Factors controlling the development of stratigraphic architecture at passive continental margins were evaluated quantitatively with reference to the New Jersey Oligocene. The following were important in the example studied: 1) progradation of siliciclastic sediments across the inner part of an existing starved carbonate-dominated shelf; 2) generally low siliciclastic flux, with in situ production of authigenic glauconite, especially during times of transgression; 3) a location landward of the hinge zone of the passive margin, with slow tectonic subsidence augmented by compaction and sediment loading; 4) low to moderate amplitudes and rates of eustatic change (10-50 m over spans of ~1-2 m.y.); and 5) an active wave climate that permitted efficient lateral transport and bypass of sediment. These factors led to the development in the Oligocene of a terraced shelf morphology, with a low-gradient coastal plain and shallow shelf (1:1,000; 0.06°) separated from a deep shelf of somewhat steeper gradient (1:500; 0.11°) by an intermediate slope (<1:100; 1.0°) with 20 to 50 m of bathymetric relief. Sediment accumulation was localized on the proximal part of the deep shelf, with well-developed offlap and minor onlap at sequence boundaries, and with little or no stratigraphic preservation on the shallow shelf. Incised valleys have not been recognized, and neither have lowstand units. This is ascribed to the maintenance during progradation of a rollover between the shallow shelf and shelf slope at paleodepths no shallower than 20 m, and to the absence along this part of the continental margin of river systems that might have provided point sources. Facies arrangements indicate that the bulk of each sequence belongs to a highstand systems tract, in spite of having accumulated seaward of the rollover in the underlying sequence boundary, and spanning an interval that encompassed much of the eustatic fall. Sequence boundaries formed gradually during prolonged progradation, and not as a result of short-lived subaerial exposure and incision of the shelf. This is the case in spite of rates of eustatic fall that at times greatly exceeded the local rate of subsidence. Our data cast doubt on the universal applicability of some widely held but qualitative concepts concerning the relation between sequence architecture and eustatic forcing.