Steve Hopkins

Rain Garden Goldfish

Hakipu’u, Hawai’i

December, 2013


This article is fairly long and there are no photos, illustrations or video.  We apologize for being long-winded but hope it sheds light on some questions, such as:

Do you need a filter?

What does a filter do?

           gas exchange

           biological filtrer with nitrifying bacteria

           consumption of dissolved organic matter by beneficial bacteria

           mechanical filtrer trapping particulate organic matter

           control of pH

How Many Kinds of Filter are There?

           box filter or corner filter

           under gravel filter

           hang-on-back filter

           sponge filter

           canister filter

           wet/dry filter

           pressurized sand filter

           bead filter

           moving bed filter

           diatomaceous earth filter

           fluidized bed filter

           Japanese mat filter

           plant filtration

           denitrification filter

           foam fractionator

           water exchange instead of filtration



Do you need a filter?

The short answer is, “No”.  Goldfish keeping began in earnest about a thousand years ago.  Having electricity in the home did not become common until about a hundred years ago.  Things like electric air pumps, powered filters, light hoods, test kits, formulated feeds and even glass aquaria are recent innovations in the history of goldfish keeping.  But, the modern conveniences have not necessarily made us better goldfish keepers.  If keeping goldfish without all the new equipment was not practical, enjoyable and successful, the hobby would not have persisted for a millennium.  The message may be that keeping goldfish does not necessarily require the newest and best equipment, but may rely more on intangibles like empathy and husbandry skills.

The more realistic answer to whether a filter is needed is, “Are you crazy.  Without a filter it is harder to manage the tank and you must have fewer fish in the same volume of water.”  The usual rule of thumb is to have ten to twenty gallons of water and adequate filtration for each goldfish in the tank.  This may seem like a lot of space.  But, it is hundreds or thousands of times less space than what the ancestral carp requires in the wild.  We try to make up for the relative lack of space by filtering the water.

This article focuses on goldfish tanks but the same principles apply to outdoor ornamental ponds.

What does a filter do?

The short answer here is, “A filter helps counteract eutrophication and pollution”.  The long answer could fill books.  We’ll touch on aspects of aquarium filters including gas exchange, biological filtration, mechanical filtration and chemical filtration.

Eutrophication is a catch-all term for the bad things that happen when excessive nutrients are added to a water body.  In the case of a goldfish tank, nutrients are added in the form of fish food.  About five to ten percent of the feed is converted into fish tissue for growth.  Goldfish are what they eat and fish growth can be thought of as long term storage of five to ten percent of the feed.  The remaining 90 to 95% of the feed is digested, metabolized and excreted into the water as waste. 

Carbon is the most abundant element in the feed.  One of the primary waste products in the fish tank is carbon dioxide.  Like other animals, fish take in oxygen and expel carbon dioxide.  Movement of water through the filter and bubble aerator enhance gas exchange with the atmosphere.  Excess carbon dioxide diffuses into the air from a micro-layer of water at the surface.  Water movement brings more carbon dioxide into this surface micro-layer so it can escape to the air.  At the same time, oxygen is moving from the air into the water across this surface micro-layer.  So, the oxygen consumed by the fish is being replenished as the carbon dioxide waste is being expelled from the water.  Were it not for this gas exchange, the fish would be suffocated by insufficient oxygen and too much carbon dioxide in the water around it.

It does not take much water movement to generate enough gas exchange.  In most goldfish tanks one weak air stone bubbler is all that is needed.  If the water surface is rippled then there is probably enough gas exchange going on.  Most think that the bubbles rising from the air stone are putting oxygen into the water directly.  Indeed, all around the shell of the bubble there is that micro-layer where gas exchange can occur as the bubble rises to the water surface.  But, at the same time, the bubble is pushing water as it rises generating a broad current that circulates water from bottom to top of the tank.  It is the constant replenishing of the micro-layer across water surface of the entire tank that accounts for most of the gas exchange; not the bubbles.

The tell-tale sign that a tank does not have enough gas exchange is fish gulping at the surface.  They are actually sucking in that surface micro-layer containing more oxygen.  These days, it is very unusual for a tank set-up operating normally to have insufficient gas exchange.  A goldfish gulping at the surface may be suffocating, but it is likely suffocating because the gills have been clogged by parasites or “burned” by a caustic substance such as chlorine or ammonia in the water.

The amount of carbon is much higher than the amount of nitrogen in both the feed and the fish waste.  Goldfish pellets contain 35 to 45% protein and protein is about 16% nitrogen so roughly 6% of the feed is nitrogen.  The amount sounds small, but the impacts are large as nitrogen waste products can be very toxic to the fish. Animals use nitrogen to form new proteins, enzymes, hormones and DNA.  All the nitrogen consumed is later excreted except for a small amount that is tied up indefinitely as new tissue when the fish grows.  

