Since the early 1980s, baseline, or minimum toxicity has been a binding and almost fundamental concept in aquatic toxicology. Most models for the prediction of the toxic 'potency' of compounds to aquatic organisms rely heavily on this concept. As early as the end of the nineteenth century, it was noted that there seemed to be a correlation between the narcotic potency of alchohols (the ability, or concentration needed, to induce a reversible lethargy in Rana tadpoles), and their solubility in oils (notably olive oil). This pioneering work in quantitative toxicology was done and described independently by Hans Horst Meyer (Meyer, 1899) and Ernest Overton (Overton, 1901; see also Lipnick, 1989), working in Germany and Switzerland respectively. This work, and the work by investigators such as Ferguson (Ferguson, 1939), Hansch (Hansch et al., 1962; Hansch & Leo, 1979; Hansch et al., 1995), and more recently Veith (Veith et al., 1979; Veith et al., 1983; Veith et al., 1984; Veith & Broderius, 1987), and Könemann (Könemann & van Leeuwen, 1980; Könemann & Musch, 1981; Könemann, 1981) led to the development of the field of QSARs in aquatic toxicology and the concept of baseline toxicity as we know it today.

In the late 1970s, Gilman Veith, a researcher at EPAs Duluth Environmental Research Laboratory, was confronted with the increasingly important problem of bioconcentration of aquatic pollutants in fish and of bioaccumulation of said pollutants in the food chain. From the observation that these pollutants are found mainly in the fatty or fat-rich tissues of fish and other aquatic species, he developed the thesis that bioconcentration should at least partially be correlated with the 'hydrophobicity' of compounds. This concept of hydrophobicity was developed decades earlier by pharmacologists and medicinal chemists to describe the behaviour of drugs in the body, more specifically their transport through the bloodstream and their affinity for their target receptors. They theorized that a compound's affinity for a receptor was governed by three major 'forces', viz. steric forces (shape), electronic forces (binding between polar regions of compounds and receptors) and hydrophobic forces (binding between apolar regions of compounds an receptors). Since in this notion the binding of a compound to a receptor was akin to a compound going from a dissolved aqueous state to a state not unlike dissolution in a (partially) non-polar solvent, they naturally ended up describing the hydrophobicity part of the drug-receptor interaction with a partition coefficient, the distribution ratio of a compound over two immiscible solvents.

It was one of the accomplishments of Corwin Hansch and Albert Leo that they decided upon the octanol/water partition coefficient to mimic the 'hydrophobicity' part of their drug affinity model, and consequently measured and recorded countless 'Pow' values for widely different compounds, many of them drugs. Through this effort, they established a de facto hydrophobicity standard that is still widely used.

It was this de facto standard, in fact the only standard around in 1975, that inspired Gilman Veith to try whether he could correlate the bioconcentration and bioaccumulation behaviour of compounds with this Pow, theorizing that if pollutants were mainly found in fatty tissues, their accumulation might be governed by a distribution over water (the aquatic environment) and fish lipids (the ultimate storage compartment for the accumulated pollutants). This research led to one of Veith's landmark publications, the 1979 paper "Measuring and estimating the bioconcentration factor of chemicals in fish", J. Fish. Res. Board Can. (36), pp. 1040-1048, by Veith, Defoe, and Bergstedt. In this paper Veith and his coinvestigators showed that indeed the bioconcentration behaviour of many industrial pollutants was directly correlated with their octanol/water partition coefficient. This was a major breakthrough in those days, since it meant that for many compounds, environmental behaviour could be predicted from a simple physicochemical parameter. Of course there were, and are many confounding factors, the most important, but by no means the only one, being the influence of biotransformation on the bioconcentration behaviour of compounds.

Veith, and independently Hans Könemann of Utrecht University, the Netherlands, then went on to show that not only the bioconcentration behaviour of many chemicals was correlated with their octanol/water partition coefficient, but also their toxicity to fish and many other aquatic organisms, such as daphnids and algae. They introduced the concept of minimum, or baseline toxicity, stating that if for the relatively inert compounds they were studying at the time, the toxicity to fish was related to their hydrophobicity, then that should mean that all compounds are at least as toxic as their hydrophobicity, or bioconcentration behaviour, dictates. In other words, in the absence of all other toxic mechanisms, compounds are at least as toxic as determined by their bioconcentration behaviour. This concept opened the way for predicting the (baseline) toxicity of a multitude of aquatic pollutants through the QSAR equations developed by Veith, Könemann, and others.

