This study demonstrates that in mammalian and nonmammalian species it is possible to deplete selectively and reversibly retinal glycinergic neurons of their content of glycine by exposure to sarcosine, a competitive inhibitor of glycine transporter 1 (glyt-1). This observation was used as a tool to test the hypothesis that uptake of glycine rather than de novo synthesis is the main determinant of glycine content in retinal neurons. Isolated retinae were depleted of immunocytochemically detectable pools of glycine. Thereafter retinae were exposed either to physiological medium containing glycine or to medium lacking glycine but containing precursors for the synthesis of glycine. Retinae exposed to glycine-containing medium rapidly recovered their content of glycine, whereas retinae exposed to medium lacking glycine but containing serine, a substrate for synthesis of glycine, showed only a slow recovery of immunoreactivity for glycine in a few amacrine cells. These data indicate that uptake of glycine is the primary determinant of glycine content in most retinal glycinergic neurons. The origins of the extracellular pools of glycine remain to be identified; however, it is suggested that such glycine may be derived from the vitreous humor and that in turn this glycine may be derived from the peripheral circulation.