14 min read

An Abridged Interpretation of Stephen Spotte’s Fish and Invertebrate Culture: Water Management in Closed Systems

Written by Russell McAndrews

Included in the following are several passages taken from the book as well as my interpretation of this most useful work.  Inadvertently this synopsis falls short of complete disclosure and as such, I would refer all questions to the book.


“Three types of filtration are used in closed system culturing: biological, mechanical, and chemical.  Of these, biological filtration is the most important.”

Biological filtration involves three fairly distinct stages:  mineralization, nitrification, and denitrification.  In an established filter bed, these stages are going on simultaneously.  However, while conditioning a bed it can be shown that they occur in the order given above.  Two classifications of bacteria are dominant in the filtration process.  Heterotrophic species convert animal wastes to simple compounds such as ammonia.  Autotrophic species convert ammonia to nitrite and nitrite to nitrate.

Mineralization, the first stage of the biological filtration process must be further simplified into two steps.  Ammonification is the first step and is defined as the chemical breakdown of proteins and nucleic acids into amino acids and organic nitrogenous bases.  The second step is deamination in which a portion of the products of ammonification as well as some organics are re-shuffled into inorganic compounds.

UREA        -->    AMMONIA
NH 2
O=         +   H2O        -->     CO2   +   2NH3

Nitrification is the second stage, which Spotte describes as the biological oxidation of ammonia to nitrite and nitrite to nitrate.  Nitrosomonas species will convert toxic ammonia to less toxic nitrite in the following reaction.

NH4  +  OH  +  1.502   -->  H+  +  NO2  +  2H2O

Nitrobacter species utilize the nitrite (NO2) given above to produce nontoxic nitrate (NO3).

NO2  +  .502  -->   NO3

The third stage in the biological filtration process is denitrification.  During denitrification, nitrites and nitrates are broken down and reformed as nitrous oxide or free nitrogen.  Apparently, this third stage is carried out by both heterotrophic and autotrophic species.  Further examples are given below.

4NO3+  3CH4  ==  2N2  +  3CO2  +  6H2O
2NO3+  2H+==  N2O  =  2O2  +  H2O

The nitrogen cycle is virtually the same in nature as it is in captivity.  Since the sheer mass of a natural environment cannot be readily duplicated in captivity, the culture animals are at the mercy of nitrogen cycle management.  As the manager, the hobbyist should note the roles of oxygen and carbon dioxide in the filtration process.  Several factors can be expected to influence a biological filter and therefore the nitrogen cycle.

Sudden salinity changes in either direction have been shown to have an adverse effect on bacteria.  Those species particularly affected are the nitrifiers (autotrophs).  The problem that arises is that these bacteria can die or go into shock, which temporarily causes a lag in the nitrogen cycle until they have had time to adapt.  This lag time can be lethal as toxic wastes have a habit of building up in short order.  According to Spotte freshwater and marine bacteria are the same and can be transferred from one to the other in salinity is changed gradually.  As a precautionary measure in marine systems, water to be added should be mixed and checked in a separate container for proper salinity in order to avoid shocking the filter bed.

A total of 90% of the nitrifying activity occurring in the system takes place in the filter bed.  Ninety percent of these nitrifiers are found attached to gravel and detritus within two inches of the top of the bed.  This implies that surface area is of primary importance.  The larger the surface area for bacterial attachment the greater the capacity of the filter.  Gravel in the system provides the bulk of the sites for attachment.  Within a fixed area, the tank bottom or filter plate, the total surface area of the filter is dependent upon the size and shape of the grains of gravel.  The lowest area per unit volume of any shape is the sphere, ergo the advisability of employing a coarse angular gravel.  Large grains also provide less area per unit volume therefore total filter surface area will increase as the size of the grains decreases.  A word of caution here as the filter will rapidly become clogged if the gravel is too fine.  Granules 2-5mm are recommended.  Uniformity of size is fairly important as well since a mixture will cause flow irregularities and cut the efficiency of the filter.  Possibly as much as 25% of the conditioned filter’s capacity is provided by detritus and the attachment sites it provides.  Peak filter efficiency requires some detritus but also detaches beneficial bacteria from the gravel surfaces.

