Breeding Coral Fishes in Closed Circuit Systems

This entry was posted on Tuesday, November 1st, 1977 and is filed under Journal of MaquaCulture.

Written & Illustrated by Graham F. Cox
(Director of Waterlife ReSearch Ltd)

originally published by The Aquarist, November, 1977
(some things just haven’t changed that much!)


The following is a brief record of the endeavours of the original Seaquariums Ltd., (1968-1970) and its later daughter-successor company Waterlife Research Ltd. (1970 to date) to breed coralfishes by the more difficult route, i.e. in closed-circuit marine aquaria using continually recirculating synthetic seawater.

The claim is made that the results currently being obtained with these methods and materials (i.e. 60-70 per cent survival working with those species of coralfish which attach adhesive eggs onto a solid underwater surface) are the first recorded large-scale successes of their kind supported by photographic evidence.

The claim is not being made that this is the first successful attempt to breed and rear coralfishes on a large scale. Full honour for that achievement must be accorded to Martin Moe jnr. and his co-workers in Tampa, Florida who as early as 1974 were rearing large numbers of demersally-spawning coralfishes by using the continual flow throughput system facilitated by their access to unlimited supplies of free, clean, warm seawater and plankton, on the shores of the Gulf of Mexico.

The Past

When the author first kept coralfishes in 1961, the most frequently heard remarks were to the effect that, although the fishes were undoubtedly the most exotic members of the entire piscean family, keeping them alive in itself was virtually impossible and breeding them would never be achieved. This attitude still persists in diehard circles to this day and it is hoped that this paper will go some way toward dispelling these beliefs.

Our first (accidental) spawnings of coralfishes occurred in 1969-70, when a large pair of Amphiprion ephippium ( = Tomato or Fire Clown) repeatedly spawned on a piece of rockwork adjacent to their anemone. Despite sustained attempts with every spawning no real rearing success was achieved.

When the eggs were left with the parents, hatching always occurred in the early hours of the morning (0300-0500 hrs.) and most attempts to remove the fry from the large-(48 in. x 18 in.x 24 in.) spawning tank in a still viable condition were abortive, Attempts were then made to hatch the last two spawnings artificially along lines previously used by the author in successful breeding work with many members of the freshwater Cichlid family. That is to say that the egg-laden rock was carefully removed from the parents 2-3 hours ‘before hatching was due to occur and transferred to a separate smaller tank filled with seawater from the spawning tank. A strong current of seawater was caused to pass over by means of a wooden aerating block, and water was treated with methylene blue. No filtration of any description was used. None of the eggs hatched. The eggs from the final spawning were similarly treated but not removed from the spawning tank to the smaller .hatching tank until hatching had actually commenced. This resulted in hatchings of only seventeen (17) fry, the remaining eggs failing to hatch and eventually turning the typically whitish-grey colour of dead eggs. The seventeen fry all succeeded in becoming free-swimming but eventually starved to death owing to their refusal to eat any of the microscopic particles of prepared food offered. Shortly after this the parents died in October of 1970 probably as a result of the debilitating effects of nitrate build-up.

At this time, we had not developed a sensitive nitrate test kit due to the then-current belief that ammonia/nitrite were the only nitrogenous toxins which one needed to monitor in closed-circuit marine biosystems. As a result of that belief the artificial seawater in this breeding tank was almost two (2) years old at the time of the parent fish’s deaths, its pH and trace element content having been maintained within acceptable parameters during this period by the use of proprietary pH buffers and trace element additives manufactured by the Company and freely available to us for this maintenance work. With the benefit of hindsight, the large amounts of fresh and prepared foods which had been necessary to keep this large pair of wild Singapore clownfishes in good spawning condition throughout this long period must have resulted in a huge nitrate concentration since neither algae-harvesting nor partial water-changes were effected throughout the entire period. It is now possible to estimate, in the light of our current knowledge, that at the conclusion of these early experiments, the nitrate content of the system’s synthetic seawater would have been of the order of 300-400 parts per million (ppm = mg/litre) of nitrate. However, the level of nitrate-N on a coral reef down to a depth of 60 feet below the water surface rarely exceeds 0.000002 ppm (equal author’s extrapolation from data published in Discovery Reports, volume 4) and is normally undetectably lower than even this minute trace.

Thus it is easily seen that our early, and admittedly primitive, attempts at seawater management produced nitrate levels which were 200 x l06 times higher, or, in other words, two hundred million times higher than those ever recorded on a normal healthy coral reef, and therefore we should not have been surprised that these early batches of eggs either failed to hatch or hatched into weak deformed fry. Equally, the deaths of the parents in seawater of such colossal concentration would not surprise us today.

