The replicate sample instrumental neutron activation method was optimized and used for the determination of selenium in foodstuffs. The method was reliable, yielding accurate results. Lower detections limits were obtained after each successive irradiation. Different irradiation conditions were used depending on the type of sample. For samples with higher selenium contents (meat, fish, eggs), the measured selenium in the first replicate is in all cases larger than the detection limit, but a better accuracy was obtained with a larger number of replicates (2-3 replicates). For samples with extremely low selenium contents (vegetable samples), at least seven replicates were necessary to obtain a concentration value two times larger than the detection limit.