Introduction

In 1989 a unique method of maintaining water quality was described and patented (Jaubert 1989, 1991; Jaubert and Gattuso 1989). The method was designed primarily for use in aquariums, although other uses were conceived (Jaubert 1991). The patent belonged to Jean Jaubert, and the method eventually became known as the Jaubert method of reefkeeping. The principles of the method described in the original patent were numerous and often speculative, but several of them are notable. First, water quality is maintained biologically. The aquarium was considered a “reactor,” and the filter was a bed of calcareous sediments separated into layers by screen and sitting atop a void water space that later became known as a plenum. The concept was that oxygen gradients were created where separate processes of nitrification and denitrification were established, allowing for low nutrient levels in the “reactor.” Microbial and faunal populations were assumed to be responsible for the various biological processes, although no rigorous investigations were performed to gather data or explain in detail what were the actual processes involved or how they worked.  The Jaubert systems at Monaco, however, were maintained for long periods of time and underwent water quality analyses, and they supported a diverse reef community, including stony corals.

One of the more obvious questions to reefkeepers is, “How were calcium and alkalinity levels maintained?” The answer Jaubert gave was that acids produced by heterotrophs in the gravel bed dissolved the gravel, releasing calcium and carbon, and the dissolving gravel needed periodic replenishment (Jaubert 1989, 1991).  No measurement of the dissolution rate, however, was ever provided to my knowledge. About five years after the description of this method I began keeping a Jaubert style reef aquarium. Few of us were actually keeping Jaubert tanks at that time because they were essentially glass boxes of water with no skimmers or other filtration; Julian Sprung was also an early user of the method (Sprung 2002). As with many subjects in the aquarium hobby, debate raged about the Jaubert method and whether it could, or would, work – mostly speculation by those who had never tried it. What did occur, though, was that many of us watched our sand beds become shallower, but later realized that this decline was probably largely from compaction and settling, and not dissolution. Yet, in any sand bed, even if a uniform grain size is added, over time finer particles appear that can be considered carbonate muds, indicating that some amount of sand decomposition is occurring, though whether this occurs by dissolution or by mechanical processes has not been determined.

Over the years I kept over a dozen Jaubert systems with a plenum, but I eventually arrived at the anecdotal conclusion that the plenum space had no real function. This conclusion was later confirmed by Toonen (2005a, 2005b) and Delbeek (2002) but see Delbeek (2006) since it appears by his comment his new view contradicts his earlier reference. Purportedly, the hypoxic plenum prevented the formation of toxic hydrogen sulfide gas in truly anaerobic reducing sediments (Jaubert 1989, 1991). It was also later found in practice and in the literature that areas of anaerobic hydrogen sulfide production released gas that became oxidized quickly as it reached the aerobic surface sediments (see Hansen et al 1918, Mitterer et al 2001, for example), making it unlikely that the toxic gas would be released into the water column except in unusual cases; e.g. it has been reported that large pockets of organic material or large sand-burrowing animals could cause such a release and deaths due to this toxic gas (Shimek, pers. comm.). The only compromise I made with regard to the Jaubert method was that I used kalkwasser to replace all my evaporative water losses because I found that calcium and alkalinity tended to drop rapidly without supplementation of these components. I failed to see how the Jaubert system could maintain the calcium and carbonates required by a moderate number of calcifying species, much less a large coral population, unless extensive sand beds existed in the tank or additional connected sand tanks were located remotely from the display tank – and even then I questioned just how much dissolution occurred.

Almost concurrently, the concept of deep sand beds was beginning to take hold in the aquarium hobby, advanced mainly by the works of Shimek (2001) and others. Soon thereafter, the availability of Miracle Mud, currently known as the “Ecosystem method,” was introduced commercially.  I commented on this extensively (Lowrie and Borneman 1998a, 1998b), but nothing quite struck me as profoundly as the graph I saw produced by its creator at MACNA X in Los Angeles, which showed a linear relationship between “success” and the use of “Miracle Mud.” Nonetheless, the more “natural” filtration had its proponents, especially after the rise of increasingly powerful protein skimmers and a tendency to limit nutrients in reef aquariums (see Paletta 2000, for example). Later, and continuing to the present, many aquarists have had nitrogen and phosphorus issues in their water columns, and they often attributed these to their deep sand beds having become “nutrient sinks” and “leaching” nutrients back into the tank. This view was common among European aquarists, who claimed that the sand beds used mainly in the United States would eventually produce a deleterious effect.

