The carrier protein for the immunization was bovine serum albumin, horseradish peroxidase conjugates were employed as enzyme tracer. For the three antisera, the optimization of detection limits and test sensitivities was our first consideration. For the mecoprop and 2,4-D antisera, the strongest influences were the pH value and the ionic strength, as much as the use of enzyme tracers with lower affinities. The MCPB antiserum reacted with 2,4-DB [4-(2,4-dichlorophenoxy)butyric acid] with equal specificity, either could be detected at 0.02 pg/L (8Oy0 B/Bo), middle of the test (50 B/Bo) lying at 0.1 pg/L. The detection limit with mecoprop antiserum was optimized to 0.02 pg/L, the prescribed limit for drinking water of 0.1 pg/L ling at 60% B/Bo. The strongest cross-reactivity was found for mecoprop methyl ester. MCPB, 2.4-DB, and dichloroprop have crossreactivities of 50%, 6.7%, and 6.3%, respectively. The 2,4-D antiserum reacts less sensitively with 2,4-D, the detection limit being 0.4 pg/L. The 2,4-D isooctyl ester and 2,4-D methyl ester demonstrate as cross-reacting compounds high cross-reactivites of 3630% and 2230%. The cross-reactions of the compounds 2,4-DB, MCPB, and MCPA [(4-chloro-2-methylphenoxy)acetic acid] lie at 520'%, 69%, and 41%. 100 ground water tests were spiked within laboratory. All positive samples were correctly identified. Falsely negative results did not appear.