We examined the effects of ginseng total saponin (GTS) and several ginsenosides injected intrathecally (i.t.) or intracerebroventricularly (i.c.v.) on the antinociception induced by cold water swimming. The tail-flick response was used as an antinociceptive parameter. We found that i.t. injection of GTS time- and dose-dependently attenuated inhibition of the tail-flick response induced by cold water swimming, although GTS given i.c.v. had no significant effect on the latency of the tail-flick response induced by cold water swimming. To identify those responsible for antagonism of GTS against cold water swimming-induced antinociception, the effects of various kinds of ginsenosides (R(b1), R(b2), R(c), R(d), R(e), and R(g1)) on inhibition of the tail-flick response induced by cold water swimming were examined. R(b1), R(d), R(e), and R(g1) effectively attenuated the inhibition of the tail-flick response induced by cold water swimming stress. Our results suggest that GTS injected spinally, but not supraspinally, reduces the antinociception induced by this stress, and that the responsible ginsenosides against antinociception induced by cold water swimming may be R(b1), R(d), R(e), and R(g1). Moreover, the possible involvement of the opioid system in the regulation of cold water swimming stress-induced antinociception by these four ginsenosides is discussed.