This article is about several concepts and an experiment. The concepts are those articles in the scientific literature that have already obtained experimental data to explain sedimentary processes and how they agree, or conflict, with the anecdotal observations of aquarium practices. They will be found in the discussion section of this article as they relate to the experimental results I will provide. The experiment I performed was designed to address the degree to which sand dissolution can maintain the calcium and alkalinity demands of coral reef aquariums as originally proposed by Jaubert almost 20 years ago; obviously, things move slowly in the aquarium hobby. Anthony Calfo, however, recently suggested that he was able to maintain coral culture systems using sand dissolution (except for overstocked aquariums filled with rapidly growing species such as Acropora garden tanks), and also suggested that certain substrates acted better than others as buffers. My previous experience with Jaubert-style tanks and reef aquariums, in general, showed that calcium and alkalinity additions were required in order to maintain adequate levels of these components. My skeptical nature got the best of me, and I decided to perform a series of simple experiments to determine the likelihood of sand dissolution being able to maintain calcium and alkalinity. It should be noted that Jaubert changed his thinking on the dissolution concept he put forward in 1989 which was speculation and never tested (Delbeek, pers. comm.). Since setting up his fragment tanks at Monaco with J.-P. Gattuso, it quickly became apparent to him that the greater calcium demand of corals could not be met by his plenum filters, so he began using supplements (ibid.)

Methods

Light and Dark Effects

To determine the biotic and abiotic effects on the calcium and alkalinity level of a water column containing carbonate sands, a stirred and homogenized mixed 2” substrate consisting of coral skeleton, special grade reef sand (CaribSea, Inc.), oolitic sand (CaribSea, Inc.), and shell fragments was added to beakers from a four-year-old established 10-gallon reef tank with all rock, corals and fish removed to leave only the sediments and any organisms encrusted on the glass or dwelling in the sand. Detrital organic material from the sand was mixed into the water column until it was uniformly turbid. Two level tablespoons of the substrate were added to 20 250ml beakers and brought to 200ml with the turbid tank water; equivalent particulates from mid-water sampling of the water column were added as the organic material to be utilized by the heterotrophs proposed by Jaubert. The beakers were covered with an inverted Petri dish top to minimize evaporation, but the beaker’s lip under the Petri dish lid allowed some minimal air exchange between the atmosphere and the beaker water. The hypothesis was that there would be no differences in measured parameters between the control and the test beakers.

To determine whether photosynthesis measurably affected the dissolution or precipitation of calcium carbonate, 10 covered beakers were placed into a cardboard box that was taped and placed into dark room. Another 10 covered beakers were placed adjacent to the established 10-gallon tank and light fixture that was moved halfway over the tank so that the tank and beakers received equivalent light with a photoperiod of 10 hours per day. PAR values were measured at 125 microeinsteins/m²/s^-1 both in the tank at the level of the beakers’ water surface, and at the beakers’ surface with a PAR meter (LiCor, Inc.). The variation in attenuation through the tank’s water was controlled by moving the light so that more of the bulb’s irradiance reached the tank through water than reached the beakers through air. Little movement was needed because the clear Petri dish covers on the beakers provided approximately the same attenuation as the water in the tank.

Water samples were taken from the tank at the beginning of the experiment as a control to determine pH, calcium and alkalinity using a two-point calibrated pH pen (Milwaukee Instruments) and using colorimetric determination of calcium and alkalinity (Salifert, Inc.). Measurements were repeated in one week to determine changes and a t-test done to determine the significance of results between each group and the control tank. Readings were taken from five of the 10 beakers of each treatment 90 minutes after the photoperiod began, and the other five of the 10 beakers were measured 90 minutes before the photoperiod ended to account for normal diurnal variations in pH due to photosynthesis. This was repeated for the beakers in the dark condition to assure consistency of measurements, though photosynthesis was not expected to be occurring in them. Results are given as means ± standard deviation.

