||We investigate the time scales of magma genesis, melt evolution, crystal growth rates and magma degassing in the Erebus volcano magmatic system using measurements of 238U–230Th–226Ra–210Pb–210Po, 232Th–228Ra–228Th and 235U–231Pa–227Ac. These are the first measurements of 231Pa–227Ac in volcanic samples and represent the first set of data in a volcanic system to examine the entire suite of relevant 238U, 235U and 232Th decay series nuclides. Our sample suite consists of 22 phonolite volcanic bombs, erupted between 1972 and 2005, and five anorthoclase megacrysts separated from bombs erupted in 1984, 1989, 1993, 2004 and 2005. The 238U–230Th, 230Th–226Ra and 235U–231Pa systems are uniform over the 34 years examined. The anorthoclase megacrysts and phonolite glasses show complementary 226Ra/230Th disequilibria with (226Ra/230Th) ?40 in the anorthoclase and ?0·75 in the phonolite glass. In all samples, (210Pb/226Ra) is in radioactive equilibrium for both phases. In two phonolite glass samples (227Ac/231Pa) is unity. For the phonolite glasses (228Ra/232Th) is in equilibrium, whereas in the anorthoclase megacrysts it is significantly greater than unity. Instantaneous crystal fractionation, with magma residence times greater than 100 years and less than 10 kyr, can account for the measured 238U–230Th–226Ra–210Pb and 235U–231Pa–227Ac. However, the significant 228Ra/232Th disequilibria in the anorthoclase megacrysts preclude this simple interpretation. To account for this apparent discrepancy we therefore developed an open-system, continuous crystallization model that incorporates both nuclide ingrowth and decay during crystallization. This open-system model successfully reproduces all of the measured 238U and 232Th disequilibria and suggests that the shallow magma reservoir at Erebus is growing. The implication of this modeling is that when the time scale of crystallization is comparable with the half-life of the daughter nuclide of interest (e.g. 226Ra) the simple isochron techniques typically used in most U-series studies can provide erroneous ages. The observation that (210Pb/226Ra) and (227Ac/231Pa) are in radioactive equilibrium suggests that the residence time of the magmas is gt;100 years. When considering the effect of 222Rn degassing on 210Pb/226Ra, the data indicate that the majority of magma degassing is deep and long before eruption, consistent with melt inclusion data. Additionally, for the 2005 lava bomb, whose eruption date (16 December 2005) is known explicitly, 210Po was not completely degassed from the magma at the time of eruption. Incomplete degassing of 210Po is atypical for subaerially erupted lavas and suggests that the Erebus shallow magma degasses about 1% of its Po per day. The combined 238U and 232Th data further indicate that the pyroclasts ejected by Strombolian eruptions at Erebus have compositions that are close to what would be expected for a near-steady-state system, reflecting inmixing of degassed magmas, crystal fractionation, and aging.