Let’s say we have a medium size goldfish weighing 30 grams in a ten gallon tank.  The amount of feed given each day is about 1.5% of the body weight of the fish or about 0.45 grams per day which is the same as 450 milligrams (1,000 milligrams = 1 gram).  Since roughly 6% of feed is nitrogen and most of the feed nitrogen is excreted into the water, we are adding about 27 milligrams of ammonia to the ten gallons (37.5 liters) of water which equals about 0.72 milligrams per liter or 0.72 parts per million of ammonia nitrogen.  No amount of ammonia is safe and 0.72 ppm is definitely toxic.  Tomorrow we will feed the fish again and the amount of ammonia will double unless it is simultaneously being removed.  Here’s where the biological filter with nitrifying bacteria comes into play.

There are naturally occurring beneficial bacteria which are hungry for the ammonia.  It is an important food source for them.  There are several types of bacteria eating ammonia such as Nitrosomonas.  Give these bacteria ammonia (NH3) and oxygen (O2) dissolved in water (H2O) and they will thrive and rapidly proliferate.  It’s a complex process, but the bacteria basically obtain their energy by de-coupling the nitrogen (N) from the three hydrogen atoms (H3) and attaching it to two oxygen atoms (O2), thereby making NO2, called nitrite. 

Nitrite is even more toxic to fish than ammonia and causes a condition called brown blood disease.  If you can measure nitrite then you have too much.  But fortunately, there is another group of naturally occurring beneficial bacteria, including Nitrobacter, which are hungry for the nitrite.  All that these bacteria need is nitrite dissolved in water to thrive and proliferate.  Again, the details are complex and poorly understood but we can imagine the bacteria create energy for growth and reproduction by adding another oxygen to nitrite (NO2) to create nitrate (NO3).  Fortunately, nitrate is much less toxic to fish than nitrite or ammonia.  No one knows for sure the safe level of nitrate, but twenty to thirty parts per million seems to be OK.

For optimum growth, these nitrifying bacteria must colonize surfaces.  They secrete a gelatinous sheath which allows them to stick to a surface and form a biofilm.  Every surface exposed to the tank water is soon colonized and covered in biofilm including the tank walls, plants, decorations and, of course, the filter media.  When the number of nitrifying bacteria is sufficient to process all the ammonia and nitrite being produced by the feed and the fish, the tanks is said to be “cycled”.  I put “cycled” in parenthesis because it is a poor choice of words to describe this steady state condition of ammonia production with simultaneous conversion to nitrate.  The complete nitrogen cycle also includes denitrification where nitrate is broken down.  More about denitrification later.

These nitrifying bacteria are not the only inhabitants of biofilm and it soon becomes an ecosystem unto itself.  Besides nitrifying bacteria the biofilm has many types of beneficial bacteria that do not rely on ammonia or nitrite.  In addition to bacteria there are fungi, algae, ciliates, other protozoan’s and even larger life forms such as nematodes, planaria and rotifers.  These creatures take nourishment from the water and also prey on each other. 

As organic matter decomposes in the tank, portions of it dissolve into the water.  This dissolved organic matter provides nutrition for a many types of free-floating bacteria, including opportunistic pathogens like Aeromonas and Columnaris.  When there are large amounts of dissolved organic matter the number of opportunistic pathogens increases and when the number of opportunistic pathogens is high there is a higher likelihood that a stressed fish will get sick.  Pathologists call this a dose response.  The sessile bacteria in the biofilm compete with free-floating opportunistic pathogens for dissolved organic matter and help keep their numbers in check.  The importance of consumption of dissolved organic matter by beneficial bacteria is often ignored or underestimated.  Trust me on this… it is very important. 

Unfortunately, it is impossible to measure dissolved organic matter at home.  There are no test kits for it.  Indications that there may be too much dissolved organic matter include cloudy water and bubbles which persist on the surface.  The cloudy water is often caused by proliferation of free-floating bacteria, some of which may be opportunistic pathogens.  They are called “opportunistic” because they are always present, but do not become a problem until their numbers become very high and/or the fish’s immune system has been weakened by stress.  Bubbles persist on the water surface longer than normal or create foam when their surface tension increases.  Think of the water surface as having a skin.  When the surface tension is high the “skin” is thicker and bubbles are reluctant to pop.  Surface tension increases when there is a lot of dissolved organic matter in the water.

While there are no test kits for dissolved organic matter, there are inexpensive conductivity meters for total dissolved solids (TDS).  The total dissolved solids includes dissolved organic matter along with all the minerals that contribute to hardness, alkalinity and salinity.  If you know that everything else is stable then an increase in total dissolved solids can be attributed to an increase in dissolved organic matter.  Everything else is never completely stable so an increase in total dissolved solids is, at best, an indication that the water is getting “old” and well used.