Effects of a CO₂-using Calcifier: Coralline Algae, Hydrolithon sp.

Five beakers were prepared as above but with the addition of a single large coralline algae crust, Hydrolithon sp., scraped and trimmed into pieces weighing exactly 60.0g each (Figure 1 below). These were added to each beaker and placed facing the light on the aquarium’s edge. A fan was set up to blow across the beakers to reduce greenhouse-heating effects. Despite the Petri dish cover, some evaporation occurred and approximately 3-5ml of double distilled water was slowly introduced into the beakers daily to maintain water levels.  PAR readings available to the coralline algae were taken using a quantum meter (Apogee Instruments) with the probe placed directly in front of the coralline crusts. PAR values are given in Table 1. The pH, carbonate alkalinity and calcium were recorded in each of the beakers, as well as the source aquarium water, at the beginning of the experiment and again after seven days, as described above. A 200ml control water sample containing only 0.45micron filtered tank water was also included. All measurements were taken at the same time of day exactly one week apart and results are given as means ± standard deviation.

Effects of a CO₃-using Calcifier: Coral, Psammacora sp.

Five beakers were prepared as above with the addition of the same coralline crusts, but with small, genetically identical clonal fragments of Psammacora sp. (Figure 2). This coral was chosen because of its rapid growth, a growth form that maintains relatively equivalent surface to volume ratios, and its ability to tolerate marginal environmental conditions such as would be present in the mostly stagnant beakers. All other conditions were as described above. The pH, carbonate alkalinity and calcium were recorded in each of the beakers and the source aquarium at the beginning of the experiment and again after seven days, as described above. A 200ml control water sample containing only 0.45micron filtered tank water was also included. All measurements were taken at the same time of day exactly one week apart and results are given as means ± standard deviation.

Results

Light and Dark Effects

The results of this experiment are shown in Table 2.

A mean decrease in pH of 0.24 ± 0.05 and 0.29 ± 0.05 units occurred in the dark condition at the early and late photoperiods, respectively. A mean increase in pH of 0.1 ± 0.07 and 0.14 ± 0.05 occurred in the lit condition at the early and late photoperiods, respectively. The increase of pH in the small water volume was greater at the end of the photoperiod than in the tank itself. Compared to the control, highly significant differences were observed in pH in all of the treatments (p<0.01), as were significant differences in the early photoperiod measurements for the lit conditions (p<0.05), despite a small mean increase. There was also a significant effect in the overall pH between lit and unlit treatment conditions (p<0.05), but not between early and late treatment conditions.

In terms of calcium, it increased in the control by 5ppt. Lit beakers gained a mean of 32 ± 5.7ppt and 27 ± 5.7ppt calcium from the initial to final measurements. Unlit beakers gained a mean of 35 ± 3.5ppt and 30 ± 53.5ppt calcium from the initial to final measurements. There was a significant increase in calcium in all treatments compared to the control (p<0.01), but no significant differences were seen between the initial and final readings.

In terms of carbonate alkalinity, there was a decrease in the control of 0.08meq/l. Lit beakers lost a mean of 0.87 ± 0.05meq/l and 0.79 ± 0.05meq/l of carbonate alkalinity from the initial to final measurements. Unlit beakers lost a mean of 0.51 ± 0.06meq/l and 0.43 ± 0.06meq/l in carbonate alkalinity from the initial to final measurements. There was a significant decrease in carbonate alkalinity in all treatments compared to the control (p<0.01), but no significant differences were seen between the initial and final readings.

After the experiment was complete, sands were dried in an oven and weighed. Despite a rough volumetric addition, their dry weights were remarkably similar, with a mean value of 42.20 ± 0.41g; 42.19 ± 0.52g for the unlit beakers and 42.20 ± 0.32g for the lit beakers. This small variation confirmed the utility of volumetric additions in subsequent experiments.

