To our knowledge, there are few recent French studies looking at the cognitive assessment of people aged over 80. Yet this population, which currently accounts for 6% of the French population and is expected to double by 2080, is particularly affected by neurocognitive disorders, the prevalence of which increases sharply with age. The influence of advanced age on episodic memory performance is frequently reported in the literature, with an accelerated decline from the age of 65 onwards. There is also a positive effect of educational level on cognitive performance in old age, as well as a gender effect, with several studies describing better performance by very old women in verbal episodic memory. To assess episodic memory, the 16-item Free Recall/Indicated Recall test (RL/RI-16 items) and the DMS48 forced-choice visual recognition test are commonly used in clinical practice. They form part of the neuropsychological assessment tests that are essential for the evaluation of cognitive functions and help to direct the clinical examination towards additional tests. French normative data have been published for these two tests but, with the exception of the study by Amieva et al (2007), the age groups of elderly subjects are broad (see for example Van der Linden et al, 2004) or very small for very elderly subjects (Barbeau et al, 2004). The aim of our work is to study the effect of gender and education variables for these two tests, in order to establish norms that can be used in clinical practice. To this end, we studied the RL/RI-16 and DMS48 performances of 107 healthy subjects aged over 82 from the FIBRATLAS cohort. The results confirm the importance of using norms corresponding to the age group for old age, and of taking into account gender and level of education for the RL/RI-16 and level of education for the DMS48.