The wide variety of life forms in the biofilm, called its biodiversity, leads to fewer fluctuations in water quality and a more stable ecosystem.  Stability is further enhanced by having a uniform daily food input.  If the amount of food or type of food is suddenly changed, the ecosystem of organisms which break down wastes must adjust their numbers to compensate.  So, when it is necessary to change the daily food ration, it should be adjusted slowly over many days to give the biological filter time to adjust in response. 

A very vulnerable point is when fish are stocked into a tank which has been “cycled” with bottled ammonia (fishless cycle).  Certainly, a fishless cycle is preferable to using guppies or other fish that could harbor parasites.  Don’t ever do that.  The fishless “cycle” is deemed to be complete when nitrifying bacteria are metabolizing ammonia and nitrite and nitrate is accumulating.  But, at this point the biofilm is mainly composed of nitrifying bacteria and lacks the diverse life forms needed to break down organic waste.  When new fish are added there will be a rapid accumulation of dissolved organic matter.  A rapid rise in the amount of dissolved organic matter often results in proliferation of free-floating opportunistic pathogens.  This happens shortly after the fish have been stressed by moving them to a new home.  You now have a high dose of opportunistic pathogens and a weak immune system in the stressed fish creating the ideal conditions for a disease outbreak.  When new fish are added, the focus should be on maintaining pristine water by feeding very sparingly and doing large frequent water changes.

Besides the living creatures, there will be dying organisms and bits of decomposing matter in the biofilm. The hard surfaces are soon covered and the biofilm begins growing upon itself in successive layers of life.  When it gets too thick, chunks will slough off and float away.  The goldfish and the goldfish keeper also interact with the biofilm.  When the biofilm begins to obscure the glass it will be cleaned away.  Goldfish are especially well adapted for foraging on biofilm and you will see them spending much of the day picking at it.  The biofilm is very nutritious and if enough of it could be harvested it could be a complete ration that may be better than the food we buy for our goldfish.  Having the fish pick at the biofilm is probably not a bad thing and there is some indication that the beneficial bacteria in biofilm performs more efficiently after the outer layer sloughs away or is picked off.

In addition to enhancing gas exchange and biological filtration through removal of nitrogenous waste and dissolved organic matter, some filters provide mechanical filtration by trapping particulate organic matter.  The trapped particulate matter is later washed away when the filter is cleaned.  Particulate organic matter, otherwise known as little bits of crud, comes from several sources but often starts as fish feces. 

Feces are made up of non-digestible and partially digested food.  Given the opportunity, goldfish will eat fibrous plant material that is not easily digested and passes through the gut somewhat intact.  When rich, easily digestible food is plentiful the fish may eat so much that the food is excreted before the gut has had time to completely digest it.  You may occasionally see a goldfish consuming fecal matter.  It sounds gross, but is perfectly natural and just the fish’s way of making use of partially digested food.  The fecal matter has microbes from the digestive system and is immediately colonized by other microbes from the tank water and biofilm.  Microbes will break down the fecal matter and essentially complete the digestion process that began in the fish gut.  When microbes attack fecal matter lying on the tank bottom they release ammonia and leave a non-digestible residue.  So, the effect of food on water quality is the same regardless of how well the goldfish utilize it. All this makes it difficult to characterize particulate organic matter.  It can be anything from fresh feces, to decomposing feces that has been colonized by microbes, to bits of non-digestible material that is coated in a thick layer of microbes.  To further complicate the issue, goldfish forage on particulate organic matter (remember that the attached microbes are nutritious too) and recycle it several to many times.  Many goldfish keepers say that goldfish are messy, but in actuality they can be very efficient at utilizing the resources available to them.

The goldfish keeper wants to remove particulate matter to the degree possible because it is unsightly and can further degrade water quality.  There are other dangers if the particulate organic matter accumulates to the point where it becomes anaerobic.  This can happen when the microbes are removing oxygen faster than it can diffuse into the accumulated sludge from the overlying water.  There will be more about anaerobic processes later.

Fish tanks filters can function as both biological filters and mechanical filters to a certain extent, but there are several innate problems that make it difficult.  It is important to remember that biological filtration and mechanical filtration cannot happen in the same space at the same time.  Beneficial bacteria will colonize every available surface and provide biological filtration as long as there is water flowing over the bacteria.  However, if that surface is also trapping particulate matter the accumulating debris will cover the biofilm and restrict water flow around it.  So, multipurpose filters usually have a layer or zone where particulate matter is captured, followed by an area for biological filtration. 

Getting the particulate debris into the filter where it can be captured is often a problem.  Most particulate debris is slightly denser than water so it will drift to the bottom and collect in deposition areas where the water current is slowest.  In a bare bottom tank, goldfish will frequently plow through these debris accumulations while foraging (a process called bioturbation) so the debris does not normally become anaerobic.  When the fish plow through the debris some of it will be re-suspended and may be taken up by the filter intake.  However, if there is gravel on the tank bottom the debris will work its way down into the spaces between pieces of gravel.  When the debris is trapped and accumulates out of reach of the fish it will become devoid or oxygen (anaerobic) and get very nasty.  Hydrogen sulfide (that rotten egg smell), a little methane, ammonia and other toxic stuff develops in the anaerobic spots.  It will slowly seep out into the overlying water.  When you try to clean it a large slug of nasty stuff is released all at once.  Disease problems often accompany a nasty anaerobic layer of gravel.  It is thought that dissolved and gaseous toxic substances weakens the immune system and makes the fish more susceptible to disease.