Effects of a CO₂-using Calcifier: Coralline Algae, Hydrolithon sp.

The results of this experiment are shown in Table 3.

The addition of the coralline algae Hydrolithon sp. to beakers containing sand raised the pH by 0.22 ± 0.11 units. Calcium in the water increased by 9 ± 8.22ppt while alkalinity decreased by 0.99 ± 0.09meq/l. The changes in pH and calcium between the test beakers and the control were significant (p<0.05) and the change in carbonate alkalinity was highly significant (p<0.01). There were no significant differences between the PAR values and any of the parameters measured.

Effects of a CO₃-using Calcifier: Coral, Psammacora sp.

The results of this experiment are shown in Table 4.

The addition of fragments of the scleractinian coral Psammacora sp. to beakers containing sand and 60g coralline algae raised the pH by 0.01 ± 0.07 units. Calcium in the water decreased by 119 ± 5.48ppt while alkalinity decreased by 1.24 ± 0.14meq/l. The changes in calcium and carbonate alkalinity between the test beakers and the control were highly significant (p<0.01) but the change in pH was not significant. There were no significant differences between the PAR values and any of the parameters measured.

Discussion

The results of the simple experiments above show that the addition of biologically active (“live”) sand, or sand with attached and interstitial microbes, algae and meiofauna, slightly increases the water’s pH in the presence of light, but decreases pH in darkness. Both light and dark treatments increased calcium levels by approximately 30ppt, but slightly decreased the carbonate alkalinity more pronouncedly in the presence of a regular photoperiod compared to darkness. The pH of the initial seawater was higher than normal, likely because of the use of kalkwasser for evaporative replacement in this aquarium. This may have affected both the relative dynamics of the carbonate alkalinity, compared to seawater with a lower pH, and the significance of the results.

When a calcifying alga that uses primarily carbon dioxide as a carbon source for calcification was added, the pH increased more than in the beakers with sand alone. Calcium was still slightly elevated from the initial reading after a week, but the presence of coralline algae reduced the net calcium gain from a mean of 33ppt to a mean of 9ppt. Alkalinity decreased slightly more than with sand alone, from 0.87meq/l to 0.99meq/l. It has been found that calcification in coralline algae is strongly coupled with photosynthesis and that decalcification (dissolution) often occurs in the dark with many coralline species despite net production (Chisholm 2000). This may explain the low net loss of calcium and alkalinity in the beakers over the study period.

When a scleractinian coral that uses primarily carbonate as a carbon source for calcification was added, the pH remained almost equivalent to the control, increasing only by a mean of 0.01 units. Calcium, however, declined dramatically by a mean of 119ppt. The carbonate alkalinity decreased by a mean of 1.24meq/l, more than with sand alone or with sand and algae together. There was a trend with increasing weight of the fragments to the amount of alkalinity decrease, but the results were not significant. There was no significant relationship between fragment size and calcium levels or pH, but the fragments were not intentionally of different sizes and, in fact, attempts to produce equivalent fragment sizes were intentional. It is likely that further tests using variably sized calcifying species would show a positive correlation between size or weight and decreasing levels of calcium and alkalinity. Zooxanthellate corals act as primary producers (like plants) during the day but as consumers (like animals) at night. In the beakers with coral, coralline algae and sedimentary flora and fauna, the pH behaved almost as a microcosm of the reef aquarium.

The pH of the small beakers of all treatments increased in the presence of light, which is expected because photosynthesis was being carried out by small algae in the sand, by the coralline algae and by the coral with symbiotic zooxanthellae. It is apparent that some amount of calcium is released into the water when calcium carbonate sand is present, though the presence of coralline algae decreased this amount by almost two thirds within a week, and the addition of a small coral fragment caused a large net decrease in the same time period. A small colony of the same genet of Psammacora used to produce the fragments was measured and weighed at about 7x6cm and 68g. This species produces relatively thick branches and its colonies maintain a surface to volume ratio roughly equivalent to fragments used in the treatments. If we assume equivalence, two small colonies of Psammacora lit by approximately 250µE/m²/s of light would reduce the levels of calcium by over 100ppt in a standard ten-gallon aquarium with a one-inch sand bed. Similarly, alkalinity dropped even further when coral fragments were present than with sand alone or with sand and algae, losing nearly 1/3 of the water’s total alkalinity.

