In the previous part, coral reproductive biology has been discussed extensively, as well as the importance of establishing sexual reproduction in aquaculture. This propagation method has become a new frontier in coral aquaculture, but is still impeded by several challenges. For some of these however, possible solutions exist. 

Establishing sexual reproduction is of vital importance for the future of coral aquaculture, as this process generates genetic diversity, which is paramount to maintaining a healthy, resilient captive population. Egg and sperm cells, also referred to as gametes, fuse together during reproduction and generate new individuals with unique characteristics (see part I). This allows a given population to withstand disturbances more effectively, and to adapt to changing conditions as only the stronger individuals survive. This long-term process is the cornerstone of evolution.  

Figure 1: A large Acropora palmata colony during the August 2008 spawning in Puerto Rico. Released egg/sperm bundles can be clearly seen. These are collected by SECORE divers with plankton nets for further rearing and research. Currently, these phenomena are rarely observed in closed aquarium systems (photograph: Ramón Villaverde).

Unfortunately, inducing captive coral spawning is highly difficult, making fragmentation the most frequently used method for propagating corals. However, for most hurdles which interfere with captive sexual reproduction, solutions exist due to extensive knowledge of coral biology and available technologies. The three major challenges at this moment may be the absence of environmental cues, a lack of proper nutrition and the usage of mechanical filtration such as protein skimmers and biofilters.

Environmental cues for reproduction

Many organisms have adapted to the erratic seasons on the planet, which are especially contrasting at latitudes outside of the Tropics of Cancer and Capricorn. Temperature, photoperiod, irradiance, precipitation and ocean tides are major abiotic factors which may fluctuate dramatically over the course of a year. For this reason, making use of the optimal season for reproduction is key to the survival of many species.

"The reason why many corals reproduce during the summer months is probably because concentrations of plankton are highest, increasing chances of offspring survival."

This has led many species to be dependent on specific seasonal cues, before they initiate reproduction. When these environmental cues are absent, organisms such as corals simply will not spawn. This phenomenon has been found to to apply to many coral species, especially to broadcasting species. The main cues or triggers for coral reproduction seem to be fluctuations in water temperature, sun- and moonlight, and possibly water movement. The involvement of hormones has also been proposed by scientists, although this remains to be elucidated¹.

- temperature

The average temperature on many coral reefs fluctuates annually between roughly 22°C (72°F) and 30°C (86°F). This forms a striking contrast with home aquaria, which are usually kept at a constant 26°C (78°F) with aquarium heaters equipped with thermostats. In these aquaria it is basically always "spring", and therefore a major reproductive trigger is absent. For several species it has been found that gametogenesis, the production of eggs and sperm, accelerates during spring. During the summer, eggs will have matured and become poised to be fertilized. This cycle is displayed by a staggering amount of plant and animal species. Temperature may well be the most critical cue for reproduction. It is of course vital not to let water temperatures exceed the 30°C (86°F) threshold, which would induce thermal bleaching and promote bacterial infections (such as with species from the genus Vibrio). The reason why many corals reproduce during the summer months is probably because concentrations of plankton are highest, which allows larvae and primary polyps to ingest ample nutrients. This of course greatly promotes their survival.

Figure 2: Swimming larvae of Favia fragum, a brooding species. These have been collected in the morning using a plankton filter mesh. The larvae have obtained zooxanthellae through the ovum from which they developed (photograph: Tim Wijgerde).

- light

Light is a factor which has been recognised by aquarists as an important cue for years, especially for fish. Lighting computers and dimmable T5 fixtures nowadays are able to mimic natural photocycles, and PL or LED lamps simulate moonlight. Unfortunately, moonlight is often applied at a constant irradiance level, which means the aquarium is always subjected to a "full moon". This likely does not stimulate corals or fish to display their natural reproductive behaviour (see water current). It is the fluctuation which makes all the difference, and this should lie around 0.01 μE/m²/s at full moon, and 0 μE/m²/s a new moon². Compare these levels with an average daytime irradiance of 200 μE/m²/s! Without a light intensity meter (PAR meter), it is difficult to mimic these exact levels, however. Where temperature should fluctuate over a course of 52 weeks, moonlight should follow a cycle of four weeks. In practical terms, this means the moonlights should increase in irradiance level for two weeks, after which they should decrease in strength for another two weeks (figure 5).