Rather, perhaps we should be amazed that the even survived for two years in this increasingly nitrate-enriched seawater, let alone spawning several times before the nitrate levels rose to such orders of magnitude above coral reef normality that the effects upon their normal metabolism became terminal.

Throughout the period 1971 to 1976, we steadily refined our water management techniques/materials and our animal husbandry methods were steadily improved to the extent that by 1975 we could determine the sex in vivo (i.e. without the need for dissection) and spawn within 3 months (of importation from the wild state) and hatch over 22 species of demersally-spawning coralfishes. These species mainly consisted of damselfishes (Family-Pomacentridae) clownfishes (same Family as above) and gobles.

However, except when using very large, fallow Biosystems which supported “accidental” populations of Plankton, we never raised more than a half dozen of the fry beyond the critical periodma concept which has been differently defined by various workers, e.g. Fabre-Domergue and Bi6trix (1897), Hjort (1914), Gulland (1965).

The present author’s preferred definition (at least in regard to its application to the reproduction of demersal spawning coralfishes) is that of the Norwegian researcher Johan Hjort and for this reason the concept is referred to here as Hjort’s critical period. This original concept, in greatly simplified form, may be represented as follows :-

That the massive mortality rate typical of marine fish reproduction in the wild state might be attributed to two principal causes as follows:- (i) inadequate populations of species of planktonic food organisms (ii) unfavourable currents, which carried the egg, larvae or fry into regions of seawater which, for whatever reason, were unfavourable to their (equal the eggs/larvae/fry) further normal development.

That as a result of the mecham’sms (i) and (ii) above acting singly or in conjunction, a critical period or catastrophic mortality stage can be identified in the reproductive cycle of marine fishes which chronologically correlates very closely with the period shortly after the yolk sac has been fully absorbed.

By 1974-75 the present author’s experimental work had shown that owing to the multiplicity of significant differences between the coralfishes reproducing in small closed circuit aquaria which we were working with and the Norwegian spring-spawning herring reproducing in the North Atlantic system which Hjort had worked with, we would have to modify Hjort’s ideas as to the causes of the critical period but not the validity of the critical period concept itself.

By a careful review of all the spawning/hatching/rearing laboratory notes accumulated since 1969, it was possible to show that with particular regard to the goal we had set ourselves (i.e. the successful reproduction of coralfishes in closed circuit aquaria using synthetic seawater), there were at least four critical periods which are listed below in what the author regards as a diminishing order of importance :

(i) The time when the developing larva had almost completely absorbed the yolk-sac and thus acquired sufficient strength to break away from the rock on which, as an egg, it had been laid and fertilized. Notes Below (NB): For logistical reasons we had, by this time, developed the system of removing the rock/shell containing the clutch of eggs away’ from the parents when microscopic examination of sample eggs revealed that hatching was due to commence in a short time. Consequently it must be recognised that, contrary to natural procedures, these developing larvae would receive no parental assistance in breaking open the egg-membrane when hatching was due.(ii) The 1-36 hour period immediately after hatching when a sufficient density of plankton of acceptable species and ingestable size must be present at all times if the fry are not to reach the “point of no return”- so named by Blaxter and Hempel to describe that condition of starved herring larvae kept in laboratory aquaria without food until they reached a stage whereat, even when normally-acceptable food organisms were made available, the young fishes would not (or could not) show any. feeding behaviour and proceeded to death by starvation.(iii) Any period of time during the first 20-30 days after hatching, if the nitrate content of the culture water was allowed to proceed beyond a concentration of 7-10ppm. despite otherwise normal water quality parameters. NB: Throughout the whole period of development, from the laying of the eggs-to the young fish reaching a total length of 15-20mm., normal water quality parameters (= “N.W.Q.P.” in notes below) were defined as follows.


Never allowed to reach detectable levels on a test kit capable of reading levels as low as 0.125ppm.


Never allowed to exceed pale-pink viewing vertically down the test vial of a commercial test kit = less than 2.0ppm.

Diurnal pH Variation

Fawn/brown/mauve on a sensitive colorimetric kit (= ph 7.9 – pH 8.3).


78° – 82°F (25.5°C – 27.7°C).

Specific Gravity

1.020-1.022 in the above temp range.

Oxygen Tension

90-100 per cent saturation.

Trace Elements

Maintained within natural seawater limits once-weekly addition of a commercial additive.

Algal Growth

Secured at a high level using “SEAGREEN” – commercial algal food.