It is apparent from these results that the dissolution of calcium carbonate sand is incapable of providing adequate calcium or carbonate alkalinity by dissolution to aquariums holding even very small amounts of coral and coralline algae. Thus, the implications that such processes are able to sustain reef aquariums with large numbers of scleractinian corals, as suggested by Jaubert (1989), must be called into question. The results here, however, are not surprising and are strongly supported by the literature and the practical experiences of most aquarists. It is also notable that sand or, more likely, the floral and faunal components along with carbonate sand (given the effects of light), have some significant effects on the pH, calcium and carbonate alkalinity of the overlying water column in small volumes of water. This effect contrasts with some existing literature below, but those differences could be attributed to the controlled conditions rather than more functional natural or aquarium conditions.

Carbonate Dynamics in Sand

The underlying assumption of Jaubert and the general aquarium population is that carbonate sediments dissolve, largely in part due to the “acids” released by benthic organisms or the acidic respiratory CO₂ they release. Yet, pH in carbonate sediments is not really acidic, and although many papers have measured and examined carbonate dynamics, I focus here on a particularly relevant paper (Ben-Yaakov 1973). Despite CO₂ levels being elevated up to 30x the seawater levels overlying sediments and reductive processes that tend to liberate H⁺ ions, the pH of marine carbonate sediments is fairly constant, ranging from pH 7.0-8.2. The addition of CO₂ to seawater significantly reduces pH, but this does not seem to occur in sediments. Why? In anoxic sediments, pH is buffered mainly around the reduction of sulfate and is predicted to be around pH 6.9, but sulfide precipitation results in actual levels around pH 7.9. Complete removal of sulfide brings the pH of sediment porewater to pH 8.3. Similar findings occur for silicate reactions. Data from reducing sediments show that major ions are fairly constant except for the depletion of calcium and strontium as carbonates. Because of interstitial sediment dynamics, calcite saturation is 20-30x that of seawater and the result is that calcium carbonate actually precipitates in the sediments. A summary of the paper concludes that the pH of anoxic sediments, purported to be responsible for carbonate dissolution by Jaubert and, in general, believed to be axiomatic in the aquarium hobby, is controlled by weak acids and bases as by-products of organic decomposition. Metal sulfides and calcium carbonate are precipitated, and net charges are transferred from nonprotolytic to protolytic species. The lower limit is determined by sulfide precipitation and the upper limit by calcium carbonate precipitation. The average interstitial pH of anoxic sediments is between 7.0 and 8.0, and the presence of organic acids has little effect on the pH value of water in the interstices between sediment grains, even in anoxic sediments, though sediments can reduce fulvic and humic acids by precipitation (Cunningham 1980).

Nutrient Dynamics in Sand

Many articles deal with the subject of nutrient dynamics in sediments - decomposition, oxidation, reduction, remineralization, sequestration and many other aspects. I focus here on works that summarize practical understanding of nutrient dynamics in coral reef carbonate sediments in hopes that those who feel sand beds are a sink (or source) for nutrients gain a better understanding of these processes. In particular, it is ironic that many aquarists state that sand beds become “nutrient sinks” that result in a “source” of nutrients in the tank; terms that are technically antithetic to each other. Entsch et al (1983) examined the dynamics of phosphorus and nitrogen in the sediments of coral reefs around Davies Reef, part of the Great Barrier Reef, and Suzumura et al (2002) examined phosphorus cycling in the reef sediments of Ishigaki Island, Japan. These, along with the studies cited therein, can be summarized as below.

