An early proposal was that for rapid ATP synthesis by the rotational ATP synthase, a specific second site must bind ADP and P(i), and for rapid ATP hydrolysis a different second site must bind ATP. Such bi-site activation was considered to occur whether or not an ADP or ATP was at a third site. In contrast, a more recent proposal is that rapid ATP hydrolysis requires that all three sites have bound ADP or ATP present. However, discovery that one second site binds ADP better than ATP, together with other data and considerations support the earlier proposal. The retention or rebinding of ADP can explain why three sites fill during hydrolysis as ATP concentration is increased although bi-site activation still prevails.