Introduction, Snafus, and Corrigenda

After the publication of the first article in this series about what is in tank water (Shimek, 2002), I received helpful commentary by Tatu Vaajalahti and Randy Holmes-Farley. They pointed out that my listing of trace element compositions was significantly out-of-date and that many new data had been gathered and that the precision of the data collection and analyses had been significantly refined. I was sent a copy of a more up-to-date listing of trace element concentrations in sea water from a recent text on chemical oceanography (Pilson, 1998). The table listed the range of values found, as well as the average value.

I converted those data from the standard molar concentrations used by chemists, to the "parts per unit" concentrations generally used by aquarists. A comparison of the previous values, used in the first article, and the new values is given in Table 1. Probably the first thing to notice is that many of the accepted concentration data have changed, in some cases very significantly. While a few elements were found in slightly higher concentrations, many more changes resulted in values significantly lower than were given in my earlier table. Often these changes resulted in differences of several orders of magnitude.

These changes were brought about by changes of analytical techniques and equipment. While oceanographers had previously been able to determine the presence or absence of some material at a given threshold value, they were unable to precisely determine the concentration, which may have a very small fraction of the threshold reading. The data were then tabulated as the threshold value, rather than the actual value. The newer methodologies allowed a much more precise determination of the actual concentrations.

In effect, the use of the newer data changed the base line of the evaluations. In many cases, these changes lowered the baselines, which resulted in significant increases in the proportion of tested elements relative to their actual concentrations in natural sea water (NSW). The revised results of these changes in proportion are shown in Figure 1. The observed range of aquarium values are plotted as a line, and the average value is shown as a tick mark within that range. To be able to encompass the values within a single graph, I had to use a logarithmic scale for the proportions, so the average values are graphically displaced from the center of the range.

Figure 1. Average tank concentrations of those elements whose concentrations were above the detection limits of the test procedure.

In the examinations of Figures 1 and 2, it is important to realize that the horizontal lines represent different things. In Figure 1, the values represent the actual concentrations of the material in ppm, whereas in Figure 2 they represent relative values compared to normal. So, in Figure 2, that a value of "1.00" from the tests indicates an average proportion in the tested tanks that is the same as the average NSW concentration. Similarly, a value crossing the line for "0.1" means the tested value was one tenth the value of the average NSW concentration and a value crossing the line for "100" is one hundred times the NSW value. In Figure 1, these values represent the actual concentrations. Additionally, if you evaluate the differences between the graphs illustrating last month's article and this one, it is important to realize that the observed changes do not reflect any change in the actual values found in the aquaria, but simply are a result of changes in the accepted values in the NSW concentrations.

Prior to trying to assess why the trace element concentrations in these tanks are different from those in NSW, it is also important to consider how we perceive them as different. The samples for these studies were evaluated by one analytical method. Another methodology might give somewhat different results. The methodology used by the lab I chose is called "Inductively Coupled Plasma Emission Spectroscopy." The methodology is reasonably sensitive and may be used for assessing a large number of elements. It is commonly used in environmental testing and assessment, and is relatively inexpensive, as each sample costs less than $200 to process. However, as in all methodologies, there were trade offs. In this case, the trade off came in the assessment of several elements where the detection limits of the test are above the levels commonly found in NSW.

Although the samples were analyzed for Beryllium, Chromium, Cadmium, Iron, Lead, Manganese, Mercury, Selenium, Silver, and Yttrium, none of these elements were detected in the samples; at least in part because the tests simply were not sensitive enough to detect them at normal and near normal concentrations. Most of these elements are quite toxic to marine organisms, but are normally found in very low concentrations and are probably of no consequence to aquarists. Iron and Manganese, however, are biologically active and important for many organisms, and it would have been preferable to have some idea of their concentrations. Nonetheless, neither of these elements was detected in any of the samples. It is important to note, this lack of detection does not mean that the materials were absent, just that the test could not detect them. Those elements will not be discussed further.

In some other cases, as illustrated by Iodide and Tin, where the detection limit for the test is above the NSW concentrations, the aquarium levels detected were all so significantly elevated over the normal NSW levels that the test was able to detect them without a problem. For example, the detection limit for Tin was on the order of 10,500 times greater than the normal level found in sea water. However, the tank concentrations for tin averaged a whopping 200,000 times the level in sea water, so the test had plenty of latitude in which to work (Figure 2).

Iodine presents a special case. Although the initial documentation from the lab indicated that the test was for Iodide ion, a discussion with the laboratory director indicated that the procedure tests for total Iodine, not just Iodide ion. Even though the detection limit for the test was above the NSW level of Iodide, it was below that for total Iodine and well below the tank levels for this material. This was another case where both the upper and lower limits of the tank concentrations were well above both the detection limits and the NSW concentrations.

After examination of these data, questions should arise as to their significance. In effect, what can we learn from such data? Several trace elements are found in elevated concentrations in aquarium water (Table 2; Figure 2). Some of these metals have extremely high concentrations relative to NSW; tin has already been mentioned as having concentrations over 200,000 times above normal, but Thallium, Titanium, Aluminum, Zinc, Cobalt, Antimony, and Copper all have concentrations of over 95 times normal. Conversely, of the detected elements, relatively few are substantially lower than normal. Although Sulfur, Boron, Strontium, Silicon and Vanadium had lower tank concentrations than in NSW, only Vanadium was present at less than about 50 % of normal levels.

In the remainder of this article, I will examine the abundance patterns of the detected chemicals, as well as some other factors, and try to determine if there is any easily evident reason for such patterns. Furthermore, I will try to assess the significance of such patterns and associations.