In the first two articles in this series, I discussed some of the dissolved components, mostly trace metals, found in the water from a survey of 23 marine reef aquaria. I chose to examine these elements for two reasons: first, many reef aquarists seem very concerned, "obsessed" is not too strong a word, about trace elements in their systems and, second, the concentrations of these materials could be analyzed relatively inexpensively. I have shown that the concentrations of many of these elements in our reef tanks bear little similarity to what is found in natural sea water (NSW). Additionally, I have shown that many of these elements have distributions that are correlated with one another. After comparing these data with the composition of the artificial salt mixes as reported in the literature (Atkinson and Bingman, 1999), it is apparent that some of the odd, and relatively very high, concentrations are due to the initial mixtures of the artificial sea water that we use in our systems. Our systems are very dynamic, however, and after a very short period bear only a passing similarity to what is found in either natural or freshly mixed artificial sea water.

That our systems are dynamic and ever changing should come as no surprise to any reef aquarist. However, the magnitude of some of these changes is quite impressive and the speed with which the systems can accommodate change is truly awesome. Most, if not all of these changes, are mediated by the organisms in the reef aquaria, and probably the most important groups of organisms in this regard are the bacteria and algae found in the aquaria (Redfield, et al., 1963). Under appropriate conditions, various bacterial and algal species will have large populations in our systems, and they have the capability for rapid metabolically induced changes of the medium. In natural reefs, about 80 percent of the non-bacterial biomass is algal (Odum and Odum, 1955). Coral reefs are, in point of fact, algal reefs with a small animal component layered over them like the thin frosting on a cheap cake. It is likely in many of our aquarium systems that the algal component is equivalently abundant; algae grow on the surfaces of virtually everything in our tanks, as well as within the rocks, corals, and sediments. When we set up our aquaria, we are actually constructing algal culturing vessels of very good design. Of course, all aquarists, myself included, then gripe loudly and vociferously about the algae "taking over" our tanks. Yeah, right... The rapidity of algal growth is amazing, as demonstrated by the dinoflagellate or diatom blooms we have had at one time or another.

For most aquarists, an unforeseen or unexpected side effect of such rapid algal growth is a significant alteration of the water chemistry; as some elements vanish from solution, while the abundance of others is hardly changed. Of course, some of the animals alter the water chemistry as well, but probably most of the alteration comes from either algae or bacteria. Such rapid alterations of water chemistry are characteristic of these two groups in natural systems, and are not at all characteristic of animals. Most animals use relatively little in the way of dissolved materials, getting most of the chemicals they need from their food. One rather peculiar animal, the reef aquarist, however, does significantly alter the composition of the fluid in reef aquaria. Aquarists can change the water chemistry of their systems either actively or passively. In the former case, they may add or remove materials from the system. In the latter case, they may simply do nothing, and in so doing, abet the changes occurring due to the actions of the various organisms in the system.

Undoubtedly as well, there are many strictly non-biological chemical reactions occurring in our systems, but with few exceptions, such as the effects of lime water or added buffers, these are likely of lesser magnitude and importance than are the biologically mediated responses.

In an attempt to estimate some of the effects that aquarists have on their systems, I asked the participants of "The Tank Water Study" to provide me with a detailed list of foods they used and their feeding schedule. I also asked for data on water changes; both if they performed them, and if so, how much water was changed at each instance and how frequently these changes were accomplished. By examining the composition of the foods that go into the aquarium, as well as the frequency of water removal, I could, rather crudely, estimate some of the factors influencing the composition of the materials dissolved in the aquarium water.

The nutrient and trace element composition of the foods was estimated by calculating the mass of the food added and comparing the foods to compositions of comparable foods in what I call "The Food and Additive Study"(Shimek, 2001). To estimate the mass of food added to the aquarium, I had to convert from a number of different odd and wonderfully esoteric measurements into metric units (Table 1). Most marine animal and plant tissue has a specific gravity of close to one, and I assumed that one milliliter of the foods derived from such tissue weighed one gram. This will introduce some errors, as some foods are a bit denser. These foods will be undervalued with regard to their contribution.

For some foods there were no available comparable data, so I estimated their values by using a similar food. For example, for an addition of mysids, I used brine shrimp as a surrogate. Likewise, for most of the algae used by the aquarists in the study, there were no data available. I used the values for dried Nori (Porphyra spp.) as a surrogate. This may over-represent the contribution of those algae, as Nori is a very rich food source. A square inch of whichever algae were used was estimated to be equivalent to 0.25 g of Nori.

Every feeding regime noted by each hobbyist was different, somewhat inconsistent, and approximate. To be comparable, I needed to bring them to the same standard, so I divided my estimate of the amount of food added in one week by seven to give "Average Daily Ration." This measure was used in all comparisons, even though many aquarists did not feed on a daily basis. The data from these calculations were tabulated and graphically compared to the concentrations of the trace elements found in natural sea water (Figures 1a - 1f).