A fourth possible function of aquarium filters is control of pH.  This only pertains to filters that contain a buffering mineral like calcium carbonate.  Respiration, nitrification and decomposition processes produce slightly acidic by-products and waste.  If left unchecked, these acids would cause the pH to fall to levels that are toxic to the fish.  But, the buffering capacity of the water counteracts changes in pH.  Alkalinity is a measure of the buffering capacity and alkalinity will decline slowly over time as the carbonates are used to neutralize acids.  Materials containing calcium carbonate will slowly dissolve and replenish alkalinity, thereby maintaining a stable pH.  Unless the alkalinity of your tap water is over 100 ppm, it is a good idea to keep some crushed oyster shell, limestone, coral or even egg shell in your filter or elsewhere in the tank.

How Many Kinds of Filter are There?

Lots, but the technology and preferences change over time. 

Sixty years ago, the home aquarium was likely to have a box filter or corner filter.  The box or corner filter is a perforated clear plastic housing containing angle hair floss and gravel.  The gravel is mainly there to weigh down the filter box and keep it from floating or moving.  Angel hair was made of fiberglass strands and, like fiberglass cloth, can be uncomfortable to work with.  It has since been replaced with polyester floss.  The box or corner filter is powered by an air lift.  The rising bubbles from an air diffuser pushes water up a larger diameter tube pulling it out of the housing.   In response, water enters the housing through perforations in the top, passes through the floss and gravel, then through a false bottom and back out through the air lift.  When the white floss becomes dark with accumulated particulate organic matter, the filter is removed from the tank and the media is cleaned.  If it is not cleaned too vigorously much of the nitrifying bacteria on the floss and gravel will be preserved.  In the heyday of the box filter, the importance of nitrifying bacteria and biological filtration was just making its way into the aquarium hobby.

Based on the principle that more is better the under gravel filter came into vogue.  The under gravel filter operates somewhat like the box filter.  A perforated false bottom plate covers the entire tank.  A layer of gravel covers the false bottom.  An air lift pulls water from under the false bottom forcing water to flow down through the gravel bed.  Under gravel filters were very popular for decades and do some things very well.  There is plenty of surface area for nitrifying bacteria.  The gravel bed traps a lot of particulate organic matter clearing the water and hiding the crud out of sight.  It is powered by an air pump so you get gas exchange and filtration with one appliance.  Part or all of the gravel may contain calcium carbonate to replenish alkalinity and stabilize the pH. 

But, under gravel filters have a serious drawback…. they are difficult to clean.  As particulates accumulate in the gravel bed the interstitial spaces between pieces of gravel become clogged in some areas and all the water begins flowing through a few remaining channels.  When water is not moving through the gravel the debris becomes anaerobic leading to a dangerous situation.  An ingenious gravel cleaning device uses a small diameter siphon hose attached to a large diameter cleaning tube.  Water flowing out of the tank through the siphon tube is strong enough to lift gravel into the large tube and allow the accumulated debris to flow out with the water.  It is a tedious process but it works to clean the gravel, although a lot of debris is left lying under the false bottom plate.  An incidental problem with under gravel filters is that the gravel is often about the same size as a goldfish mouth.  When the fish are foraging on the bottom for tid-bits to eat a piece of gravel can become lodged in their throat choking them.

Under gravel filters have been largely replaced with the hang-on-back filter, also called hang-on-tank filter.  Personally, I have never been fond of these things but they are tremendously popular and are usually what the pet store salesperson will offer with a new tank.  They have the advantage of being an all-in-one appliance that is easy to set up, quite and not too conspicuous.  The media is often enclosed in easily removable and replaceable cartridges.  Depending on the type and configuration, the cartridges can provide mechanical filtration, biological filtration, or chemical filtration using activated carbon. The present generation of hang-on-back filters usually have a small magnetic drive pump build into the bottom of the housing.  The pump pulls water into the filter housing; water passes through the media and then overflows back into the tank.  Because the pump only has to push against several inches of head (lift), the electrical efficiency is OK to good.  The media cartridges are convenient, but are a continuing expense.

