A MULTIVARIATE BIOLOGICAL AND CHEMICAL CHARACTERIZATION OF INDUSTRIALEFFLUENTS CONNECTED TO MUNICIPAL SEWAGE-TREATMENT PLANTS

Effluents from middle-sized industries, connected to municipal treatme nt plants in two cities in Sweden were sampled daily during 1 week and were characterized chemically and biologically. The results were eval uated mainly with multivariate statistics to And relations between che mistry and toxicity. The principal component analysis (PCA) overview o f the chemical variables displayed three main clusters: nitrogen fract ions, metals, and organic parameters. An overview of the biological re sponses showed that inhibition of nitrification was not correlated to the other biological tests. Bivariate statistics found significant cor relations existing between different endpoints of Microtox(R). In the partial least squares in latent structures (PLS) models created for th e biological toxicity tests (Microtox, inhibition of nitrification, an d algal growth test with Selenastrum capricornutum), two significant g roups causing toxicity were detected, namely inorganic metals and orga nic pollutants detected by adsorbable organic halogens, chemical oxyge n demand, and total organic carbon. When characterizing these industri al effluents, the chemical determinations and the various biological t oxicity tests complement each other and none can be excluded.