Stream snails are known to exhibit upstream migration in temperate and tropical regions (see reviews by Huryn & Denny, 1997; Kappes & Haase, 2012). This behaviour can be driven by a variety of factors depending on species and circumstances, including the search for food, predator avoidance and compensation for washout or catastrophic drift during spates (e.g. Schneider & Frost, 1986; Schneider & Lyons, 1993; Snider & Gilliam, 2008). Recolonization of upstream habitats via migration can replenish losses due to spate-induced washout (Kopp, Jeschke & Gabriel, 2001) and could be an important mechanism contributing to population persistence (Speirs & Gurney, 2001; Pachepsky et al., 2005). Upstream migration may also be a mechanical consequence of snails orientating themselves to reduce hydrodynamic drag (Huryn & Denny, 1997; Covich, 2010). Within a species, migration patterns vary in relation to individual size and growth (e.g. Schneider & Lyons, 1993; Shigemiya & Kato, 2001; Snider & Gilliam, 2008; BlancoLibreros & Arroyave-Rincon, 2009). Upstream migrations of snails in tropical and subtropical streams have been mostly documented in amphidromous Neritidae (Blanco & Scatena, 2005, and references therein), which have plankotrophic larvae that require saline conditions for development. Juveniles must therefore migrate upstream from estuaries to adult habitats where they mature and breed (McDowall, 2007; Crandall, Taffel & Barber, 2010; Gorbach et al., 2012). The caenogastropod pachychilid Sulcospira hainanensis (Brot, 1872) is widespread in tropical streams in southern China (including Hong Kong, Guangdong and Hainan) and northern Vietnam (Kohler & Glaubrecht, 2001; Yeung & Dudgeon, 2014b). Sulcospira hainanensis is not amphidromous; it spends its entire life in fresh water and larvae are incubated in a cephalic brood pouch prior to their release in the form of miniature adults (Kohler & Glaubrecht, 2001). Wet-season declines in abundance have been recorded in this species, likely due to washout during spates (as high as 85%; Yeung & Dudgeon, 2014b). Upstream migration may therefore be necessary to compensate for this downstream displacement, which might otherwise result in long-term decline in snail densities in the upper courses of streams. There are few studies of migration by non-amphidromous caenogastropds in tropical streams, but Cochliopina tryoniana (Pilsbry) (Hydrobiidae) in Costa Rican streams exhibits upstream migratory behaviour, moving at rates of up to 110 m d (Schneider & Lyons, 1993). These streams receive frequent spates in the wet season, analogous to those in Hong Kong, and we consider it likely that the upstream migration of C. tryoniana is a compensatory response to downstream washout during this period, rather than having the life-history role seen in freshwater neritids (Schneider & Lyons, 1993). We hypothesized that S. hainanensis would show upstream migrations as a compensatory response to spate-induced washout. This was tested using a year-long mark-and-recapture study of three populations in Hong Kong. The study populations inhabited second-order stream at Tai Po Kau (TPK), and third-order streams at Pak Ngau Shek (PNS) and Shing Mun (SM) in the mainland New Territories of the Hong Kong Special Administrative Region of the People’s Republic of China (Yeung & Dudgeon, 2013: fig. 1). The TPK site (elevation: 180 m) was situated at mid-altitude and other sites (PNS: 110 m; SM: 300 m) in the lower half in their respective drainage basins. These sites were thus not close to the uppermost distributional limits of snails, which allowed the detection of any active upstream migration. Snails were present both upstream, and at various distances downstream of the study reaches, and physical conditions upand downstream were similar. Hong Kong has a tropical monsoonal climate, with the wet season (May to August) in 2011 (the year when the study was conducted) receiving 42% (1,007 mm) of the mean annual rainfall of 2,399 mm, which is typical of the seasonal pattern of precipitation (Hong Kong Observatory, 2011). Spates (.50 mm rainfall in 1 d) were frequent during this period (Hong Kong Observatory, 2011). The study sites encompassed a wide range of canopy closure (PNS: 40%; TPK: 80%; SM: 90%), with consequent differences in algal food availability (Yeung & Dudgeon, 2014a). Most other physical characteristics were similar in these streams, where the waters were slightly acidic and welloxygenated (see Yeung & Dudgeon, 2013; Table 1). Therefore, the sites chosen broadly represent the stream environment that the snails inhabit in Hong Kong. Monthly collections of 80–100 S. hainanensis individuals were made from January 2011 to January 2012 (except July 2011) along a 20-m study reach at each of the three sites. They were marked individually by uniquely identifiable bee tags (Opalith Numberplates for Queens, Item 112,902, Swienty A/S Company, Sonderborg, Denmark) attached to the shell with Aron Alpha Instant Glue super glue gel (Toagosei America Inc., West Jefferson, United States). Individuals ,1.5 mm aperture width