Aluminum (Al) is unquestionably neurotoxic in both experimental animals and certain human diseases. Minute quantities injected intracerebrally into rabbits will induce severe neurological symptoms and neuropathological features of neurodegeneration. Hyper-aluminemia often develops in patients with renal failure being treated with intermittent hemodialysis on a chronic basis, and in severe cases results in an encephalopathy. Uremic adults and premature infants not on dialysis treatment also can develop encephalopathy due to Al toxicity, as is the case when large amounts of alum are used as a urinary bladder irrigant. There are many other examples of Al-induced neurotoxicity; however, the question as to whether Al presents a health hazard to humans as a contributing factor to Alzheimer's disease is still the subject of debate. Several lines of evidence are presented that have formed the basis of the debate concerning the possible pathogenic role for Al in Alzheimer's disease. Important evidence for an Al-Alzheimer's causal relationship is the observation by laser microprobe mass analysis (LMMS) of the presence of Al in neurofibrillary tangles, although there are conflicting data on the extent of the Al deposition. The relatively poor sensitivity of some of the analytical instruments available for these challenging in situ microanalyses could explain the discrepant results, although LMMS and perhaps secondary ion mass spectrometry (SIMS) appear to be sufficiently sensitive. Harmonization of the techniques is an essential next step. There is new evidence that exposure to Al from drinking water might result in cognitive impairment and an increased incidence of Alzheimer's disease. However, these epidemiological studies have inherent problems that must be scrutinized to determine if an association really does exist. An understanding of a possible enhanced bioavailability of Al in this type of exposure, versus other exposures such as antacid intake or industrial exposure, needs to be considered and explored. There has been one promising clinical trial of the treatment of Alzheimer's disease patients with the Al chelator desferrioxamine (DFO). Further studies are needed, and if confirmation is forthcoming then such data could also support an Al-Alzheimer's disease link as well as suggesting that DFO offers potential as a therapeutic agent. The possibility that iron might be the offending agent needs to be considered since DFO is a very strong iron chelator. The significance of Al-induced neurofibrillary degeneration in experimental animals should be assessed especially in light of new data showing that this model exhibits abnormally phosphorylated tau protein structures in the neuronal perikarya. Thus the key questions that must be answered before it can be asserted that Al possesses causal relationship to Alzheimer's disease, are as follows and are addressed in this present discussion: (1) Are there elevations of the concentration of Al in the brains of Alzheimer's disease patients? (2) Is there a relationship between environmental exposure to Al, particularly in drinking water, and an increased risk of Alzheimer's disease? (3) Is treatment with DFO a potentially useful therapeutic approach and to what extent might beneficial effects of DFO implicate Al in the etiology of Alzheimer's disease? (4) Are there similarities between the experimental animal studies and Alzheimer's disease particularly in the development of abnormal forms of tau seen in neurofibrillary tangles? (5) Does Al promote the deposition of the A beta peptide in Alzheimer's disease? (6) Does hyperaluminemia associated with long-term hemodialysis treatment induce neurofibrillary degeneration? If the answer to each of these six questions is yes, then does this assert that Al possesses a causal relationship to Alzheimer's disease? On the other hand, must all six be met to be able to make this assertion?