But, the amount of surface area for nitrifying bacteria in a hang-on-back filter is usually small.  Some models try (in vain) to overcome this by incorporating a biowheel.  The biowheel looks like the paddlewheel on an old time steamer.  It is a type of wet/dry filter where the nitrifying bacteria attached to its surface are alternately submerged in water and then exposed to air as the wheel turns. Wet/dry filters tend to process more ammonia per unit surface area of biofilm than continuously submerged media but some biowheel designs require some fiddling to keep the wheel turning.  The mechanical filter in a hang-on-back must be cleaned often, but is fairly easy to do owing to the removable cartridge that can be rinsed in the sink.  The activated charcoal in a hang-on-back filter can create a false sense of security.  Activated carbon has a huge number of microscopic pours and ion exchange sites that absorb dissolved substances.  Its use is often called chemical filtration as activated carbon attracts binds and, in some cases, chemically alters substances dissolved in the water.  These include chlorine, ammonia and certain types of dissolved organic matter.  In a new tank with new water, a new filter and new fish the nitrifying bacteria are not yet established and activated carbon can be very useful in preventing ammonia from reaching toxic levels.  But, then again, with little or no ammonia in the water the nitrifying bacteria cannot proliferate and the tank will be slow to become “cycled”.  Using activated carbon in this way is a crutch for the uninformed beginner at best.  Removing chlorine from tap water would also be of considerable benefit if the fish were not already in the tank.  Using a water conditioner for chlorine removal is faster, safer and less expensive.  As mentioned, activated carbon also removes certain types of dissolved organic matter.  Removing dissolved organic matter is always a good thing.  The trouble with activated carbon is that you never know when the bulk of the absorption sites are used up and you may think it is doing something constructive long after it is expended.  In most fish tanks, the activated carbon is probably active for only a week or so.  Once the absorption sites are used or clogged the activated carbon is nothing more than an attachment surface for nitrifying bacteria.

The water flow through a hang-on-back filter is fairly rapid.  When the filter is placed on a tank of the rated size, the tank volume is pumped through the filter three or four times per hour.  The faster water flows over the biofilm, the more efficient the biofilm is in converting ammonia to nitrate.  The more water that is passed through the mechanical filter the more particulate matter it will capture.  Also, the high flow rate helps keep particulate matter suspended in the water (rather than settling to the bottom) so there is more debris available for capture in the mechanical filter.  This all works great for some fish species.  However, the high water flow may not be so good for goldfish. Many goldfish varieties are bred to have very large fins and tail.  The goldfish will be marginally strong enough and athletic enough to handle the exaggerated tail.  When placed in a strong water current, the tail acts like a sail catching the current and the fish has to work harder to maintain its position mid-water in the tank.  The extra effort may eventually exhaust the goldfish and it will sit on the bottom or hang in a corner rather than fight the current.  Other goldfish may be strong enough to fight the current, but the fins and tail collapse or become frayed.  As a general rule of thumb, goldfish should be able to comfortably hover in mid-water without being swept about by the current.  Because goldfish are often believed to be more “messy” than other species (a characterization that I would argue), some goldfish keepers use a hang-on-back filter rated for twice the tank size, or use two filters instead of one.  There is often some sort of air diffuser as well in case the filter pump becomes jammed.  The larger the filter and the more filters and air diffusers you have the stronger will be the water current.  You cannot have too much filtration, but you can easily have too much water current.  Some filters have provisions for reducing or dispersing the water flow back into the tank.  Try it, but remember the suggestion that the fish should be able to hover mid-water without being swept about by the current.

The sponge filter is simple and, in many ways, reminiscent of the old box filter.  The sponge filter has a weighted base and an upright air lift tube.  Part of the air lift tube is perforated and encased in a block of open cell foam.  The open cell foam or sponge is porous and allows water to flow through it.  Nitrifying bacteria will colonize the pours throughout the block of foam and particulate matter will also become entrapped.  Sponge filters are relatively inexpensive and come in a wide variety of shapes and sizes.  They are fairly easy to clean, but may need to be cleaned often.  Because they are driven by an air diffuser, they usually do a good job of gas exchange.  They do not accommodate calcium carbonate rock for replenishing alkalinity and stabilizing pH.  The main problem with sponge filters is that biological filtration and mechanical filtration cannot occur in the same space at the same time.  The accumulating debris will suffocate the biofilm.

A canister filter typically has media for biological filtration, media for mechanical filtration and, perhaps, activated carbon as well.  Canister filters are known for having ample space for filter media and the flexibility to accommodate many types of media including something with calcium carbonate to replenish alkalinity and stabilize pH.  They are generally larger than a hang-on-back filter and are often situated under the tank where they are inconspicuous.  A hose runs over the side of the tank into the bottom of the canister filter housing.  Water flows up through the various layers of media and is picked up by a pump in the top which moves water back to the tank through a second hose.  Since the canister filter housing is completely closed and water tight, the standing water level in the tubing when the pump is turned off is the same as the tank water level.  So, the pump static head (lift) is only a few inches and the pump efficiency is good.  The biggest complaint about canister filters is that they are expensive and difficult to clean.  Since the canister housing and hoses are always full of water, there must be a provision to open the housing to clean the media without spilling water. Occasionally, the inlet and outlet hoses themselves become fouled and obstructed and these are notoriously difficult to clean.   Because the housing must be water tight and pressurized it requires sophisticated seals that tend to increase the overall price.  Because the housing is air tight, canister filters do not provide much gas exchange unless the return hose in the tank is fitted with a spray bar.  There is always a possibility of having too much current in the tank when using a canister filter.