Figure 3: Scientists cover gravid colonies of Stylophora pistillata with plankton nets in the Gulf of Eilat (Gulf of Aqaba), Red Sea. The following morning the collection beakers are harvested, which often contain released larvae (photograph: Dr. Keren-or Amar).  

Due to the daily rhythm of any aquarist, the aquarium is mostly viewed during the day or evening. At night, however, there is also much to be seen; corals for example expand their tentacles to catch particles, which is a beautiful phenomenon. The release of gametes or larvae also takes place during the night, save for some exceptions such as Pocillopora meandrina in Hawaii³. The reason why corals do this is probably to reduce predation on their offspring, as less predators are active at night. This poses yet another difficulty for aquarists, as gametes and larvae are readily trapped by filtration systems. Furthermore, it is often desirable to rear offspring in separate aquaria. This problem may be solved by placing plankton nets or filtration mesh over suspected gravid colonies (figures 3 and 4). This can be done in the aquarium as well as in situ, such as in the Gulf of Eilat where researchers place plankton nets over gravid Stylophora pistillata colonies (figure 3). The collection beakers are harvested the following morning, which often contain released larvae as this is a brooding species. S. pistillata is known to release larvae mainly during springtime (April - May), but other species such as the Caribbean Favia fragum may release larvae all year round.

"The release of gametes or larvae usually takes place during the night. Corals probably do this to reduce predation on offspring."

Another possibility is reversing the photo cycle of the aquarium, allowing for gamete collection during the day. It may take a while before corals acclimatize to a new photo cycle, however there is evidence that corals are able to do this. As an example, aquarists regularly see corals expanding their tentacles during the day as a response to food addition.

Light intensity and photoperiod also fluctuate over the course of a year, especially on reefs which are located outside of the Tropics. With an advanced light computer, 24 hour cycles and changes in photoperiod/irradiance can be simulated. Some data indicate that these factors may also control coral spawning⁴.

- water current

Closely linked to moonlight are the tides, which strongly influence water current velocity. In the aquarium, water current cycles should following a two-week period. This is because on the reef, water current is strongest during new and full moon. This means that controllers should ideally increase water flow for one week, which is followed by another week of decreasing current (figure 5). Moonlight and water current also have to be properly synchronised, where water flow peaks at maximal (full moon) and minimal (new moon) moonlight irradiance. Why are moonlight and water current so important to corals?

For years it has been known that broadcasting species mainly reproduce about a week after full moon, during the night. This is possibly because at this time point, water current is lowest. A week after full moon, during the third quarter, the sun and moon partially cancel out their forces of gravity, thereby decreasing tidal forces. During this moment, called neap tide, chances of egg fertilization by sperm are highest as less dilution of gametes occurs. In 2007, scientists discovered corals are able to detect blue light which triggers production of cryptochromes⁵. These proteins are light sensitive, and are known to regulate Circadian rhythms. After increased production of these proteins, changes in the physiological state of coral polyps may occur through cellular processes, such as the onset of final gamete/larvae maturation and subsequent release.

When fluctuations in water temperature, (moon)light and water current are applied in home aquaria, the natural cycles of marine organisms are maintained. This will, in a number of cases, lead to reproduction of fish, corals and other invertebrates.

Figure 5: Recommendation for fluctuations in water temperature, moonlight intensity and water current velocity in the aquarium over a one-year period. Temperature is shown in degrees Celsius on the left y-axis, moonlight and water current are displayed in arbitrary units on the right y-axis. It is difficult to quantify these last two values without sophisticated equipment, however the essential point here is that strong fluctuations occur (graph: Tim Wijgerde).

Nutrition

Coral nutrition has been neglected by many aquarists for years; the view that fish require feeding, and that corals simply need strong lighting still persists. This view is only in part correct, as corals also need to obtain nutrients from other sources than photosynthesis. This part of coral feeding is called heterotrophy, which can be subdivided in uptake of particles and dissolved nutrients. Particles consist of various types of plankton (including bacteria and protozoa) and detritus. Dissolved nutrients consist of DOC such as amino acids, and inorganic molecules/elements such as nitrate, phosphate, calcium, magnesium, bicarbonates, potassium, strontium and iodine.