As is shown earlier the filter is living, consuming oxygen, and producing various metabolites.  The rate at which these processes take place is dependent upon the concentration of oxygen available to the processor.  For this reason, a turn-over rate of one gallon per square foot per minute (1gsfm) is necessary in order to maintain oxygen concentration at or near saturation throughout the filter bed.  While both heterotrophic and autotrophic species can process waste at a low oxygen concentration or even under anaerobic conditions the compounds produced are very different and often toxic.  Acids, carbon dioxide, ammonia, methane, and hydrogen sulfide are all products of anaerobic bacterial metabolisis.  It is important to remember when aerating a system that biological oxygen demand is increased by an active filter bed.  For the reasons given above and some to follow an airlift, system is superior to a mechanical pump.  Versatility, simplicity, low cost, easy maintenance, greater efficiency, and ease of flow regulation are only some of the advantages cited by Spotte.  The efficiency of an airlift can be calculated as percent submergence but is also dependent on the extent of diffusion of the air entering the lift.  In simpler terms 100%, submergence is defined by the lift commencing at the lowest point in the culture system and terminating at the water’s surface.

“The under or sub gravel filter consists of a perforated plate suspending the gravel bed above the bottom of the tank.  When used in combination with an air lift pump, no means yet devised can surpass its efficiency for sustaining biological filtration in the filter bed.”

Another contributing factor to filter efficiency is the pH of the culture water.  In marine systems, the optimum range was found to be 7.5-8.3.  Ideal freshwater system pH ranged from 7.1-7.8 with optimum ammonia conversion at 7.8 and the best nitrite conversion at 7.1.

Temperature is essentially unimportant to the filter bed.  A direct correlation can be drawn between temperature and the metabolic rates of bacteria, however, the optimum temperature was found to be 30C.  This is higher than most culture animals can withstand.  In any case, the bacteria seems to be far more tolerant of even rapid changes than the animals in the system.

Any number of toxic additives or poisons have been shown to have a detrimental effect on the operation of the filter.  Copper, tobacco, insecticides, antibiotics, and other medications either kill the bed outright or depress its activities sufficiently to inhibit its oxidizing capacity.  This loss of filtration, temporary or otherwise, can bring about spikes in the concentrations of ammonia and other waste by-products.  Events of this nature might very well be lethal and at the very least increase stress on the culture.

At this point in the text, two pages are devoted to a formula by Hirayama (1966) for determining the carrying capacity of a marine system.  As most hobbyists do not have the means of measuring the variables necessary for calculation, the formula is omitted from this article.  The gist of it is that ten one ounce fish produce substantially more load on the system than one ten ounce fish.

The conditioned filter system is one that is in equilibrium and hence operates at peak efficiency.  In the process of establishing a biological filter bed the sequence and separation of filtration stages is more obvious.  Initially ammonia will peak within two weeks.  The peak of nitrite concentration may take six weeks.  This lag in the cycle takes place as the Nitrobacter population builds.  Unfortunately, the elevated ammonia concentration inhibits the growth of Nitrobacter and lengthens the delay.  Even after these toxins have peaked, the system is not conditioned, as another two weeks may be necessary for the bacterial population to settle out at a stable level.  A small amount of unwashed gravel from a previously conditioned system can be used as a population seed to help speed the process in a new filter bed.  It is recommended that only hardy animals be used to condition a system, as these animals will be subjected to one stress after another.