Both teams found that surface water typically was extremely devoid of nutrients while the sediments contained an inorganic phosphorus pool (300ppm by weight in Entsch et al 1983). Earlier works had supposed that carbonate substrates strongly bound or absorbed phosphates. Their sediment analyses contained Halimeda, coral, molluscs, coralline algae, forams and other carbonate sediments, similar to those found in many aquariums. They (and others) showed no change in phosphate with water that contacted reef substrates. Entsch et al (1983) examined nitrogen and phosphorous ratios (N:P) of phytoplankton and benthic macroalgae and found the algae to contain higher nitrogen and in some cases, phosphorus, than the phytoplankton whose composition resembled those found in the water column. The majority of phosphorus was found in the inorganic form (>80%), indicating long-term concentration of phosphorus in the sediments representing a nearly uniform pool across depth gradients to 5 meters (also found to be the case at other reefs). The results of Suzumura et al (2002) were similar. The conclusion was that sediments did act as a sink for phosphate.

What about soluble phosphorus and sediments acting as a “source” or, as aquarists often state, sand beds “leaching” phosphates into the water? Including nitrogen and phosphorus, the researchers hypothesized that the large nutrient pool must be large compared to the total biomass, though plants and animals can have 10 and 100 fold higher concentrations. Phosphate solubility is complex in the sediments, and they noted the distinct layering of oxic over anoxic (reducing) sediments 5cm from the sediment/water interface. This distance varies and can be as little as a few millimeters in organically enriched fine sediments. In sediments less than 10cm deep, as would be found in most aquariums, soluble phosphorus was thought to depend mainly on the species of calcium phosphate present and pH, and also to cause nitrogen to occur mainly as ammonia (as was found to be the case). The concentration of phosphate was 30-50 times higher in the shallow interstitial water of the sediments than the water column and thus available to algae in the sediments. Entsch et al (1983) examined desorption of phosphate from sediments and found that only 1-3% was lost as soluble material, although rapid cycling but containment was found by Suzumura et al (2002). Water moving through the sediments and the subsequent washing away of organisms binding or sequestering phosphorus can increase, even indirectly, phosphorus in the water column. Yet, this increase is not significant in the overlying water in most cases. Inputs to the sediment nutrient pool are likely the accumulation of particulate material (detritus) and its accumulation into reef biota that form carbonate sediments, with biological pathways, probably benthic algae (microphytobenthic communities), retaining the phosphorus absorption onto calcium carbonate with relatively rapid adsorption and desorption resulting in low net loss of phosphate from the sediments as they are taken up during photosynthesis. These benthic algae are then used as food for grazing species. Higher levels of particulate material that could provide input to the sediments are likely minimized by the consumption of those particles by benthic species such as corals and other filter-feeding invertebrates. Furthermore, the loss of nitrogen to the overlying water of sediments is mediated by benthic species’ uptake, and perhaps by never escaping the anoxic areas. Traditional views of  the nitrification-denitrification pathway involves the production of both ammonium and nitrate, either of which may escape sediments into the water column. Thamdrup and Dalsgaard (2002), however, showed that with coupled ammonium oxidation-nitrate reduction, ammonium may be largely consumed in the anoxic zone of the sediment and reduce its escape into the overlying water.
 This will be covered in relation to aquariums in an upcoming article examining the material removed by protein skimmers.

Summary

It is apparent from the simple experiments done using sand, sand and calcifying algae, and coral fragments that the dissolution of carbonate sediments is not adequate to maintain the calcium or carbonate alkalinity of aquariums housing even very low numbers of calcifying organisms. My results conform with previous works (Shimek 2001, Toonen 2005a, 2005b), as well as existing scientific literature cited. My work contradicts Jaubert’s findings, but correlates well with the experiences of most reef aquarists, my past experiences included, in the frequent to daily needs to supplement both calcium and alkalinity to maintain seawater’s calcium and alkalinity levels for normal rates of calcification. Studies, including those cited in this article, explain why significant dissolution of carbonates does not occur because the dynamics and pH of sedimentary interstitial water is not acidic enough to allow for significant dissolution of sand, if any net change occurs at all. Similarly, the nitrogen and phosphorus dynamics of carbonate sediments explained in the studies above echo the experiences of many reef aquarists, me included, that sand beds do not, under normal conditions, become a nutrient source to tank water. The likely and most parsimonious explanation for those who do not find sand beds to be a long-term sink for nitrogen and phosphorus is probably the lack of biological mediation by sediment associated microalgae, macroalgae and herbivory. In the near future I will quickly test the relative claimed  “buffer capacity” of various substrates available in the marine aquarium trade that aquarists commonly use, in order to determine any differences in solubility under similar conditions, and by using acid to dissolve a known weight and volume of substrate as an estimate of solubility and porosity.

