A new experimental microculture approach was developed to investigate the creation and transmission of differing traditions in small communities of young children. Four playgroups, with a total of 88 children, participated. In each of 2 playgroups, a single child was shown how to use 1 of 2 alternative methods of tool use, "lift" or "poke," to extract a reward from an artificial "foraging" device (the "panpipes") used in earlier diffusion experiments with chimpanzees. Each of these proficient models then participated in his or her playgroup during free play for 5 days, with the panpipes available to all. Compared with a condition in which no model was witnessed, where only 18% of children successfully gained rewards and the lift technique never appeared, 66% of children in the open diffusion conditions (83% of those who attempted the task) were successful. Each of the 2 different seeded approaches initially spread strongly in their respective groups. These seeded differences eroded over time as modifications were spontaneously invented, but social learning played a dominant role throughout, with a majority of children adopting the technique they witnessed most commonly, whether initially seeded or resulting from other children's innovations. A majority of children thus fell into 1 of several categories of "follower," relying primarily on social learning, with a minority displaying 1 of several other categories of innovation. One of the techniques was modified into a distinctively different form that was then socially transmitted further, allowing us to document the microevolution of small-scale traditions in this cultural microcosm. (Contains 3 tables and 6 figures.)