Detritus is probably the most misunderstood factor in the biological filtration process; where it comes from, its role in the nitrogen cycle, and what needs to be done about it.  It’s defined as an accumulation of loosely collected material in the culture system.  Not all the detritus visible in the system is due to solid fish wastes.  Heterotrophic bacteria help form detritus from dissolved organic and inorganic substances.  Air bubbles also aid in it’s formation as organics tend to form a film around a bubble much as they form a film at the water’s surface.  As the bubbles burst the coagulation of organic molecules sinks to the bottom.  These tiny bits of detritus cling to each other and collect into a larger more visible aggregate.  The active biological filter bed relies on detritus for attachment sites for bacteria, which also rely on the detritus for a source of nutrients.  A conditioned filter bed heavily laden with detritus will tend to acidify the system water.  This is due to the increased oxidizing capacity furnished by the detritus.  The conclusion is that some detritus is very beneficial but too much can cause other problems including decreased flow through the bed.


Mechanical filtration is defined as the separation and concentration of suspended particulate matter.  Mechanical filters in general are efficient at removing turbidity, organic particulates, and accumulated detritus from the culture system water.  These qualities make it a supreme secondary filter.  By removing solid wastes which will eventually clog a biological or chemical filter the run cycle of these filters can be extended considerably.  The mechanical filtration medium is utilized as attachment sites for bacteria performing biological filtration.  However, every time the filter is cleaned or unplugged for any length of time the biological effect is wiped out.  For this reason, the mechanical filter is a very poor primary filter.

Two mechanisms are employed by a mechanical filter to accomplish its task. Suspended particulates either become trapped in the interses of the filter or become statically attracted to the filter material.  Gravel is a good material because of its low cost and re-usability.  Several factors affecting filter efficiency are gravel size, detritus, gravel shape, gravel grading, and flow.  As mentioned earlier gravel size is inversely proportional to surface area, and likewise filter efficiency.  Along the same lines, the accumulation of detritus enables the filter to trap finer particles.  Rough angular shaped gravel is best because it resists deep penetration of debris.  The irregular surfaces increase the electrostatic potential of the bed.  In back flushable filters, the gravel should be graded coarse - fine, top-bottom to facilitate cleaning.  All other sand filters should employ a uniform grade for maximum efficiency.  A major factor in figuring efficiency of a mechanical filter is flow rate.  In larger systems 100%, turn over daily is a minimum.  With smaller systems, the 1gsfm rule should be used as a minimum.  Also, efficiency is highest when using an even distribution of gravel to prevent flow distortion.

Spotte discusses the mechanical aspects of the biological filter calling it a “slow sand filter” versus a rapid sand filter.  Again, many of the same principles and numbers are repeated.  Efficiency is proportional to surface area.  Gravel grains of 2-5mm should be evenly dispersed to a minimum depth of two inches.  A flow rate of one gallon per square foot per minute should be maintained in order to permit bacterial oxidation, which helps keep detritus from caking.  Cleaning should not be an attempt to remove all detritus from the bed as it aids in both mechanical and biological filtration.  Vigorous stirring of the bed would only serve to dislodge beneficial bacteria.  Stir the bed just enough to ensure that no areas have become caked with detritus.  “Schmutzdecke” is an appropriate term.

The rapid sand filter utilizes some different working principles and a couple of common ones.  Filter surface area is of lesser importance due to the higher flow rate.  The electrostatic potential of the bed is proportional to flow rate.  An intermediate system is created by placing a power head on the airlift of an under-gravel (sand vacuum filter).  Sand pressure filters are canisters such as pool filters. They are good filters for cleaning up turbid water.  When cleaning a rapid sand filter stirring of the bed is inadequate due to the deeper penetration of detritus. Back flushing is a must when flow through the bed begins to decline.