The concept of a wet/dry filter was briefly described above.  These are also called a trickle filters.  Rather than having the biological filter media submerged, the media is supported above the water level and water trickles over it.  Conversion of ammonia to nitrate is more efficient in biofilms that are wet yet exposed to the air.  The gas exchange is excellent.  There may, or may not, be some provision for particulate removal.  It is possible to use a calcium carbonate media for replenishing alkalinity and stabilizing pH.

There are two basic configurations for wet/dry filters.  The one most commonly used on aquaria is a vessel holding the media and a sump placed under the fish tank.  Water flows by gravity from the fish tank and is dispersed over the media.  After flowing over the media, water collects in a sump at a lower level and is pumped back up to the fish tank.  Because the static head (lift) of the pump is several feet, the energy efficiency is not very good.  Also, there must be a provision for power outages because if the pump shuts off water will continue to flow down from the tank and the sump will overflow.  The common way to accommodate this is to draw water from the surface in the fish tank.  If the pump shuts off the fish tank will soon stop draining; hopefully before the sump overflows.  When using a wet/dry filter below the tank, there is typically a bulkhead fitting through the tank wall.  Also, the pump flow must be regulated to keep from running the pump dry and there must be some way to dissipate the flow so there is not too much current. Commercially available wet/dry filters of this type often use plastic bioballs (trademarked name) for media.  Home made versions may use Matala mat or something less expensive like plastic scrub pads or shredded plastic soda bottles.

The second configuration of a wet/dry filter is to have the media supported above the tank.  These are home made filters and a popular design is to use a plastic show box or tray with a perforated bottom to hold the media.  Water is pumped from the tank with a small power head and is dispersed over the media.  Water trickles through the media and drains through the holes in the bottom of the tray directly back into the tank.  Gas exchange is excellent.  Particulates are not captured unless a sheet of floss batting or sponge is laid on top of the media.  Energy efficiency is a little less than that of a hang-on-back or canister filter owing to the height of the media tray but efficiency is much better than a wet/dry filter placed below the tank.  A calcium carbonate media is easily accommodated for replenishing alkalinity.  Because of the way water rains back into the tank there is little danger of having too much current.  The main drawback of a trickle filter over the tank is that they can be unsightly and noisy.  If there is enough light, plants can be grown in the filter tray as this set-up is identical to some aquaponic systems.  More about plant filters later.  The noise comes from the water raining from the tray into the tank.

For what it’s worth, a wet/dry filter above the tank is my favorite type of filter.  For media I use golf ball size pieces of coral rubble.  The large media pieces do not trap debris.  A very small power head pumps water from the tank into the media tray sitting on slats across the top of the tank.  We clean the filter only when the fish are moved and the tank is completely restarted.  This happens every six to twelve months.  Debris accumulations are siphoned off the tank bottom as needed which is usually every month or so.  It is a very simple and inexpensive set-up and requires less time for tank management than anything else I’ve tried.  It looks terrible but the fish do not seem to care.  We use a scaled-up version for our ponds.

Finally, tanks outfitted with wet/dry filters that use a porous media such as the ceramic bakki shower-type media or coral rubble often have lower than expected nitrate levels.  It is a mystery how this could happen, but some have postulated that there is denitrification going on inside the porous media.  More about denitrification later.

There are numerous other, less popular, types of filters that are seldom used for goldfish tanks and ponds.  The pressurized sand filter was developed for pools and spas but is sometimes used for fish tanks and ponds.  A globe-shaped sealed vessel is filled with a deep layer of sharp sand.  Water is pumped in the top and forced out through screened laterals at the bottom.  The sand traps particulate matter and when the sand becomes clogged it is back flushed by reversing the flow.  The water velocity is supposed to fluidize the sand so the light debris can be flushed away while leaving the heavy sand behind.  These can work in the somewhat sterile water of a pool or spa but are a disaster with fish water.  The same organisms that create biofilm cake the sand particles together and the filter cannot be adequately back flushed.

A bead filter also uses a sealed housing that is either globe shape or hourglass shaped.  The media is a layer small floating plastic beads.  Debris becomes trapped in the beads.  When the filter is back flushed to remove the debris, the floating layer of beads is fluidized by air bubbles, water velocity or, in very large models, a propeller.  The plastic beads are easier to fluidize than sand so they are somewhat less prone to caking.  Bead filters are often advertised as being both mechanical filters and biological filters but we now know that biological filtration and mechanical filtration cannot occur in the same space at the same time.  So, they behave more like a biological filter when back flushed frequently, and behave more like a mechanical filter when back flushed less frequently.  If not properly cleaned the plastic beads can cake together making them non-functional.

