Abstract : The principal flows in and out of the East China Sea (ECS) are through gaps in the Ryukyu Ridge. Since about 20 Sverdrups (Sv) of Kuroshio mean flow enters and exits through two of them, those two are especially well known: the East Taiwan Channel (sill depth 775 m) at the ridge's southwestern end, and the Tokara Strait (sill depth 690 m) near its northeastern end. But the deepest channel connecting the ECS to the surrounding ocean is near the mid-point of the ridge. This is the Kerama Gap, about 50 km wide with sill depth 1050 m. Little is reliably known about the flow through the Kerama Gap. Mean-flow estimates from measurements and models range from 0.9 Sv out of the East China Sea to 7.2 Sv into it. Knowledge of the flow variability is even more uncertain, but there is evidence of transport variations with magnitude of a few Sv caused by impingement of Philippine Sea eddies from the east at intervals of a few months. Our main purpose in this project is to make a reliable determination of the varying flow through the Kerama Gap. On time scales ranging from two days to two years, our main objectives are to test the following three hypotheses: (1) increase in transport through the Kerama Gap is associated with (a) an increase in ECS-Kuroshio transport across the PN-line, north of Okinawa, and (b) a decrease in Ryukyu Current transport east of Okinawa about two months earlier; (2) the arrival of anticyclonic (cyclonic) eddies at the eastern side of the Kerama Gap is associated with an increase (decrease) of transport through the Gap; and (3) variations in wind stress over the local region cause variation in the flow through the Kerama Gap.