For some time, onset of the Antarctic Circumpolar Current (ACC) was considered to have caused or stabilised full Antarctic glaciation. Recently, however, the importance of the ACC in this role has been questioned. In order to understand the relationship between the ACC and Antarctic glaciation, and thence the importance of ocean circulation to palaeoclimate, we need to determine the development history of both processes. To this end, we summarise all published estimates of ACC onset. The time of onset, of shallow circulation or deep, is uncertain, whether based on tectonic studies or the interpretation of changes in the sediment record. Two potential final barriers to circumpolar flow have been identified; south of Tasmania and south of South America. The former is well constrained by tectonics and marine geology to before 32 Ma for a deep gap, with a shallow gap in place by 35.5 Ma at the latest. These ages fit nicely with the onset of full Antarctic glaciation at 33-34 Ma, although some workers question the causality. Estimates of the time of opening of the latter range widely, whether based on tectonics or sedimentary geology, from as recently as 6 Ma to as early as 41 Ma, with the gap depth uncertain also. Resolution of the tectonics-based uncertainties by additional survey being most probably both time-consuming and inconclusive, and the geological estimates being open to alternative interpretations, we define an optimal strategy for additional sampling and measurement, designed to resolve the time of onset more certainly, possibly also resolving between deep and shallow opening, and thereby constraining the ACC role. Sample sites would have to be close to likely final barriers, to avoid extraneous influence, and within modern zones of ACC influence, ideally would form a depth transect, and would have continuous, mixed terrigenous and biogenic sections. A wide range of carefully selected parameters would be measured at each.