Scinto et al (Science 1994; 266: 1051-4) observed that a single drop of 0.01% tropicamide elicited more than a 13% pupil dilatation in 19 individuals with probable or possible Alzheimer's disease (AD) and 3 cognitive abnormal elderly without dementia, but not in 30 out of 32 normal elderly and 3 out of 4 patients suffering from other dementias. It could be an easy and bloodless test to help in diagnosing AD. We have administered the same test to 10 patients with probable AD (NINCDS-ADRDA criteria) and twenty 40-90 year-old control subjects (10 of whom were sons or daughters of AD patients and 10 without a family history of dementia). The researcher does not know which group the subject belongs to. He administers a single drop of 0.01% tropicamide in eye and one drop of 0.45% normal saline solution in the other (without knowing the contents of either vial) and measures the pupil diameter by means of a Goldmann pupilometer in basal condition and 10, 15, 25, 30, 35, 45, and 55 minutes after. The results show that it is necessary to measure the pupils at least between the minutes 25 and 55 to detect tha maximum pupil dilation in every case. The cutoff point to consider the result positive must be located between 43 and 50%. If we establish the cutoff point in 50% of pupil dilation, 90% of AD patients and 35% of control subjects show a positive response. There was not a statistically significant difference between both control groups. Our results from this test show a sensitivity of 90% and a specificity of 65%. The positive responses of some control subjects may express a weak specificity, or perhaps they mean that we have a marker of the pre-clinical stage of the disease before us.