A moving bed filter provides a large amount of biological filtration in a small amount of space and is nearly maintenance free.  The media, called Kaldness or K1, is little extruded buoyant plastic pieces with a complex shape that has a lot of surface area.  An air diffuser is used to keep the plastic media in constant motion so it does not collect solids. 

A static Kaldness filter uses the same K1 media but the media does not move.  Static Kaldness is efficient at trapping solids and it is easy to then fluidize the media with air, water or by hand and then wash the particulate matter away.  A static Kaldness filter is much like a bead filter but requires less energy to clean and there is less possibility of caking.

A fluidized bed filter uses very fine sand as media.  The sand is placed in a tall, narrow cylinder and is kept suspended in the water by a strong water flow from bottom to top.  The movement prevents the sand from trapping particulate matter.  Because the sand grains are so small, fluidized bed filters provide an enormous amount of surface area and the most biological filtration in the smallest amount of space.  However, they are finicky as the water flow must be precisely controlled and if it ever shuts off it may not be possible to re-suspend the sand.

Diatomaceous earth filters are really good at removing even the smallest particles of debris.  Diatomaceous earth is a mineral deposit made up of the silica cell walls of diatoms (a single-cell algae) in ancient sea beds or lake beds.  It is mined from deposits in Europe, the U.S.A and elsewhere.  Diatomaceous earth crumbles into a very fine powder. 

A diatomaceous earth filter has perforated tubes or plates covered in fine plastic mesh on the filter outlet.  The diatomaceous earth powder is poured into the filter housing where it “cakes” on the plastic mesh and is held in place by the water pressure.  Particles of debris as small as a few microns are then captured on the surface of the cake as water passes through the pours.  The filter soon becomes clogged with debris and is cleaned by backflushing the cake off the plastic mesh and discharging the diatomaceous earth and the debris it has trapped.  When they clog you go from ultra filtration of fine particles to no filtration at all.

The Japanese mat filter functions as a biological filter when it is kept clean, but functions as a mechanical filter if mat with a finer weave is used.  The material is a semi-rigid spun plastic.  A similar product is Matala mat.  The material is sold in large sheets for do-it-yourself applications or precut to fit in a commercial filter housing.  Some have cut the mat to fit hang-on-back or canister filters.  Some use the round pads sold for institutional floor cleaning machines as an inexpensive alternative to Japanese or Matala mat.

Plant filtration is as old as goldfish keeping itself.  The idea is that the plants will remove nitrate and/or ammonia from the water and utilize it as a nutrient source.  Submerged plants are also supposed to be oxygenators because they take up carbon dioxide and release oxygen into the water as a by-product of photosynthesis.  Indeed, these processes happen but if you do the math you may find that the overall effect is often negligible.  Plants use various forms of nitrogen, phosphorus, carbon and other elements to build more plant tissue and grow.  But later, when that new leaf has served its purpose and started to die, much of those nutrients are released back into the water.  So, plants do not really remove nutrients, they just store them for a while.  To truly remove the nutrients, the plant must be pruned and the cuttings thrown away. 

In the presence of sufficient light, plants take up carbon dioxide, use it to create sugars for energy, and expel oxygen as a by-product.  Simultaneously, the plant is respiring.  In respiration, the sugars and oxygen are “burned” to create energy for growth and metabolism, and carbon dioxide is expelled.  Photosynthesis happens during the day (or as long as the lights are on) while respiration continues around the clock.  Once the oxygen consumed in respiration is subtracted from the oxygen produced through photosynthesis, there is not a lot of excess.  This small net gain in oxygen was important in the days before electricity and air pumps, but is insignificant in a modern aquarium with moving water and a lot of atmospheric diffusion of oxygen.  But, with that said, I think plants are a great addition to a goldfish tank.  Having plants to pick at does wonders for goldfish health, color and overall appearance.  Plants in the tank or growing in a filter compartment are aesthetically pleasing and create an additional layer of interest.

Denitrification filters are seldom used and they’re tricky.  In denitrification, nitrate (NO3) is converted to nitrite (NO2), the nitrite is then converted to nitric oxide (NO) and/or nitrous oxide (N2O) which, in turn, is converted to nitrogen gas (N2).  Nitrous oxide is a potent greenhouse gas but, fortunately, not much of it escapes to the atmosphere before it is converted to nitrogen gas.  Nitrogen gas makes up 78% of the air we breathe while only 21% is oxygen. 

Like the nitrification processes in a biological filter, denitrification is performed by various types of bacteria.  Nitrification bacteria require oxygen to survive, but denitrification only occurs in the absence of oxygen.  Denitrifying bacteria are essentially using the oxygen (O3) attached to the nitrogen (N) in nitrate (NO3) because there is no free oxygen in their environment.  But, as mentioned above, bad things tend to happen when the fish are exposed to dissolved and gaseous substances produced during anaerobic decay.  The stuff may not be toxic or concentrated enough to kill the fish outright, but it can still weaken the fish enough to make it susceptible to disease.