Nitrite/ammonia/ammonium levels were controlled by filtration over beds of nitrifying bacteria. Nitrate and pH levels were controlled by the twin procedures of algae-harvesting (Caulerpa species) and partial water changes which were effected when, despite the addition of a proprietary brand of pH buffering solution to the seawater once weekly, the pH of the seawater did not remain in the range pH 7.9-8.3 for more than 24 hours after addition of the buffer to the system – thus showing that the alkaline reserve of the seawater had diminished to below the level at which it would be economically sensible to continue pH buffering at normal levels.

(iv) The transitional phase of adapting to and striking at a “new,” larger species of zooplankton after the fry had become accustomed to eating the “old,” smaller species. To a large extent this critical period can be reduced to extremely low limits by overlapping the availability of the smaller zooplankton with the larger species for 48/72 hours.

Finally, by late 1975 we were resigned to the prospect of having to develop plankton culturing techniques and we had temporarily abandoned our search for a prepared-food substitute for living phytoplankton and zooplankton on the grounds that our post-egg-yolk fry had never been seen to strike at any particles of the many and varied blends which we had produced or bought, and fry mortalities were massive using these materials. Never more than 3 fry were raised from a batch of 400-450 eggs by these means ( = approx. 99.5 per cent mortality).

The Present

We have now refined the equipment, test kits, synthetic sea-water and water-additives to the point where, given adequately large and matured cultures of the six species of plankton necessary ( = 3 species of phytoplankton and 3 species of zooplankton), we are able to reduce total mortality to as little as 30 per cent.

The following is an unedited report from the Company’s records of an experimental rearing conducted to evaluate the efficacy of a new type of rearing box.

Spawning No. 83 Amphiprion percula

0500 hrs, Tuesday 7.6.77. (Jubilee Day). Larvae began hatching during early hours 0630 hrs. still hatching apace and too count. Begin adding fresh Synthetica/Natura water at 0815
hrs. N.W.Q.P. Because phyto/zoo cultures are at an “unfortunate period in their cycle, will have very little Phyto 2 and 3 and zoo 2 to play with. Accordingly 0830 hrs’. transferred 10 fry to experimental rearing box in Tank 2 and 10 fry as a control in small stainless tank.

1200 hrs. Tuesday.7~6.: N.W.Q.P. 9 of fry which were placed in experimental, rearing box dead – latter now rejected for this purpose but Memo investigate its usefulness at the Artemis stage, when fry would not be so easily trapped in it.

1540 hrs. Tuesday 7/6. Remaining 10 fry in stainless tank still looking good and occasionally striking at plankton introduced at 1530 hrs. though not seen to ingest any. N.W.Q.P. but No’s climbed slightly !

2350 hrs. Tuesday 7/6. Only 9 fry found in stainless but all with fat abdomens – definitely now striking and ingesting. 20 per cent water change to reduce No’s.

0200 hrs. Wednesday 8/6. Still 9 fry~all “hopping” with usual jerky motion and occasionally resting for up to 10 minutes on tank floor before setting off in search of food. N.W.Q.P. Bed.

0630 hrs. Wednesday 8/6. Still nine good ones. Size differentiation already possible – two larger ones now not resting on bottom but swimming continuously though still with jerky motion. N.W.Q.P.

0900 hrs. Wednesday 8/6. Still nine. Two superfry no longer hopping but cruise steadily through water in total 3D spatial control of movement. Other 7 fry still hopping jerkily but now no longer resting on tank floor at all N.W.Q.P.

2355 hrs. Wednesday 8/6. Can only count 5 fry but suspect we’ve entered secretive phase as usual – anyway too tired to search N.W.Q.P. Bed.

0745 hrs. Thursday 9/6. Prolonged search – found 7 fry all of which healthy and have clearly graduated to zoo 2 at which they snap on average once each 5-8 minutes. Suspect other 3 fry bought it during critical period. Have become very secretive and difficult to find. N.W.Q.P.

2350 hrs. Thursday 9/6. 2 largest fry now 3.5 to 4.00mm. long overall and pale orange all over – both taking zoo 3 at rate of one zoo 3 each 69 minutes. These 2 superfry show no trace of melanin pigmentation whatsoever though other 5 weaklings still heavily melanin pigmented on all areas except cranial area. N.W.Q.P. Bed.

0120 hrs. Sunday 12/6. Quite definitely seven fry remaining of original 10 introduced. Largest of the 2 superfry now swimming with a faint Clown waggle. N.W.Q.P. Bed.

