Contrary to predictions based on earlier observations, it is now found that label from DL-[6-14C]lysine and from [6-14C]-Δ1-piperideine does not enter the three C5 segments of lupanine with equal efficiency. One of the C5 segments is labelled at a significantly higher level than the other two C5 segments, which show identical levels of labelling. A new biogenetic model is proposed, which is consistent with these new findings as well as with all other regio- and stereochemical evidence on the mode of incorporation of precursors that is available.