One hundred and five normal subjects from 20 to 89 years old, and with a minimum of six years of formal education were selected. Subjects were divided into seven age ranges, with 15 subjects in each group. The following tests were individually administered: (1) Wechsler Memory Scale (WMS) (Wechsler, 1945); (2) Serial Verbal Learning (Ardila, Rosselli & Puente, 1994; Ardila & Ostrosky, 1991); (3) The Rey-Osterrieth Complex Figure (ROCF) (Osterrieth, 1944) (copy, immediate, and delayed reproduction); (4) Corsi's blocks (Milner, 1971) (forwards and backwards); and (5) Rivermead Behavioural Memory Test (Wilson, Cockburn, & Baddeley, 1985). This last rest may be considered as a functional or ecological memory test. All test scores declined across ages between 4.1% and 76.6%. In all the tests, excepting Digits, standard deviations increased with age. It was observed that the youngest subjects could retain about three-fourths of the information presented, whereas the oldest retained around one-third of this information. About half of the rest significantly correlated with the Rivermead Behavioral Memory Test. A factor analysis revealed four different factors accounting for 63.3% of the variance. A single nonverbal memory factor accounted for over 40% of the variance.