During the past 2,000 yr, vertical crustal movements of as much as several meters have occurred in Naples Bay, located along the west coast of southern Italy. In this Special Paper, recent observations of Roman coastal ruins around Naples Bay have been combined with historical records and results of oceanographic surveys, geologic studies, and repeated geodetic surveys to determine the history, and identify the possible causes, of these movements. Where it is possible to determine the position of ancient sea level along the west coast of Italy, most Roman ruins indicate a relative rise in sea level of about 0.5 m, probably caused by a global rise in sea level. Throughout most of Naples Bay, however, the relative sea level rise since the Roman Age has been about 2 m. This 2-m rise is mainly the result of crustal subsidence, probably caused by regional extension related to opening of the Tyrrhenian Sea. Interpretation of two stratigraphic layers identified in seismic reflection profiles support the recent formation of Naples Bay by crustal subsidence. Specifically, an evaporitic layer formed during the last major drying of the Mediterranean Sea 5 to 7 m.y. ago is absent within, but is present south of, Naples Bay. Also, an erosional surface, formed during the last glacial period about 15,000 yr ago, has subsided 50 m in Naples Bay with respect to its position south of France and along the east coast of the United States. Since both the Roman Age and the last glacial maximum, the average subsidence rate around Naples Bay has been a few millimeters per year. In one region along the north coast of Naples Bay, within Campi Flegrei, an active volcanic caldera forming Pozzuoli Bay, Roman ruins are submerged several meters below present sea level. Sea level measurements made within Campi Flegrei from 1819 to 1968 give an average subsidence rate of 14 mm/yr. Thermal cooling and contraction of a large magmatic body beneath this caldera have been proposed as a possible cause for local subsidence within Campi Flegrei. Calculations suggest, however, that the cooling rate of a magma body would be too slow to account for the measured subsidence rate. Alternative explanations involve removal of fluids by flow of groundwater or gas emissions or by self-compaction of a thick porous deposit. These two alternative explanations may be related: the force of gravity might be responsible for removal of fluids by self-compaction of a porous, 2-km-thick deposit known, from a Bouguer gravity survey, to underlie Campi Flegrei. Episodic uplift is clearly indicated in a small region within Campi Flegrei. A study of uplifted marine deposits showed that a small region in the center of Campi Flegrei has undergone net uplift of about 40 m during the last 4,000 to 8,000 yr. Roman ruins along this same segment of coastline have been uplifted several meters. Repeated leveling surveys have recorded two uplift episodes since 1968 that resulted in uplift of as much as 3.00 m. Because this region contains the most recent eruptive vents, including the vent for the one historical eruption in 1538, and was the epicentral area for shallow earthquake swarms from 1983 to 1984, we suggest that uplift at Campi Flegrei is caused by periodic reactivation of a shallow magmatic body, possibly by intrusion of additional magma. Further support for a magmatic source for this uplift is the rapid uplift rate of a few millimeters per day recorded in the center of Campi Flegrei during the two recent uplift episodes and the abrupt start and end to these episodes, lasting less than a few weeks. A hydrothermal mechanism probably cannot account for these uplift rates or for the rapid change in uplift rate at the start and end of these episodes.