Various mutations in the amyloid protein precursor and presenilin genes can lead to early onset, autosomal Alzheimer's disease. A series of mis-sense mutations (with one exception) in each of these genes has been shown to cause disease in a fully penetrant fashion. It has recently been shown, both in vivo and in model systems, that tissues expressing these mutations have increased production of amyloid (A beta) ending at residue 42. It has also recently been shown that this form of A beta is deposited early and selectively in the disease process and is more fibrillogenic in vitro. It is argued that these genetic and molecular biological data provide strong support for the veracity of the 'amyloid cascade hypothesis' for disease pathogenesis, and that this hypothesis offers a coherent framework for drug discovery.