hox genes and ParaHox genes conservation? Expression of an ems class gene in Hy- in the Cnidaria. Evolution & Development 1: 16-23 dractinia symbiolongicarpus (Cnidaria: Hydrozoa). Gauchat D, Mazet F, Berney C, Schummer M, Kreger S, Proc Natl Acad Sci USA 95: 3673-3678 Pawlowski J, Galliot B (2000) Evolution of Antp- Nuclear Receptors Nomenclature Committee (1999). A class genes and differential expression of Hydra Hox/ unified nomenclature system for the nuclear receptor para Hox genes in anterior patterning. Proc Natl Acad superfamily. Cell 97: 161-163 Sci USA 97: 4493-4498 Samuel G, Miller D, Saint R (2001) Conservation of a Gilmour J (1999) Experimental investigation into the DPP/BMP signaling pathway in the nonbilateral cni- effects of suspended sediment on fertilisation, larval darian Acropora millepora. Evolution and Deve- survival and settlement in a scleractinian coral. Mar lopment 3: 1-10 Biol 135: 451-462 Schummer M, Scheurlen I, Schaller C, Galliot B (1992) Grens A, Gee L, Fisher DA, Bode HR (1996) CnNK-2, HOM/HOX homeobox genes are present in hydra an NK-2 homeobox gene, has a role in patterning the (Chlorohydra viridissima) and are differentially ex- basal end of the axis in hydra. Dev Biol 180: 473-488 pressed during regeneration. EMBO J 11: 1815-1823 Harrison PL (1988) Comparative ultrastructure of Scle- Schwarz JA, Krupp DA, Weis VM (1999) Late larval ractinian spermatozoa and its evolutionary impli- development and onset of symbiosis in the sclerac- cations. PhD. Thesis. James Cook University of North tinian coral Fungia scutaria. Biol Bull 196: 70-79 Queensland: 282 pp Scott MP (1992) Vertebrate Homeobox Gene Nomen- Harrison PL, Wallace CC (1990) Reproduction, dispersal clature. Cell 71: 551-553 and recruitment of scleractinian corals. In: Dubinsky Shenk MA, Bode HR, Steele RE (1993) Expression of Z (ed) Ecosystems of the World-Coral Reefs vol 25, Cnox-2, a HOM/HOX homeobox gene in hydra, is Elsevier, Amsterdam: 133-207 correlated with axial pattern formation. Development Harrison PL, Babcock RC, Bull GD, Oliver JK, Wallace 117: 657-667 CC, Willis BL (1984) Mass spawning in tropical reef Slack JMW, Holland PWH, Graham CF (1993) The corals. Science 223: 1186-1189 zootype and the phylotypic stage. Nature 361: 490- Hayashibara T, Ohike S, Kakinuma Y (1997) Embryonic 492 and larval development and planula metamorphosis of

Smith KM, Gee L, Blitz IL, Bode HR (1999) CnOtx, a molecules and embryos: new perspectives on the member of the otx gene family, has a role in cell Cambrian explosion. Development 126: 851-859 movement in Hydra. Dev Biol 212: 392-404 Van Reempts J, Borgers M (1975) A simple poly- Swofford DL (2000) PAUP* (Phylogenetic analysis using chrome stain for conventionally fixed epon-embedded parsimony, and other methods), version 4.0b3. Si- tissues. Stain Technol 50: 19-23 nauer, Sunderland Wallace CC (1999) Staghorn corals of the world. CSIRO Tranter PRG, Nicholson DN, Kinchington D (1982) A Publ, Melbourne: 422 pp description of the spawning and post-gastrula Wright N (1986) Aspects of reproduction and planula develop-ment of the cool temperate coral Caryophillia development in the reef coral Cyphastrea ocellina. In smithii (Stokes and Broderip). J Mar Biol Assoc UK PL Jokiel, RH Richmond and RA Rogers (Editors) 62: 845-854 Coral Reef Population Biology. Hawaii Inst. Mar. Valentine JW, Jablonski D, Erwin DH (1999) Fossils, Biol. Tech. Rep. No 37 : 179-192