Abstract: Rich circumstantial evidence suggests that the extensive behavioural diversity recorded in wild great apes reflects a complexity of cultural variation unmatched by species other than our own. However, the capacity for cultural transmission assumed by this interpretation has remained difficult to test rigorously in the field, where the scope for controlled experimentation is limited. Here we show that experimentally introduced technologies will spread within different ape communities. Unobserved by group mates, we first trained a high-ranking female from each of two groups of captive chimpanzees to adopt one of two different tool-use techniques for obtaining food from the same 'Pan-pipe' apparatus, then re-introduced each female to her respective group. All but two of 32 chimpanzees mastered the new technique under the influence of their local expert, whereas none did so in a third population lacking an expert. Most chimpanzees adopted the method seeded in their group, and these traditions continued to diverge over time. A subset of chimpanzees that discovered the alternative method nevertheless went on to match the predominant approach of their companions, showing a conformity bias that is regarded as a hallmark of human culture.