Diatomaceous earth filters have the same basic working principles as other mechanical filters.  Microscopic skeletal diatoms caked to a support mesh act as the filtering material.  Diatom filters are capable of removing suspended particles as small as one tenth of one micron.  This makes the diatom filter an extremely important, albeit scarcely needed tool.  The ability to remove such fine particles makes it possible to actually filter microorganisms from the culture water.  One of two different set-ups could be the popular canister type or the placement of the filtering sleeve in an isolable bay with open access to the system.  The latter type must have isolation capability for the coating process, hence the popularity of the canister.  The four main factors affecting the diatomaceous earth (DE) filter’s efficiency are the pre-coat, body feed, sleeve surface area, and the thoroughness of the cleaning process involved.  Pre-coat is a process whereby a layer of diatomaceous earth is caked to the outside of the support sleeve.  Obviously, a sufficient quantity of DE is required to thoroughly coat the sleeve preventing short-circuiting through an area of less resistance.  The filter cake protects the filter sleeve from becoming clogged with suspended particulate matter, which could be much more difficult to remove.  After pre-coating flow must not be interrupted since this would release the DE cake and inject the loose particles into the culture water upon the return of flow.  Body feed is a means of prolonging the run cycle by the addition of porosity of the cake by preventing the compressible detritus from restricting water flow.  The size and duration of the filtering job dictate the sleeve surface are necessary to achieve an appropriate run cycle.  Cleaning is relatively simple.  Back flushing or rinsing with a strong stream of water is all that’s needed to rinse off the old cake.  Periodically the sleeve should be removed, inspected, and cleansed or organic build up.  Be sure and rinse or soak as appropriate to remove chemical cleansers.  Shortened run cycle, loss of flow, and bare patches on the sleeve after pre-coating are indications of a clogged filter sleeve.  In fresh water systems, the primary cause is organic build up but in marine systems, the culprits also include iron oxide, mineral scale, or algae.

An evaluation of mechanical filtration methods reveals that sand vacuum filters are superior to rapid sand filters simply because of the added biological factor. Maintenance of a DE filtration system is prohibitively costly in time as well as money.  However, the ability to remove all suspended particles can be very useful.


Chemical filtration is defined as the removal of substances from solution on a molecular level by adsorption, ion exchange, oxidation, and chemical breakdown. The word adsorption refers to the use of charcoal or activated carbon.  As in other forms of filtration the efficiency is directly proportional to the surface area of the filter medium.  A carbon filter will remove odor, color, turbidity, suspended solids, organic carbon, and utilizable organic compounds which assimilate oxygen from the system.  Carbon has shown that it adsorbs twenty to thirty percent it’s own weight of the substances mentioned above.
Factors which affect the capacity of a chemical filter of this nature are pH, temperature, time, and surface area available.  Optimum pH for adsorption is 7.0. As pH drops the adsorption rate of negative ions drops and as pH climbs the positive ions react in the same manner.  Temperature and rate of adsorption are directly proportional.  The rate of adsorption over time starts at a peak and declines logarithmically.  As with gravel the total surface area of a carbon filter is inversely related to granule size ergo, the finer the better.  The adsorption rate of carbon is greater than that of charcoal.  While admitting that estimates of carbon’s useful life vary considerably, Spotte recommends two to three months as a usable run cycle.  Let me point out that this filter would require substantially more carbon than the average hobbyist uses in a standard power filter.

Ion exchange resins are electrochemically charged resin beads that remove an ion from solution by exchanging it for another.  A filter of this sort is impractical for small systems and very costly for large systems.

Air stripping is the removal of dissolved organics by foam fractionation.  The common term used in the aquarium industry is protein skimmer.  Air stripping reduces the chemical oxygen demand on the system by 40%, the ammonia by over 90%, and removes CO2 causing an increase in pH.  Air strippers operate by oxidizing and coagulating dissolved organics.  As mentioned earlier some organics have a tendency to come out of solution to form a film around the surface of a bubble.  The bubbles rising to the top of the stripper column accumulate as froth and are separated from the system in an upper chamber.  Efficiency is proportional to the amount of air diffusion and the column height as it pertains to exposure time.

Dissolved organics can also be removed by inducing oxidation.  Ozone and ultraviolet are the two methods discussed.  Ozone (O3) destroys micro-organisms by oxidizing their internal protoplasm into a variety of compounds including acids.  A concentration of only 1.5ppm provides 100% sterilization within 5 minutes.  The use of O3 increases oxidation beyond the capability of air or O2.  However, O3 is incapable of oxidizing compounds which have reached saturation.  An ultraviolet (UV) sterilizer reduces bacteria, protozoans, and viruses by 99%.  When suspended over the surface of the water it will produce O3.  The UV’s efficiency increases with length of exposure time and drops with turbidity.  By decreasing flow rate exposure time can be lengthened.

