Bioconcentration curves of 2,4-dinitrotoluene(2,4-DNT) in carps (whole fish, liver, intestine and muscle) were investigated using semistatic system. For whole fish, its curve could be described as a gentle peak which began with a rise in concentration to summit or steady state, then declined and reached lower level followed by another steady state. For liver and intestine, their curves both contained two successive peaks, with the second peak followed by slight fluctuation. Bioconcentration factors of 2,4-DNT in whole fish during the first and second steady state were 9.15 and 4.15,(97.86 and 44.39, based on lipid content), respectively. By logarithmic plotting, two straight-lines with different slopes(3.6 and 0.1 d-1) were measured for elimination. According to peaky curves of 2,4-DNT in whole fish, liver and intestine, smaller BCFs than calculated BCFs based on the regression equations for inert chemicals, and large rate constants of elimination, biotransformation was inferred to have happened in tissues such as liver, intestine, and other tissues. Two metabolites were separated from liver and identified as 4-amino-2-nitrotoluene(4A2NT) and 2,4-diamino-toluene(2,4-DAT) on HPLC, their retention times were 23.1 and 8.8 min, respectively. In bioconcentration test of 2,4-DNT in liver, two metabolites and parent were determined at the same time at intervals, higher concentrations of 4A2NT and 2,4-DAT were found when level of 2,4-DNT declined. Such results demonstrated our inference that metabolism caused the declines in bioconcentration curves. A one-compartment model was set up to simulate the bioconcentration, in which biotransformation adhered to Delayed Enzyme-Catalytic Logarithmic Kinetics. Good fit of model curves with measured values could be observed.