“…In threeand four-dose assays, with 12 and 16 tubes respectively, equally divided between standard and unknown, a factorial design (6) allowed the use of factorial coefficients, simplifying tests for validity and calculations of potency and confidence limits as discussed (8,9). Two such assays yielded the following: common slope bc, 8.7 and 9.7; confidence limits, 88 to 114% (12 tubes, 2X3 assay) and 94 to 106% (16 tubes, 2X4 assay).…”