The development of self-disclosure within an acquaintance exercise was examined. A sample of 80 subjects was assigned to same-sex or mixed-sex dyads. In half of the dyads, one member was asked, through a short statement in the written instructions, to encourage maximal disclosure in his/her partner. Dyad partners took turns discussing self-selected topics from a 72-item list of intimacy-rated choices. The resulting design allowed for testing the effects of sex-pairing, leader instructions, and subject sex on the intimacy of disclosure and the extent of reciprocity. Sex-pairing had no effect upon any measures. Leader instruction dyads showed significant linear and cubic trends in increasing intimacy over time, while the non-leader instruction dyads showed no temporal trends. Male subjects tended to disclose at levels somewhat higher than females. Generally, subject reciprocity was found unrelated to experimental factors, although one significant interaction between subject sex and leader condition is discussed. (Author)