The new DFG-approach (1991) for an estimation of the limit of detection and the limit of determination via calibration curve procedure (DFG-Konzept ,,Ableitung von Nachweisgrcnze und Bestimmungsgrenze nach dem Eichkurvenverfahren'', 1991) has been tested in several laboratories performing residue analysis at the Tierhygienisches Institut Freiburg (Veterinary State Institute). It proved to be applicable not only in the residue analysis of organochlorine pesticides, but of heavy metals, anabolics and chemotherapeutics as well. In comparison with the results derived from the mean blanc values' standard deviations (NG3SD = MW+3.SD, BG6SD = MW+6.SD) according to obligatory European Community's decisions, and with results of other statistical procedures the new DFG-approach leads to limits of detection (NG(DFG)) an limits of determination (BG(DFG)) being 2,5 to 3 times higher. In few cases, checking of given regulatory limits could thus be affected. Several reasons for these findings are being discussed. The position of the lowest fortification level relative to NG(DFG), as well as the width of the fartification intervall seems to be of major influence on the hight of the NG(DFG) and BG(DFG) values. With a calibration curve's intercept differing from ,,zero'', the test of a 70%-recovery should not be performed by checking the corresponding slope but by checking the recovery on each single fortification level. The time demand for application of the DFG-approach varies from 2 to 10 h per compound, depending on the analytical method. The new DFG-approach turns out to be suited as an efficiency test of an analytical procedure, delivering information which cannot be gained by just measuring blanc values or from fortification experiments on a single fortification level. From the present results, additional standardizing measures seem to be necessary to improve the comparability of the derived parameters and therewith achieve a general applicability of the procedure.
