Spherical alginate gel beads containing pindolol were prepared using three types of sodium alginate with different molecular size. The rate of gelation of sodium alginate in calcium chloride solution was in the range of 1.0 to 1.3 h-1 among the used three alginates, but the amount of water squeezed from the alginate gel beads during gelation increased from 5 to 40% with increasing molecular size of the alginate. The beads prepared were similar in diameter (1.2 mm after drying), weight (0.9 mg/bead), calcium content (27-29 micrograms/bead) and pindolol content (40-45%). Pindolol was rapidly released from all the alginate gel beads at pH 1.2 owing to the high solubility of pindolol, in spite of non-swelling of beads. On the other hand, pindolol release from alginate gel beads at pH 6.8 was dependent on the swelling of the beads and was significantly depressed compared to drug powder. Interestingly, the release rate of pindolol and the swelling rate of beads were markedly slow for gel beads prepared by low molecular size alginate. However, when the alginate gel beads were administered orally to beagle dogs, the serum levels of pindolol showed sustained-release profiles, depending on the molecular size of the alginate. The in vivo absorption of pindolol from alginate gel beads did not reflect their in vitro release profiles, because of a physical strength of beads in the intestinal tract. Furthermore, the in vivo and in vitro release of pindolol from alginate gel beads were compared with a commercial sustained-release tablet, Carvisken showed a rapid release of 50% of content in pH 1.2 fluid and residual 50% of pindolol were easily dissolved at pH 6.8. Although the release characteristics of pindolol from Carvisken and the alginate gel beads were completely different, the serum levels of pindolol in human volunteers were comparable.