Next to maintaining optimal water parameters, food particles should be dosed in adequate levels. Artemia nauplii, rotifers (Brachionus plicatilis) and phytoplankton such as Tetraselmis suecica and Phaeodactylum tricornutum are ideal food sources for many marine invertebrates. A rough distinction between carnivorous and herbivorous corals can be made here. In general, stony corals and gorgonians seem to be mainly feeding on zooplankton, whereas soft corals have been found to possess plant-digesting enzymes^6,7.

For captive Tridacna clams it has been found that these are often male, and usually only release sperm cells in the aquarium^10,11. This is possibly due to a lack of proper nutrition with marine plankton. Similarly, for some Fungiids it has been found that smaller or weakened specimens, with less energy reserves, mostly have active male gonads. This poses as problem, as both gamete types are required simultaneously for successful fertilization and thus reproduction. Parthenogenic and simultaneous hermaphrodite corals may be an exception, although even these species may have difficulty with maintaining oogenesis when a lack of nutrients exist.

Numerous physiological processes are strongly stimulated after sufficient feeding, such as tissue buildup, calcium carbonate deposition which builds the coral skeleton, a more effective usage of light due to zooxanthellate chlorophyll buildup, and reproduction^8,9.

Filtration

A third major hurdle for captive sexual reproduction is aquarium filtration, which is not particularly plankton-friendly. The most well-known filtration system is the Berlin-method, named after its German origin. This system allows many aquarists to maintain a marine aquarium with relative ease. By means of foam fractionation, organic molecules and particles are removed from the water. This reduces bacterial breakdown of food and animal wastes, thereby inhibiting the accumulation of (in)organic molecules such as nitrate and phosphate. The major disadvantage of this method is that plankton, which includes gametes and larvae of many marine animals, is removed as well. The same problem exists when using biofilters, which are often loaded with substrates having very small pore sizes. Gametes and larvae are trapped by this, and die off as a result. Aquarium (circulation) pumps also have a destructive effect on planktonic life, although no real alternatives for the home market are currently available. This problem is so fundamental that active removal of gametes and larvae with plankton nets is required for most species (figures 3 and 4).

Alternatives for protein skimmers and biofilters have existed for years. Actually, these were even necessary before the (re)introduction of the protein skimmer in the 1980's. These methods have been in decline, as they can be technically challenging, labour intensive and sometimes do not yield the desired result. Examples of such filtration methods are the deep sand bed (DSB), including a remote version of this which is placed in an external compartment (R-DSB), and the Jaubert system. A sand bed utilizes denitrifying processes which occur under anaerobic conditions, resulting in water of high quality. Heterotrophic bacteria use nitrate (NO₃^-) as an oxidator instead of oxygen (O₂) to burn carbon sources. Phosphates are converted into bacterial biomass, and precipitated as apatite when calcium carbonate is used as a substrate.

This filter has yielded positive results for many, although long-term problems often occur. The main problem which presents itself after a year or so, is congestion of the sand bed. This results in patches which are completely anoxic, with redox values in the range of -400 mV. Under these conditions, sulfate (SO₄^2-­) instead of nitrate is reduced as this is more energetically favourable for the bacteria. This leads to the production of toxic hydrogen sulfide gas (H₂S), which often breaks through the sand bed. This sometimes has dire consequences for the marine life in the aquarium. 

A system which is similar to the DSB has existed for many years, and was developed by Prof. Jean Jaubert. This system uses a plenum, which is a layer of stagnant water situated under a screen which supports the substrate. Ideally, the system is stocked with Cirratulid worms, which consume detrital matter precipitating on top of the sand bed. After ingestion, they transfer fecal matter to the hypoxic layer. Here, bacteria can use the carbon present in the worm feces to reduce nitrate. Jaubert has obtained impressive results with this system at the Oceanographic Museum of Monaco.

Nowadays, new filtration techniques exist which have been significantly improved. One of these is the Dymico system, which is short for Dynamic Mineral Control. This is a computerized DSB, which makes use of several important innovations. With membrane pumps, active communication between the water column and the sand bed is established. This allows for a much more efficient breakdown of nitrate, as the system is no longer dependent on diffusion processes. A carbon source (acetate or ethanol) is also injected into the substrate, ensuring nitrate reduction is not carbon-limited. Finally, redox and pH probes in the sand bed allow for careful monitoring of the system. Redox values between -200 and -100 mV are important for optimal rates of denitrification, and this value can be manipulated by controlling the amount of injected carbon and flow through the sand bed. When the pH level of the sand bed is kept around 6,5 - 7, it also functions as a calcium reactor, as calcium carbonate or crushed coral is used as a substrate. All of these features make the system highly versatile.