References

Ben-Yaakov S. 1973. pH buffering of recent anoxic marine sediments. Limnol Oceanogr 18: 86-94.

Chisholm, J.R.M. 2000. Calcification by crustose coralline algae on the northern Great Barrier Reef, Australia. Limnol Oceanogr 45: 1476-1484.

Cunnigham, R. 1980. Organic-inorganic interactions of marine humic substances from carbonate sediments: metal binding and adsorption studies. Ph.D. thesis, University of Texas, Dallas.

Delbeek J.C. 2006. Plenums: A path toward thriving tanks. Marine Fish and Reef USA 2007 Annual 9:16-22.

Delbeek J.C. 2002. A three year study on plenum activity. MACNA XIII, Baltimore, MD.

Entsch, B. K.G. Boto, R.G. Sim, and J.T. Wellington. 1983. Phosphorus and nitrogen in coral reef sediments. Limonol Oceanogr 28: 465-476.

Hansen, M.H., K. Ingvorsen and B.B. Jorgensen. 1978. Mechanisms of hydrogen sulfide release from coastal marine sediments to the atmosphere. Limnol Oceanogr 23: 68-76.

Jaubert, J. 1989. An integrated nitrifying-denitrifying biological system capable of purifying sea water in a closed circuit system. Bulletin de l'Institut Oceanographique, Monaco, no special 5: 101-106.

Jaubert, J. and J-P. Gattuso. 1989. Changements de forme provoques par la lumiere, observes, en aquarium, chez des coraux (scleractiniaires a zooxanthelles). Bulletin de l'Institut Oceanographique, Monaco, no special 5: 195-204.

Jaubert, J. 1991. System for biological purification of water containing organic materials and derivative products. United States Patent Number 4,995,980, Feb. 26, 1991.

Lowrie, J. and E. Borneman. 1998a. Demystifing mud (sediment filtration).  Part 1.  FAMA 21(6): 96-102.

Lowrie, J. and E. Borneman. 1998b. Demystifing mud (sediment filtration).  Part 2: an effective sediment. FAMA 21(7): 160-164.

Mitterer, R.M., M.J. Malone, G.A. Goodfriend, P.K. Swart, U.G. Wortmann, G.A. Logan, G.D.A. Feary, and A.C. Hine. 2001. Co-Generation of hydrogen sulfide and methane in marine carbonate sediments. Geophysics Research Letters 28: 3931-3934.

Paletta, M. 2000. The EcoSystem aquarium revisited: checking in after two years. Aquarium Fish Magazine, February issue.

Shimek, R. 2001. Sand Bed Secrets: The common-sense way to biological filtration. Marc Weiss Companies, Inc.

Sprung, J. 2002. Jaubert's method, the "Monaco System," defined and refined. Advanced Aquarist 1(9).

Suzumura, M., T. Miyajima, H. Hata, Y. Umezawa, H. Kayanne and I. Koike. 2002. Cycling of phosphorus maintains the production of microphytobenthic communities in carbonate sediments of a coral reef. Limnol Oceanogr 47: 771-781.

Thamdrup, B. and T. Dalsgaard. 2002. Production of N₂ through anaerobic ammonium oxidation coupled to nitrate reduction in marine sediments. Appl Environ Microbiol 68: 1312-1318.

Toonen R.J. and C. Wee. 2005a. An experimental comparison of sandbed and plenum-based systems. Part 1: controlled lab dosing experiments. Advanced Aquarist 4(6).

Toonen, R.J. and C. Wee. 2005b. An experimental comparison of sandbed and plenum-based systems: Part 2: live animal experiments. Advanced Aquarist 4(7).