Thomas, Carolyn L.2014-08-132014-08-131985http://hdl.handle.net/10919/49941The purpose of this research was to develop a test system for screening toxic substances by predicting their aquatic ecosystem effects. The system studied was a static, one liter microcosm with a diverse species assemblage. Teh microcosm was composed of biotic inoculum, chemically defined medium and sediment. The biotic inoculum cotained primary producers, grazers, carnivores and decomposers. The chemical medium used was Taub #82 modified by adding sodium bicarbonate. Three different types of sediment were studied: sand, clay and clay plus sand. Four organic chemicals: phenol, triethylene glycol (TEG), quinoline and naphtoquinone were evaluated with this test system. The toxicities of TEG, quinoline and naphthoquinone were compared for each sediment type. Toxicity was evluated in terms of the chemicals' effects on primary productivity and heterotrophic activity though other effects are also noted. The toxicity of the chemicals in this study was compared to those from other toxicological evaluations based on threshhold toxicity values (EC20 and LC50). The screening test sytem evaluated in this study did not demonstrate significanly different threshhold toxicity values than the other screening systems to which it was compared. The ranking of the toxicants based on EC20 values was different for the two ecosystem properties, net production and heterotrophic activity. Naphthoquinone concentration exhibited no correlation between ecosystems property values and therefore, could not be ranked. Phenol exhibited the greatest toxicity to net production immediately after the toxicant addition. Quinoline was most toxic to net production over the longer time scale. TEG exhibited the least toxicity to net production, however, TEG exhibited higher toxicity to heterotrophic activity than either quinoline or phenol. Although the type of sediment used in the nicrocosms did not change the relative toxicities of the chemicals, the microcosms with clay sediment always were observed to exhibit lower net production and higher variability. Nonparametric statistical analyses are recommended for microcosm studies because of the lack of normally distributed data. Confidence limits of 80% are recommended because of the need for biologically conservative estimates of ecosystem toxicity.xii, 185 leavesapplication/pdfenIn CopyrightLD5655.V856 1985.T565Organic compounds -- ToxicologyToxicity testingAquatic ecologyDissertation