Another method of filtration is based on the addition of a carbon source to the aquarium water, which stimulates the buildup of bacterial biomass (thereby reducing nitrate and phosphate levels) and denitrification in anaerobic pockets of live rock and substrates. This method is both cheap and effective.

Algae filters are popular as well, which may be stocked with Chaetomorpha, Caulerpa or Ulva algae. Phosphate reactors also serve well to reduce inorganic orthophosphate (PO₄^3-, also called DIP) levels, which use iron or aluminium as substrates. In the end, water changes are the most simple method of water purification, although this is often regarded as costly and time-consuming.

Whatever filtration methods are used, the essence is that these should be plankton-saving if sexual coral reproduction is to be successful without intervention, such as with plankton nets. It is clear that sexual coral reproduction is not an easy task, as it requires detailed knowledge, dedication and time. The same principle goes for breeding higher animals, such as marine fish.

Colony size and fertility

In the aquarium trade, people often speak about "mother colonies", and "fragments" or "frags". The mother or parent colony usually has become too large for the home aquaria, after which it is pruned. Often, these colonies are nowhere near their natural maximum sizes. For example, Acropora, Montipora and Euphyllia colonies can grow up to several meters in diameter.

Figure 6: Echinopora lamellosa, Montipora sp. and Seriatopora caliendrum colonies which have reached impressive sizes, growing in an artificial lagoon. Some coral species only reproduce once they have attained sizes beyond the reach of the average aquarium (photograph: Tim Wijgerde, NAUSICAA, France).

The point here is that corals simply do not start reproducing before they have attained a minimum size. This principle goes for all plants and animals, which first invest energy into growth before they reproduce (an apple tree will first grow up to a large size, before initiating fruiting). In biology, these two phases are called the vegetative and generative phases, respectively. This phenomenon also prevents many colonies from spawning in the aquarium, as they are still juveniles. Exceptions to this rule exist, such as the species Pocillopora damicornis, Favia fragum and Tubastrea coccinea. These species already reproduce at very small colony sizes. These are brooding corals, which in addition makes it less difficult to sexually reproduce them (see part I).

Additional obstacles

- "mating pairs"

As described in part I, gonochoric species exist which display separate male and female individuals. This applies to roughly 25% of all coral species, and this poses a challenge for aquarists. Species from the genera Turbinaria and Dendrophyllia (family Dendrophylliidae) for example will therefore have to be present in both sexes. For broadcasters such as Turbinaria sp., gametes will also have to be released in concert.

- settling of larvae

If inducing coral spawning and gamete/larvae collection have been successful, the next challenge is to have the offspring settle onto a substrate. If this occurs in the aquarium, which is common for brooding species, this process can be left to nature. In some cases however, it is desirable to collect gametes or larvae, which is especially true for broadcasters. The eggs can then be fertilized by mixing them with sperm, with an incubation time of 30 minutes being sufficient for at least some Acropora species¹¹. Next, the eggs can be transferred to small aquaria which are gently aerated with hoses.

Kreisel systems are also ideal, which allows for gentle stirring of larvae. Excess sperm cells will have to be removed within several hours after fertilization, as these may cause a decline in water quality and oxygen levels. Depending on species and temperature, larvae will develop from the eggs within 1-4 days. As soon as the swimming stage has been reached, the larvae will settle within several days to weeks. Brooding species release large larvae with short competency periods, allowing them to settle within a day. Larvae which develop from eggs externally often have long competency periods, allowing them to settle even after several weeks, in some cases. To survive for this long, these larvae develop mouths during their planktonic stage, allowing them to feed on particles and ingest zooxanthellae (see part I).