An evaluation of chemical filtration suggests that an activated carbon filter is the only one which could act as a primary system filter.  This is attributed to the capability of adsorption to work over a broader spectrum than fractionation or oxidation.  In an overall evaluation of the various filtration methods biological is the first choice with either mechanical or chemical systems making good secondary filters.


The carbon dioxide system involves the chemical interactions of carbon dioxide (CO2), water and carbonates (CO3).  A few terms which may need some clarification are:  BUFFER - any substance which inhibits a change in hydrogen ion concentration (reserve alkalinity); ALKALI RESERVE - carbonate and bicarbonate (HCO3) negative ions which neutralize positive hydrogen ions when an acid is added to the water; HARDNESS - a measure of cations (positive ions) or calcium carbonate (CaCO3) concentration;

pH - negative log of the hydrogen cation concentration.

pH = log 1/H+

Hydrogen cations or hedonism ions (H+ & H3O+) make the water more acid while hydroxyl ions (OH-) make the water more basic.  In a freshwater system the primary buffers are CO3 and HCO3 ions.  These can be derived from three sources.  First is an equilibrium reaction between CO2 and water.

Carbon dioxide and water combine to form carbonic acid which disassociates as the pH climbs producing hydronium and bicarbonate ions.  If required the HCO3 buffer can disassociate further into CO3.  The second source for HCO3 is also an equilibrium reaction.  Calcium carbonate plus CO2 and water reform to produce free Ca and HCO3 ions.

CaCO3 + CO2 + H2O == Ca++ + 2HCO3

Finally CO3 -- and HCO3 - can also be produced by bacteria during reduction processes.  With a pH of less than 9.0 CO2 can combine with ammonia (NH3) in an aqueous solution to form; a) ionized ammonia (NH4) and bicarbonate ions, b) NH4 and carbonate ions.

CO2 + NH3 + H2O -- NH4 + HCO3
CO2 + 2NH3 + H2O -- 2NH4 + CO3

These natural processes when combined with poor turn over may lead to localized areas of alkalinity in which the pH could exceed 9.0.  Under this extreme pH, ammonia will react with Ca and HCO3 ions causing ionized ammonia and CaCO3 which precipitates out.

Several factors affect the solubility of mineral carbonates.  PH is inversely proportional to solubility.  Magnesium concentration seems to be directly related to the solubility of Ca ions.  Magnesium in solution will readily exchange with fixed Ca.  Because of this tendency, gravels which contain Mg make better buffers.  Good buffering gravels include crushed oyster shell, crushed coral, limestone and dolomite.  Of these dolomite is the best due to it’s chemical makeup, CaMg(CO3)2.  Free CO2 is directly proportional to the solubility of carbonates because of it’s reverse effect on pH.  Dissolved organics can prevent solubility altogether by coating carbonate particles.  This can rapidly lead to a total lack of buffering capability.  The last factor affecting carbonate  solubility is the build up of organic phosphates which have been shown to inhibit the precipitation of mineral carbonates.

A couple of problems can arise to complicate maintaining a stable pH.  Bacteria produce fatty acids which react with NH3 to precipitate Ca out of solution.  Dissolved organics, as stated previously, can coat the gravel and drastically reduce the Ca and Mg exchange sites which reduces buffering capacity.  Nitrification and oxidation reactions have an acidifying effect due to the production of free CO2.  Poor surface agitation or low turn over will drive CO2 concentration up In a weekly buffered system this could cause pH to drop over night.  A rising CO2 level will drop the pH and a dropping pH can boost the CO2 level.  This implies stress.  Gradual changes in pH are not dangerous.  Remember when adding a chemical buffer that this only temporarily removes the symptoms but not the underlying problem  The build up of nitrates or organic compounds which produce the CO2 must be removed.