To estimate the effect of the added foods in the tank, I asked each participant to give their system's size in gallons, and then to estimate the total volume of the rockwork and the sand. I assumed that the rock and sand totally displaced water. This is patently false, there is pore water in any sediment bed and there is water inside of the rock. The volume of this "internal" water is difficult to measure and even harder to estimate. However, in both of these cases, exchange with the main water volume of the tank will be slow, and these areas may be considered to be effectively isolated from immediate changes. In any case, however, the amount of water in each system was probably somewhat underestimated, maybe by a factor of as much as twenty percent. Such a reduction has the effect of increasing the calculated concentrations (Table 2).

The composition of the material fed was estimated by multiplying the components of the average daily ration by data from "The Food and Additive Study" (Shimek, 2001). In some cases, I was able to use the data directly from that study, as some of the participants indicated that they used foods that were examined in that study. In other cases, I averaged the values for a given component of food. For example, to estimate what was in a "generic" flake food, I used the average of the dried prepared foods examined in "The Food and Additive Study" (Shimek, 2001). For each of the foods, values for about 30 elemental concentrations were estimated. These were summed over all the foods. This sum was "The Average Daily Food Ration (Table 3).

For each aquarium, the proportion that the average daily food ration contributed to a water volume equivalent to that system was calculated relative to the same NSW volume (Pilson, 1998) (Table 3). These data are rather interesting. The average DAILY food ration contains enough aluminum, if it were totally dissolved in the tank water, to raise the aluminum concentration of an average tank from zero to 1.572 times normal. It contains enough iron to raise the concentration of iron in an average tank from nothing to 35.141 times the level found in natural sea water. This bears repeating: each day, each aquarist, added enough iron to an average tank to raise the iron concentration to over 35 times the normal sea water concentration. And yet, the tank water contains no detectable iron (Shimek, 2002a, 2002b).

Tank-by-Tank Comparisons of the Input from Feeding

No tank is average, of course, and it is useful to examine the input from feeding on a tank-by-tank basis. In these graphs, the average daily feeding ration calculated specifically for each tank is given, along with the tested values for that particular tank's water. Please note that I used the average daily feeding ration in these graphs, while many of the aquarists did not feed daily. That simply means I took their periodic feeding ration and averaged it evenly over each day. As these graphs are specific for each tank, and each tank's daily feeding ration, the graphs represent some of the variability seen in reef tanks, at least as expressed in this study (Figures 1a -1f). The graphs in these figures have a pair of colored bars for each element listed on the horizontal axis. The height of each bar indicates the concentration of the elements as a proportion of its normal NSW concentration. In each of these figures, the yellow bars indicate the average daily additions of trace elements listed on the "x" or horizontal axis, while the pink bars indicate the tank water concentration for that element as determined in this study. Note that the values along the "y" or vertical axis are logarithmic and are fractions or proportions of the NSW value. The horizontal line labeled "1" in the middle of each graph represents the NSW value, any yellow or pink bars near that line indicate values close to those found in natural sea water. Each line indicates values 100 times greater than those indicated by the line beneath it.

Click here to view the figures.

As I indicated in last month's article (Shimek, 2002b), several of the elements examined in the tank waters were below the detection limits for the tests. These elements are clearly recognizable in the graphs by the lack of a pink vertical bar paired with a yellow vertical bar. Interestingly enough, all of them are sufficiently concentrated in the food to be apparent in the feeding rations. In some cases they appear to be quite concentrated in the average ration. These are elements such as beryllium, iron, lead, manganese, mercury, and silver. Unfortunately, some of these materials, such as mercury, lead, and silver are quite poisonous. The values vary from tank to tank, of course, but it is interesting to note that the values for iron, lead, manganese and mercury all are in the range of one. Put another way, each average feeding puts enough of each of these elements into the system, if they were completely dissolved, to bring their concentrations from zero up to around NSW levels.

Last month, I showed the data indicating that some of the materials in the tanks were at very high levels, and some of the reason appears to be evident in the composition of the various average feeding rations. Cobalt, copper, iodine, phosphorus, titanium, tin, thallium and zinc are concentrated in the foods, and are entering the tanks in relatively high concentrations. These concentrations are sufficiently high that if these materials were all converted to soluble form, the water volume of the tank would receive enough of them to be raised from zero to around NSW levels, EACH DAY, EVERY DAY.

From these data it comes as absolutely no surprise that the concentrations of tin, thallium, titanium zinc, cobalt and copper are high in the water from these tanks. The tanks are getting significant daily transfusions of these metals. Perhaps, what is surprising is that the concentrations of many of the other elements are not equally high.