So, the trick is to take advantage of the benefits of denitrification while avoiding the detrimental effects.  The usual approach is to feed the denitrification filter from the fish tank, but isolate the effluent from the denitrification filter until it can be aerated and made safe.  A popular filter configuration is the coil denitrator.  This is nothing more than a very long section of tubing that is coiled to make it compact.  Tank water slowly flows through the coil of tubing; very slowly.  The water flows so slowly that aerobic microbes use all the oxygen and then anaerobic microbes converte the nitrate to nitrite, then nitric and nitrous oxide, and finally nitrogen gas.  If water flows through the coil too fast then the process is not completed and there could be nitrite flowing out the other end.  Another denitrification configuration is a plenum which is nothing more than a separate tank with a deep layer of sand.  Water slowly migrates into and out of the anaerobic sand bed.  Water exiting the coil denitrator or plenum is heavily aerated before it is returned to the fish tank.  The aeration drives off toxic gases and kills most anaerobic bacteria.

Denitrification filters are sometimes used by gadget conscious marine reef keepers, but it is usually easier to just change the water in freshwater tanks.

Foam fractionators or protein skimmers are a type of chemical filter that removes large molecule dissolved organic compounds like dissolved proteins and amino acids, fats and fatty acids, carbohydrates and some metals such as copper.  These substances have an ionic charge which makes them surface active so they concentrate at the air-water interface.  In this way, the dissolved substances are collected on the skin of air bubbles as they rise to the surface.  The smallest possible bubble is used as they will have the largest possible surface area per unit volume of air.  The increased surface tension makes bubbles persist when they reach the surface creating a foam.  This foam containing pollutants is trapped by the filter, collected and disposed of.  Some particuate matter becomes trapped in the foam as well so, to some extent, foam fractionators also function as mechanical filters.

Dissolved organic compounds provide nourishment for free-floating bacteria.  When bacteria proliferate and create a bacterial "bloom" the water will appear milky.  Some of these bacteria may be opportunistic pathogens.  As dissolved organics are metabolized by bacteria ammonia is created.  So, removing dissolved organic does much to improve the goldfish environment.  It is too bad that foam fractionators are not more efficient at what they do.  By the time there is enough dissolved organics to create a firm foam that can be collected and removed, there is already too much dissolved organic pollution.  Perhaps you should have changed the water before it got this bad.  There have been numerous design improvements in recent decades, but the process is still unpredictable and the units require a lot of fiddling to get the most out of them.  They tend to work better in high salinity water but are not-so-good in freshwater.

Using water exchange instead of filtration is an alternate and/or complimentary means of managing the water quality. In most aquaria, the nitrate continues to accumulate over time.  The toxicity of nitrate to fish is low and levels of twenty to thirty parts per million are common.  But, fish health tends to deteriorate when nitrate gets very high.  We do not know if, 1) these problems result from the toxicity of nitrate itself, or 2) nitrate stresses the fish, weakens the immune system and makes the fish more susceptible to disease, or 3) nitrate is just a convenient indicator of other pollutants or pathogens that accumulate at the same time nitrate does.  In any case, changing water on a regular basis seems to avoid the problem.  If you exchange enough water, no filtration is needed at all to successfully keep goldfish.

The quality of your tap water or well water and how you handle it is important.  Some water sources, particularly those near a lot of intensive agriculture, contain nitrate right out of the tap.  Some water is naturally very soft, has little natural buffering capacity, and requires alkalinity adjustment.  Some water needs remediation because it has had the pH adjusted downward by the water authority to prevent lead from leaching out of old plumbing systems.  Some water is very hard but goldfish do not seem to mind even when it is 300 ppm.  Most municipal water has been disinfected with chlorine or chloramines.  Before water is used in the goldfish tank a water conditioner should be used to remove chlorine or chloramines, the water should be aerated, and the temperature should be adjusted to that of the tank.  If the pH has been adjusted downward it should be bought back up to pH 7 or above before use.  Alkalinity can be augmented after the new water is added to the tank.


The primary message here may be that there is no one way or best way to keep the water clean for our goldfish.  It depends on how much you are willing to spend on equipment and electricity, how much time you are willing to invest, how much space you have and what you want it to look like.  Everyone must find their own way.

Hopefully, another important message is that there is more to water quality than ammonia, nitrite and nitrate.  Having a filter that is “cycled” does not mean nothing can go wrong.  Of particular concern are dissolved organic matter, alkalinity and proliferation of free-floating bacteria.

Finally, remember that water change can resolve many problems.  If you have good tap or well water to work with, then you cannot have too much water change.  If you suspect there is a problem, change the water.



contact us at raingarden@hawaii.rr.com 

49-041 Kamehameha Highway

Kaneohe, Hawaii 96744 USA

808-294-3973 for goldfish

808-294-1921 for flowers & vacation rental