The sensitivity and practicability of in situ hybridization methods utilizing isotopic or non-radioactive labeling were compared. The aim of this study was to determine whether digoxigenin-labeled riboprobes are as sensitive as 35S-labeled probes to detect changes in type I and IV procollagen expression in an animal model of rat gastric ulcer. Both labeling and detection methods yielded similar results, with a superimposable signal distribution in the specimens. High levels of procollagen type I and IV transcripts were observed in spindle-shaped cells, presumably fibroblasts or myofibroblasts, localized in the ulcer base and rim. The increased expression of these collagen types suggests a remarkable upregulation of collagen expression during the healing of gastric ulcers. Liver tissue adhering to perforated ulcers displayed signals related to non-parenchymal cells, with hepatocytes demonstrating no detectable transcripts of type I or IV collagen genes. Due to the identical pattern of signal distribution by both hybridization techniques it is concluded that non-radioactive in situ hybridization is of value in monitoring highly expressed genes and yields results similar to those achieved with radioactive probes. In these cases, non-radioactive techniques are preferable because they are performed more rapidly and do not require handling of isotopes.