In addition to trace elements, the daily food ration added another substance... Food. The analysis of the foods, or conventional nutrients, indicates just how little mass is being added to the tank on a daily basis. The total weight of the food was 15.39 ± 15.90 grams, or roughly a half of an ounce of food, however, 11.64 ± 12.13 grams of this was moisture, leaving only 3.75 ± 3.77 grams, or about 0.13 ounce of actual solid food. Ash comprised 0.31 ± 0.32 grams, and many of the trace elements were found in that constituent. Only 1.72 ± 1.84 grams of protein, 1.36 ± 1.75 grams of carbohydrates, and 0.36 ± 0.34 grams of fat constituted the rest. The food provided a total of 15.75 ± 15.61 Calories of nutrient energy. In this study, an average tank contained an estimated volume of about 191 liters, so if the food was totally dissolved and dispersed throughout it, the average tank daily ration would be an overall tank volume of about 20 ppm of food.

And the average aquarist often thinks they are overfeeding their animals!

Starving them would be a better description. Production of some foods within the tank, takes place, primarily carbohydrates in corals and algae, however, the protein-rich foods necessary for tissue growth are largely generated from to food additions; see Hamner, et al.(1988) for a discussion of how much food is actually available to corals and coral reef animals.

To estimate the effects of water changes, I used the data provided by the participants to calculate an average period between water changes, and an average volume of water removed. From this, it is easy to calculate the amount of any dissolved material removed. If the replacement water is made from the same artificial mix, and presuming that the mix hasn't changed, in such calculations, one only needs to worry about the increases caused by feeding, as additions and deletions of the artificial salt water cancel themselves out. Additionally, of course, there would be changes in biologically utilized materials such as calcium, which are added by the hobbyist. But ignoring these for the moment, such calculations can provide an estimate of the "tank load" of the other materials (Table 4).

One might assume that significant amount of these trace elements would be removed with each water change.  If one did assume that, one would initially be dead wrong, the changes do not seem significant.  Using the values for the frequency of water changes and the average amount changed each time, it is apparent that the initial water changes barely slow the rate of increases.  However, as the water gets more and more concentrated, the relative amount removed by each change increases.   By the time 20 water changes have been done, the concentrations tend stabilize albeit at values far in excess of NSW.   In Table 4 the values for Instant Ocean should be added to each of the other columns to give the total concentrations.  The columns as given indicate the accumulations due to feeding alone.  The numbers in the columns of Table 4 are proportional values, or values times natural sea water. 

Some of these values are so large as to be absurd.  Consider the last two columns which would represent a tank that is 4.55 years old.  Enough iron, for example, would have been added over that period to boost the system’s water iron concentration, presuming it had all dissolved, to a level approximately 2,975 times NSW, if water changes had been occurring, or 58,181 times NSW, if water changes had not been occurring.  Many of these levels are probably above the saturated concentrations for the elements in question.

Calcium is a special case amongst all of the elements examined, as it is the only one that hobbyists generally take a consistent and vigorous approach to managing. I thought it would be worthwhile to examine some of the factors that go into maintaining calcium concentrations. Natural sea water has calcium concentrations ranging around 400 to 410 ppm. Most hobbyists specifically try to maintain calcium at least at those levels, if not higher, to boost coral growth and calcification. At calcium concentration levels of about 525 ppm, Swart (1981) found that the rate of coral calcification reaches its maximum, and any additional calcium concentration is immaterial. More recent studies indicate the saturation of calcium may occur at concentrations around 360 ppm (Tabutte, et al., 1996). Consequently, it would seem that most aquarists would attempt to keep their systems within the 360ppm to 525ppm range.

The major tool that aquarists have for ascertaining the calcium concentrations of the water in their systems is a calcium concentration test kit, and there are several of these on the market. To test the validity of using these test kits, I asked the participants in the survey to indicate if they used a test kit for calcium and, if so, what was the tested value for the calcium concentration in the water sample. I also asked the participants which test kit they used. These tested values were compared to the actual values (Table 5).

Amongst the users of test kits the average calcium concentration was 395.4 ppm, so the overall average was pretty good. Unfortunately, if these aquarists had to depend on the test kit values as being anything other than approximations they would be out-of-luck. The average difference between the test kit value and the actual value was 56.2 ppm. It would appear that the combination of aquarist and test kit gives the correct value ± about 56ppm. Aquarists using test kits should aim to keep their tanks calcium levels in the middle of the normal calcium range, and if they use the kits at all, they should use them to assure themselves that the readings were neither extremely low nor extremely high, and not worry about the actual value of the reading as it relates to NSW.

The calcium concentration of natural sea water varies a bit within the oceans, but around coral reefs a value near 400-410 ppm is probably a fair estimate of the concentration. I tabulated the differences between the aquarium calcium concentrations for those who tested their calcium levels and those who indicated they did not. I tested these two groups to determine if the average system calcium levels were statistically significantly different from one another (Table 6).