These concepts ‹bioconcentration being dependent on hydrophobicity, and toxicity being dependent on bioconcentration‹ led investigators like McCarty, Opperhuizen, and Hermens ( McCarty, 1987; McCarty et al., 1993b; McCarty et al., 1993a; Opperhuizen et al., 1988; Hermens et al., 1985; Hermens, 1989) to formulate a theory of lethal body burdens, or critical body residues, that states that for baseline toxicity compounds, there is one, compound-independent, internal concentration that determines the toxic effect. In fact, what they were saying was that there was one lipid-tissue-concentration for baseline chemicals that defined lethality (or other toxic effect level).

The rationale behind this lethal body burden, or critical body residue concept is that the combined lipid tissues in an organism (the 'lipid fraction') are in 'thermodynamic equilibrium' with their surroundings, not a far fetched assumption for aquatic organisms swimming in water with dissolved organic pollutants. This lipid fraction constitutes a sort of 'surrogate' target organ. Building on work by Ferguson, and Franks and Lieb (Franks & Lieb, 1990), among others, the still prevailing theory is that the anaesthetic (neurodepressing) effect of many chemicals, including the gases used for inducing general anaesthesia, but also 'inert' solvents and comparable compounds, is due to the dissolution of their molecules in the cell membranes (or association with certain apolar structures in these cell membranes) of CNS or peripheral NS cells up to a certain threshold concentration. If an organism is in thermodynamic equilibrium with its contaminated surroundings, and if the polarity of these cell membranes (or membrane structures) is similar to the polarity of the averaged lipids in an organism, then there should be a correlation between the concentration attained in these lipid tissues, and the concentration attained in the membranes. In other words, if a certain threshold concentration is reached in the actual target for these narcotic chemicals, a similar, but not necessarily equal 'critical concentration' is reached in the bulk lipids of an organism. For a thorough overview of this subject, see van Wezel and Opperhuizen (1995).

We can therefore formulate a generalized theory of the aquatic toxicity of inert compounds. This theory states that any compound that enters the body of an aquatic organism, be it across the gills, or via food, in fish, or by direct diffusion over the cuticula or cell wall/membrane in the case of daphnia or algae, will be distributed over the different 'structures' in this organism as determined by its hydrophobicity. Eventually, a thermodynamic equilibrium will be reached in which the concentration of the compound in essential lipoid tissues, such as CNS membranes, is directly related to the external concentration and the compound's hydrophobicity, such as its log Kow. If this equilibrium concentration exceeds a threshold value, a toxic effect will be observed (Hermens, 1989; Verhaar, 1995; Verhaar et al., 1995).

Of course there are many additional mechanisms and confounding factors that may enter the equation, so to say, such as metabolism and extretion (a detoxification pathway), metabolism into more toxic species, specific toxicity, etc., but the fact remains that this theory predicts that compounds will, generally, be not less toxic than determined by their hydrophobicity. This theory was the basis for the classification system described by Verhaar, van Leeuwen and Hermens (Verhaar et al., 1992), in which they presented a rule based system to classify organic compounds into four toxic classes, viz. inert compounds (real 'baseline' compounds, class 1), less inert compounds (mainly phenols and anilines, compounds that appear slightly more toxic than baseline predicts, but still follow all the other basic rules, class 2), reactive compounds (class 3), and compounds that act through specific, receptor-based mechanisms (class 4). They also gave ranges of so-called Toxic Ratios, that roughly indicated what the aquatic toxicity of a compound should be, based on its predicted baseline toxicity and its class membership. These TR ranges are:

• Class 1: 1
• Class 2: 5-10
• Class 3: 10-104
• Class 4: 10-104

Verhaar et al. (2000) showed that comparison of independently determined 96h LC50 values for fish generally coincided quite well with the predicted toxicity ranges based on a baseline toxicity QSAR equation and the above-mentioned class-specific ranges.