For larval settlement, suitable substrates are required such as ceramic tiles. These are first conditioned in the aquarium, after which a biofilm of bacteria and calcareous algae develops over the tiles. Without this vital step, the larvae will simply not settle. The biofilm may provide necessary settling cues for the larvae, such as signal molecules released by bacteria. Bacteria are known to be part of the coral holobiont, the complete assemblage of the coral and its symbiotic organisms such as zooxanthellae and bacteria, and may provide the coral with immunological defenses against other harmful microorganisms. An example of such harmful organisms are bacteria from the Vibrio genus, which have been linked to coral diseases such as white band syndrome and yellow blotch¹³.

Figure 7: Ceramic tiles with grooves form an ideal substrate for coral larvae, if they have been preconditioned in an aquarium for several weeks. This results in a biofilm with bacteria and algae, which may provide necessary cues for larvae to settle (photograph: A three month-old Acropora palmata colony, courtesy of Mitch Carl, Omaha's Henry Doorly Zoo, VS).

- zooxanthellae uptake

As described in part I, about 15% of all coral species receives zooxanthellae from its parent colony, a process which is called vertical transmission. The majority of all species, however, has to acquire these symbiotic algae early in life each generation. This uptake occurs during the larval stage, or after metamorphosis at the primary polyp stage. Several species are not yet competent to take up zooxanthellae until after metamorphosis into a primary polyp, as their larvae do not develop an oral pore during the planktonic stage. This is true for the Caribbean Elkhorn coral, Acropora palmata. Primary polyps can be inoculated by adding concentrated zooxanthellae cultures to the aquarium water, which have been previously isolated from coral tissue. They may also be able to take up algae which have been expelled by other corals, through mucus for example, a process which may also take place in the wild. Knowledge of the onset of the coral-algal symbiosis for each species is useful to promote survival chances for larvae or primary polyps.

"Although sexual coral reproduction is not an easy task, several degrees of difficulty exist. This depends on coral sexuality and the mode in which ova are fertilized."

- rearing offspring

Rearing young colonies can be a difficult task, as they are quite sensitive to overgrowth by algae and other organisms. Offspring should therefore be kept clean, by introduction of cleanup crews such as snails and (hermit) crabs, and by regular cleaning of the tiles.

Figure 8: Setup for rearing young A. palmata colonies. According to Mitch Carl, a well-known American aquarist, only 10% of the larvae settles onto a tile to metamorphose into a primary polyp. Young colonies from this species grow very slowly, and should be kept clean at all times (photograph: Mitch Carl, Omaha's Henry Doorly Zoo, USA).

Degrees of difficulty: an overview

Although sexual coral reproduction is not an easy task, several degrees of difficulty exist. This  mostly depends on coral sexuality (parthenogenesis, hermaphroditism or gonochorism) and the mode in which ova are fertilized (broadcasting or brooding). The easiest species are parthenogenic and hermaphroditic brooders. For these species, only one colony is required, as in many cases selfing may occur. Egg cells are also internally fertilized and developed, which circumvents the problems of removal by mechanical filtration. Furthermore, larvae released by brooding species often display negative buoyancy and short competency periods, which again reduces filtration-related losses. When plankton nets are placed over brooding colonies, larvae can often be collected the following morning for rearing in separate aquaria.

"The development of a closed system aimed at large-scale sexual coral reproduction would be a major step in the field of marine aquaculture."

The most difficult species are hermaphroditic and gonochoric broadcast spawners. These corals require environmental stimuli such as strong fluctuations in water temperature, (moon)light and possibly water movement. For gonochoric species, colonies of both sexes are also required. Finally, released gametes will have to be collected before they are trapped by filtration systems, at least when current aquarium systems are utilized. Table 1 provides an overview of the various categories and their degrees of difficulty.

Table 1: Various groups of coral species, categorized in degrees of sexual reproduction difficulty. Only several species per group are presented (based on Riddle, 2008 and Fadlallah, 1983).

1: classification from A: easy to G: highly difficult

2: sufficient nutrition and light required for maintaining oogenesis

3: usage of plankton nets required if protein skimmers and biofilters are applied, to prevent mechanical removal of gametes and larvae

4: environmental cues such as fluctuations in water temperature, (moon)light and water current required

5: a minimum of two colonies required

6: two sexes required

The future

The field of coral aquaculture has made significant advances over the last decades, with the advent of protein skimmers, calcium reactors, powerful pumps, high quality synthetic sea salts and a thorough knowledge of water chemistry. The husbandry of any exotic plant or animal always seems to follow three distinct phases; survival (> 1 yr), growth and finally reproduction. The first two phases have been successfully achieved for many marine species, save for non-photosynthetic corals such as Dendronephthya sp., feather stars, tunicates, many bivalves and sponges. However, when the final phase is considered, reproduction, it becomes clear that coral aquaculture it still faced with interesting challenges.

