Recently, a population of neurones was discovered in the monkey's ( Macaca nemestrina) ventrolateral part of the pre-motor cortex (area F5). It is specialised for recognising object-oriented actions, regardless of whether these actions are performed or observed by the monkey. The latter observation led to the term mirror-neurones, and because these cells respond to both observed and executed actions, it seems likely that neurones of that type became co-opted during hominid evolution to serve the imitative behaviours that are so prevalent in our species. There is recent physiological evidence that Broca's area, the human ( Homo sapiens) homologue of monkey's area F5, is involved in the imitation of finger movements. However, concluding that human imitation is based on a mirror-neurone system is premature, because: (1) imitation in monkeys does not reach the same level as in humans or apes and (2) monkeys' mirror-neurones are specialised for object-oriented actions. This specialisation has not yet been demonstrated in adult humans. We investigated the role of objects in human imitation behaviour in a response time experiment. Subjects had to imitate downward movements of an index finger. In one condition, the observed finger touched one of two dots either ipsi- or contralaterally. In the other condition, the very same movements had to be imitated. However, there were no dots on the table. The presence of dots had a decisive influence on error patterns and on response times, but did not influence the movement proper. Dots specifically reduced the onset latency of ipsilateral finger movements and they specifically increased the use of the wrong finger, when contralateral movements were required. In general, results showed that objects also drive human imitation behaviour. Hence, it is very likely that imitation emerged from the mirror-neurone system of the common ancestor of monkeys and humans.