The oligosaccharide structures of blood group antigens are not the primary gene products; they are constructed in a stepwise manner by adding particular sugar to precursor oligosaccharides via several glycosyltransferases coded for by different blood group genes (Watkins 1966, 1978, 1980). Consequently, final profiles of antigens expressed in each cell type are influenced by many different factors such as the intrinsic composition of glycosyltransferase species which are defined by the genotype of the individuals, relative activity or amount of these enzymes (repression, derepression or induction of the enzymes), competition between enzymes with overlapping substrate specificity, the organization of the enzymes in membranes, utilizability of precursors and specific substrate sugars, and the activity level of degradating enzymes. Changes in the antigen profiles during maturation, differentiation and malignant transformation are thought to be intimately related to the variability of these factors. Although great importance attaches to histo- and cytochemical information on the distribution and levels of glycosyltransferases and messenger RNA corresponding to the relevant enzyme, detailed and precise localization of the blood group antigens and their variants is the base line for analyzing these complex factors. On the basis of individual genotype and histochemical findings about the antigen distribution and the interrelationship between cells and cellular components producing different antigenic structures (cellular and subcellular mosaicism), we can deduce precursor oligosaccharide levels as well as the status of gene activation and its primary product, glycosyltransferases. Thus, these findings are a prerequisite for further analysis at the molecular genetic level. As emphasized in this article, lectin staining or immunostaining methods with MAbs combined with glycosidase digestion procedures are powerful tools for in situ analysis of carbohydrate structures in histochemical systems. Although in some cases valuable results have been obtained by applying the technique, our knowledge concerning the distribution of complex carbohydrate structures is still far from satisfactory. Along with well defined MAbs and lectins, the key to developing our methods further is successful introduction of glycosidases, in particular, endoglycosidases since these reagents are indispensable for analyzing the inner core structures and glycoconjugate species of the blood group antigens. Application of these techniques at the ultrastructural level is an alluring possibility, even though many difficulties must be overcome. Although their functional roles have not yet been determined, a diverse array of macromolecules is known to be decorated with blood group-related antigens.(ABSTRACT TRUNCATED AT 400 WORDS)