The compression of peat beneath its own weight (autocompaction) is discussed, and it is shown that because of this process radiocarbondated samples of salt-marsh peat or peaty sediment, other than very thin samples cut from the base of the deposit, cannot be correlated with sea level without construction of a sea-level change curve from other types of data. With rising sea level there is a maximum thickness of salt-marsh peat for any given productivity of marsh grass. Because of this limitation, most marshes older than about 5500 years B.P. have been drowned. An important effect of peat autocompaction is the intrusion of wood into older peat horizons and the juxtaposition of wood of different ages. Wood, therefore, should be avoided in the dating of peat profiles. The roots of salt-marsh plants generally descend a foot or more beneath the rhizomes, contaminating older horizons insofar as radiocarbon dating is concerned. This contamination tends to make dates from marsh samples err on the young side. Sixteen radiocarbon dates from the Boston area are used to construct a sea-level curve going back to 14,000 years B.P. Relative sea level was at +60 feet or higher 14,000 years B.P., dropping sharply to approximately -70 feet about 10,000 years B.P. From a low of -70 feet, sea level rose steadily to about -2 feet approximately 3000 years B.P. Since then sea level appears to have kept close to its present level, probably fluctuating about a foot during the course of the stillstand. A crustal movement curve, based on the relative sea-level curve for Boston and the eustatic curve, indicates that about 290 feet of crustal rise occurred between 14,000-6000 years B.P., with a maximum rate of uplift of about 0.2 foot per year at 12,750 years B.P., and that from 6000 to 3000 years B.P., crustal subsidence occurred at Boston. © 1964, The Geological Society of America, Inc.