E.M. Thurman D.S. Aga D. A. Goolsby M.L. Pomes 1998 <div class="hlFld-Abstract"><div id="abstractBox"><p class="articleBody_abstractText">Triazine and chloroacetanilide concentrations in rainfall samples collected from a 23-state region of the United States were analyzed with microtiter-plate enzyme-linked immunosorbent assay (ELISA). Thirty-six percent of rainfall samples (2072 out of 5691) were confirmed using gas chro matography/mass spectrometry (GC/MS) to evaluate the operating performance of ELISA as a screening test. Comparison of ELISA to GC/MS results showed that the two ELISA methods accurately reported GC/MS results (<i>m</i><span>&nbsp;</span>= 1), but with more variability evident with the triazine than with the chloroacetanilide ELISA. Bayes's rule, a standardized method to report the results of screening tests, indicated that the two ELISA methods yielded comparable predictive values (80%), but the triazine ELISA yielded a false-positive rate of 11.8% and the chloroacetanilide ELISA yielded a false-negative rate of 23.1%. The false-positive rate for the triazine ELISA may arise from cross reactivity with an unknown triazine or metabolite. The false-negative rate of the chloroacetanilide ELISA probably resulted from a combination of low sensitivity at the reporting limit of 0.15 μg/L and a distribution characterized by 75% of the samples at or below the reporting limit of 0.15 μg/L.</p></div></div><div class="hlFld-Fulltext"><br data-mce-bogus="1"></div> application/pdf 10.1021/es970462l en ACS Evaluation of microtiter-plate enzyme-linked immunosorbent assay for the analysis of triazine and chloroacetanilide herbicides in rainfall article