E A m h i v e s o f n v i r o n m e n t a l c o n t a m i n a t i o n a n d | oxicology© 1992 Springer-Verlag New York Inc.
Reliability of Heavy Metal Pollution Monitoring Utilizing Aquatic Animals Versus Statistical Evaluation MethodsZoltfin Mastala, Katalin V.-Balogh, and J~inos Sal4nkiBalaton Limnological Research Institute of the Hungarian Academy of Sciences, Tihany, P.O. B. 35, Hungary Abstract. Heavy metal concentrations in two snail and two fish species were used to compare the pollution of four sites on the River Zala and on the Kis-Balaton Reservoir (Hungary). The effectiveness of different statistical methods (t-and Bartlett's test, analysis of variance and analysis of covariance) were compared. The most reliable results were obtained by using analysis of covariance, with body weight as the covariate, taking the seasonality in metal concentrations into consideration. The different species showed the same relations among sampling stations if analysis of covariance was applied. It is essential to use statistical methods that account for the most important effects (c0variates); otherwise, the reliability of the information supported by biomonitors is compromised.Heavy metal contamination in the aquatic environment is usually monitored by measuring concentrations in the water, sediment and biota. Determination of metal concentrations in surface water samples is problematical because concentrations are low and display wide fluctuations. Metal concentrations in sediments are high and relatively invariant with time. However, they are influenced by the organic matter, Mn and Fe contents, pH and oxidation-reduction potential, grain size distribution, etc. (F6rstner and Wittmann 1983;F6rstner 1984), which make direct comparisons between locations difficult. The main problem of utilizing aquatic animals is that metal concentrations in organs or in tissues depend not only on concentrations in the environment, but on several factors, such as temperature (Pentreath 1976a), salinity (Unlu and Fowler 1979), pH (Campbell and Stokes 1985;Wiener 1987), organic ligands (Gjessing 1981), seasonal changes in metal accumulation (Pentreath 1976b) and organism size (Phillips 1980;Beumer and Bacher 1982).Among them, the body weight dependence was the most critical factor, and there is no simple answer for its elimination from monitoring surveys (Phillips 1980). Although, Boyden's power model (Boyden 1974(Boyden , 1977 is used primarily to describe the effect of body weight on metal concentrations, it is an empirical relationship without a clearly defined, underlying mechanism. Boyden hypothesized that the exponents of the power functions were generally constant for some speciesmetal combinations. If this were so, the effects of size on metal levels could be easily eliminated using an exponent with universal validity. This hypothesis was supported by Smock (1983), but was refuted by Cossa et al. (1980) as well as by Strong and Luoma (1981).Another alternative is the deliberate selection of a large range of sizes at all locat...