Sexual reproduction of fish, corals and other species is a time-consuming endeavour which is not within reach of most aquarists or hobbyists. Nevertheless, for the average hobbyist there still is room for much improvement. A fresh view on environmental reproductive cues, coral nutrition and the usage of alternative filtration systems may yield new results. Zoos, public aquaria, universities and research institutions have taken on a leading role in the field of coral reproduction, with projects such as SECORE (www.secore.org) and CORALZOO (www.coralzoo.org). In Japan, much has been done in this field as well. For example, coral spawnings in Japanese public aquaria are more frequent, such as at the Okinawa Churaumi Aquarium. Coral larvae are also used around Okinawa to restore entire reef sections, by pumping large numbers into plankton nets which are placed over demised parts of the reef. These larvae subsequently settle onto local substrates, after they will start building new colonies.

Although impressive and pioneering, these efforts all depend on in situ field work or open aquarium systems. The development of a closed system aimed at large-scale sexual reproduction would be a major step forward in the field of coral aquaculture. New projects are in development at several institutions, and it is quite possible that captive sexual coral reproduction will become a standard practice in the near future.

Visit www.reefbase.org and www.secore.org for more information about coral reproduction and reef restoration. For a comprehensive list of stony corals and their reproduction methods, the reader is referred to the article of Dana Riddle in Advanced Aquarist's Online Magazine (2008).

Acknowledgements

The author wishes to thank Dana Riddle for critically evaluating and improving the manuscript.

References:

Atkinson S. & M.J. Atkinson, 1992. Detection of estradiol-17ß during a mass coral spawn. Coral Reefs 11:33–35

Jokiel P.L., R.Y. Ito and P.M. Liu, 1985. Night irradiance and synchronization of lunar release of planula larvae in the reef coral Pocillopora damicornis, Marine Biology 88:167-174

Riddle D., pers. obs.

van Woesik R., 2008. Solar Insolation Drives Reproductive Schedules Of Reef Corals: A Long Road in The Search For Proximate And Ultimate Cues, 11th ICRS, Fort Lauderdale, USA

Levy, O., L. Appelbaum, W. Leggat, Y. Gothlif, D.C. Hayward, D.J. Miller and O. Hoegh-Guldberg, 2007. Light-Responsive Cryptochromes from a Simple Multicellular Animal, the Coral Acropora millepora, Science 318:467-470

Elyakova L.A., N.M. Shevchenko, S.M. Avaeva, 1981. A comparative study of carbohydrase activities in marine invertebrates, Comp. Biochem Physiol. 69B:905-908

Fabricius K.E., G. Yahel & A. Genin, 1995. Herbivory in asymbiotic soft corals, Science 268:90–93

Wijgerde T., F. Houlbrèque and C. Ferrier-Pagès, 2009. How corals feed, Coral Science (www.coralscience.org)

Houlbrèque F., C. Ferrier-Pagès, 2009. Heterotrophy in tropical scleractinian corals, Biological Reviews 84:1-17

Anonymous aquarist, 2008, personal communication

Tom Verhoeven, 2008, personal communication

Iguchi A., M. Morita, Y. Nakajima, A. Nishikawa and D. Miller, 2009. In vitro fertilization efficiency in coral Acropora digitifera, Zygote 17:225-227

Rosenberg E., O. Koren, L. Reshef, R. Efrony and I. Zilber-Rosenberg, 2007. The role of microorganisms in coral health, disease and evolution, Nature Reviews Microbiology 5:355-362

Fadlallah Y.H., 1983. Sexual reproduction, development and larval biology in scleractinian corals, Coral Reefs 2:129-150

Riddle, D., 2008. Feature Article: Coral Reproduction, Part Three: Stony Coral Sexuality, Reproduction Modes, Puberty Size, Sex Ratios and Life Spans, Advanced Aquarist's Online Magazine (www.advancedaquarist.com), 7(9)