This article begins with a review of gastric emptying, small intestine transit, and colonic transit of drug delivery systems with special attention paid to the different physiological processes involved in stomach emptying and to the cut-off size of nondigestible solids for passage through the gastroduodenal junction during the digestive phase. Then, the proposed means for prolonging the gastric residence time (GRT) of drug delivery systems are reviewed and analyzed with special emphasis on floating (F) dosage forms. The following means are discussed: the use of passage-delaying agents, large single-unit dosage forms, bioadhesive drug delivery systems, "heavy" pellets, and buoyant forms. In the section devoted to bioadhesive forms, the influence of the turnover time of the intestinal mucus gel layer on the performance of mucoadhesive preparations is pointed out to explain the poor results obtained in humans with such peroral products. The use of a specifically designed apparatus for measuring the total force acting vertically on an object immersed in a liquid is presented as a methodology for selecting optimized buoyant formations in vitro. Scintigraphic studies are described in nonfasting human volunteers either in upright or in supine posture, who concurrently were given one optimized F and one nonfloating (NF) hydrophilic matrix capsules of the same size, for three different sizes (small, medium, and large). In upright subjects, the F forms stayed continuously above the gastric contents irrespective of their size, whereas the NF ones sank rapidly after administration and never rose back to the surface thereafter. Consequently, the F forms show prolonged and more reproducible GRTs compared to the NF ones. The significance and extent of this prolongation are the most marked for the small size units (p < 0.001) but gradually lessen as the dosage form size increases (p < 0.05 for the medium size units), to become insignificant for the large size units (p > 0.05). Moreover, there is no significant difference between the mean GRTs of the small, medium, and large F units (p > 0.05). This indirectly confirms that the intragastric buoyancy of the F forms is the main process determining their prolonged GRT and protecting them from random gastric emptying related to antral peristaltism. Thus, their GRT depends mainly on the occurrence of the end point of digestion. To the contrary, the lasting retention of the NF forms in the stomach is only size dependent.(ABSTRACT TRUNCATED AT 400 WORDS)