For marine benthos with planktonic larval stages, demographic metapopulation and empirical metacommunity studies are scarce. Asink nature was shown for a local intertidal sandflat population of the gastropod Umbonium moniliferumin an estuarine to coastal oceanic region in Amakusa, Kyushu, Japan. The gastropod has a 3-day planktonic larval duration (PLD). It inhabited 11 sandflats, which were shared by the ghost shrimp Nihonotrypaea harmandi. The shrimp is a burrow-dwelling bioturbator with a 30-day PLD. On one sandflat (hereafter U0) in Amakusa, the density and distribution of the two species was monitored from 1979 to 2014. All their main local populations were surveyed in 1998. In July 1979, the gastropod and shrimp were distributed on the lower and upper zones of U0 separately, with mean densities of 1,740 and 170 ind/m2. The gastropod population declined to extinction in 1986, associated with the increase in shrimp density to 440 ind/m2 and distribution to the lowest shore. Shrimp sediment destabilization was most likely to inhibit gastropod recruitment. The gastropod population began to recover in 1997 owing to shrimp population shrink. The regional survey results showed that (1) eight local populations (U0–U7 sandflats) of the two species along a 25-km shoreline of Amakusa, presumably, constitute each substantial metapopulation, and (2) their largest populations were at the shoreline opposite ends. These features would be generated by the interaction of speciesspecific larval dispersal traits and shoreline physiography. The threshold shrimp density of 160 ind/m2 precluding gastropod recruitment was detected from U1–U7. In 2001 and 2009, gastropod population density on U0 reached respective peaks comparable to that in 1979. This population fluctuation was due to varying recruit abundance, which corresponded well with the year-to-year change in shrimp population density across that threshold. The recolonization of the U0 gastropod population ought to be made by allochthonous larvae from the U1–U7 populations, especially U6–U7, suggesting a rescue effect in the source—sink relationship. The degree of self-persistency of the U0 gastropod population was estimated by calculating net reproductive rate (R0) based on the parameters about pre-settlement larval return-home and survival rates and post-settlement juvenile and adult survival rates. The availability of data set restricted the estimation to potential R0 for four year-segments, with each 1-y parameter set extrapolated to the gastropod 4-y life span. On U0 in 1979 to 1981, despite with minimum shrimp influences, the gastropod population could still be a demographic sink probably due to a limited larval return-home rate and due to a low post-settlement survival rate of juveniles caused by epibenthic predation or storm for their first 1 y to reach sexual maturity.