Sinapine leakage to detect seed germination potential on a single-seed basis in Brassica has been developed as a rapid test. In this test, sinapine leakage predicts that a seed is non-germinable; however, the major source of errors in this method are false-negative (F–)—i.e., the method predicted a seed was germinable because the seed did not leak, and it did not germinate. The sinapine leakage index (SLI) was used to asses the F– for any seed lot by dividing the number of non-germinable seeds that leaked sinapine by the total number of non-germinable seeds. Seed lots including cabbage, cauliflower, and broccoli (B. oleracea L., Captitata, Botrytis, and Italica groups, respectively) were used to examine the F–. The leakage rate as measured by T50, the time for 50% of heat-killed seeds to leak, was linearly correlated to SLI. Cabbage seeds were viewed by scanning electronic microscopy and leaking non-germinable seeds either had cracks or were shrunken. NaOCl pretreatment has been found to increase the rate of sinapine leakage and SLI. The mode of NaOCl was due to high pH altering the seed coat permeability. Chemical analysis was conducted on isolated seed coats for pectin, tannins, hemicellulose, cellulose, phenolic lignin, and cutin. It was found that the higher SLI (more permeable) lots contained lower amounts of cutin, suggesting that cutin may restrict the diffusion of sinapine through the testa.