0220 hrs. Monday 13/6. 7 fry still remain. Thin white headband now clearly visible on 5 largest fry. 2 smallest fry still looking good but no headband yet. N.W.Q.P. Bed.

1200 hrs. Tuesday 14/6. Can only count 6 fry although since missing one is the largest specimen, I suspect he is hiding – all fry now extremely secretive and taking many minutes to locate for a head count as usual. N.W.Q.P. but No’s still mounting faster than algal removal coping. Memo 20 per cent change tomorrow AM.

2345 hrs. Wednesday 15/6. Definitely seven fry and possibly 8 – accurate count impossible because of usually secretive behavior. 5 fry now have white waist band as well as a head band and have clearly visible first doesal fin. All fry bright orange with vivid blue/white bands. N.W.Q.P. Bed.


The author did not maintain written records of the subsequent development of these young fish since they would exactly parallel the development of many, preceding spawnings for which adequate records already existed. However, their progress was recorded photographically (see accompanying prints) for the purpose of this publication and, at the time of writing they measure 20-28mm. in total length, and are fully pigmented as per the adults of the species. All being well they will join other coralfishes of our breeding for second generation reproduction experiments in 1978.

The Future Our short and longterm objectives are as follows : A. To convince our coralfish suppliers in the tropics that, just as the tropical freshwater aquarium hobby only assumed appreciable proportions when neon tetras could be sold for pennies instead of pounds, so will the equally fascinating tropical marine hobby only become available to the man in the street when coralfishes are bred on a wholesale scale at correspondingly low prices instead of being hunted for, one at a time, on frequently dangerous coral reefs. B. To either (i) successfully complete our development work on a prepared liquid suspension food which will produce at least a 50 per cent survival rate with coralfish fry, or (ii) greatly simplify our presently very sophisticated and expensive methods of culturing phytoplankton and zooplankton. The effects of achieving (i) and/or (ii) above would be to enable the more advanced home marine hobbyist to breed his own coralfishes. C. To assist in a small way in Man’s advancement along the lengthy road he has traveled towards making this planet a beautiful farm instead of preserving it as a ferocious jungle – a jungle where the fisherman/whaler is now the most vicious and feared hunter out of all Earth’s creatures.

Acknowledgements I would extend my grateful thanks to the following :- Dr. Richard Lincoln, his wife and Mr. Stephen Baynes for their invaluable assistance in supplying cultures of many species of marine plankton and thus enabling suitable food organisms to be isolated and cultured. Mr. Martyn Haywood, Livestock Manager, Waterlife Research Ltd. Mr. David Craske, Senior Aquarist with this Company during the period 1976-77 for his devoted assistance and support during the latter period of this work. To all the members of Staff of this Company past and present for their unending support and encouragement, and finally to the members of my family for their tolerance of all the missing hours.


Blaxter, J.H.S., 1965 “The feeding of herring larvae and their ecology in relation to feeding.” Calif. Coop. Oceanic Fish Invest., Rep. 10, 79-88.

Blaxter, J.H.S and Staines, M.E., 1971 “Food searching potential in marine fish larvae. Proc. Fourth European Marine Biol. Symposium, 467-485.

Bowers, A.B. and Williamson, D.I., 1951 “Food of larval and early post-larval stages of autumn spawned herring in Manx waters. Red. Mar. biol.stat. Pt. Erin 63, 17-26.

Fabre-Domergue and Bietrix, E., 1897 “Role de la v6sicule vitelline dans la nutrition larvaire des poisson marins.” C.R. Mem. Soc. Biol. (Paris) 10 Ser., Tome 5, 466-468,

Gulland, J. A., 1965 “Survival of the youngest stages of fish, and its relation to year-class strength.” Spec. Publ. ICNAF 6, 363-371.

Harvey, H.W., 1955 “The chemistry and fertility of seawater.” Cambridge University Press.

Hjort, J., 1914 “Fluctuations in the great fisheries of northern Europe in the light of biological research.” Rapport Process – Verbaux Réunion Couseil Perm. Intern. Exploration Mer. 20, 1-288.

May, R.C., 1973 “Larval mortality in marine fishes and the .critical period concept.” Proc. Inter. Symp. Scot. Mar.. Biol. Assoc., 3-19. Webmaster’s note: Amphiprion ephippium today is common known as the Red Saddleback clownfish. A. frenatus is commonly referred to as the Tomato clownfish. Webmaster’s note: No’s = nitrate


Breeder's Registry staff are all volunteers and appreciate any comments or corrections on any of our posts.

Leave a Reply

Your email address will not be published. Required fields are marked *