This test examines the averages of the two groups and calculates the probability that the two averages could be randomly drawn from a larger group. Note that this means it is immaterial whether or not I chose 400 ppm or 410 ppm, just so long as the comparison was the same for both groups. The probability that these two samples could be taken from the same group is 92.1 percent. In other words, there was no significant difference in the calcium concentrations in the systems of aquarists who tested for calcium and those who did not. Using a test kit did not allow the aquarists who used it in this study to have tanks with a calcium concentration that was any closer to NSW than did those aquarists who never tested for calcium.

First and most importantly, aquarists must keep in mind that while groups of these elements are acting similarly, each is being metabolized or processed in its own unique fashion. One simply cannot speak of "trace element usages" in a meaningful manner. The element in question must be specified. Second, it is obvious that the whole story is not yet told. While we now have information about what is in the foods, and the system water, there are no comparable data about what is in the substrates, animals or exports. All of this notwithstanding, there is enough information to put together at least some partially reasonable explanations about the dynamics of these systems, and there is enough information to pose testable hypotheses about what is happening in our systems.

Although aquarists tend to think that all the food they feed their aquaria passes immediately or relatively soon into the water, this is definitely not the case. Some of it may effectively never become soluble, and instead will remain bound into the tissues of some organisms, or physically or chemically bound into some aspect of the substrate. Pathways leading to low tank water concentrations are seen to be active with regard to two important groups of trace elements; those which are required nutrients, and those which may become acutely toxic. For most of the trace elements in the food, the most likely way that they can become soluble is through the action of organisms. When food is added to a tank, only the liquid components are immediately soluble; the remainder becomes available to the water column only through the action of organisms.

Some essential, or presumably essential, nutrients such as iron and manganese are being added in amounts that are very high compared to what is found in natural sea water and tank water. The average daily food ration contains these materials in sufficient quantity that, with regular feeding, the water concentrations should really be off the chart. Nonetheless, they are not. Organisms are likely taking up these materials as soon as they become soluble, and they will effectively remain bound into organisms indefinitely.

Some very nasty and toxic materials such as cadmium and mercury are likely being bound by organisms into protein complexes called metallothineins. Many invertebrates bind toxic materials such as heavy metals into these and structural proteins (scleroproteins), and in doing so, they render the toxic material insoluble, and safe (for some examples of this see: Cherian, et al, 1994; Dallinger, 1994; Cosson, & Vivier, 1997). Purely inorganic processes may also work to cause the precipitation of many of these elements. Although these are all natural processes, they are potentially troublesome. Scleroproteins, metallothineins and inorganic precipitates may accumulate in a system, and there is the possibility that they may become toxic at some future time. Another possibility is that these toxic materials are being exported in some manner, either directly by exiting the tank in some filtrate, or by being bound into or onto some organism that leaves the tank. At present, we have no way of distinguishing between these pathways, or estimating their importance. Presently, I am conducting an analysis of tank exports which may help in elucidating the relative importance of these exports in maintaining low abundances of toxic trace metals.

Finally, a few trace elements, materials like titanium and tin appear to very soluble and present in concentrations far above those in NSW. The extent to which these elements may cause problems is simply unknown, but when a metal such as tin reaches concentrations 200,000 times normal, it must be considered to be a potential hazard.

Aquarists have used water changes for decades as a means of reducing the build up of toxic materials, but as we can see from Table 4, water changes only slightly slow the accumulation of many of these trace elements. Initially, much of the mass of these potentially toxic materials enters the system as components in the artificial sea water mixes. As time passes, however, the contribution from foods becomes quite significant, particularly after a couple of years.

The short answer to the question above is, "No, not with any precision." The long answer is, "Yes, we probably will be able to get a reasonable handle on the processes involved, and that will allow us to modify our tank husbandry accordingly." This study and the preceding food and additive study provide ways to assess what is in an average tank, and what is added to it when we feed (Shimek, 2001; 2002a, 2002b; present article). I will be doing another study on some of the tank exports. When the results from all of these data are considered together, it will become possible to estimate, in some cases quite closely, I hope, the movement of materials through an aquarium as well as determining which elements are not moving through the aquarium and building are up in it.

A couple of things are pretty certain, however; the first is that it is unlikely that we need any sort of additives to our tanks. With a normal feeding program the amount of trace elements entering the systems is really very high. Furthermore, they are entering as food, which is the most useable pathway. The second is that hobbyists probably won't be able to test in a cheap and meaningful way for any of the materials. This study showed that "the hobbyist-calcium test kit combination" is not one that leads reliably to any meaningful answer. Unless either better tests or better hobbyists are developed, it is unlikely that inexpensive test kits will provide any real tracking of any of the trace materials in our tanks.

If you have any questions about this article, please visit my author forum on Reef Central.