The combination of a divinylbenzene-based reversed-phase (RP) column and acetic acid gradients in water as mobile phase described in the accompanying paper was used for characterizing the extractable polypeptides from the normal and the diabetic human pancreas. The pancreas was lyophilized, minced and extracted three times in 3 M acetic acid. After mechanical clarification, the raw extracts were applied directly to the RP column. Alternatively, the extracts were lyophilized and subjected to size-exclusion chromatography on Sephadex G-50 in 3 M acetic acid. Two fractions with mol. wt. > 6000 dalton (Peak I) or with mol. wt. ⩽ 6000 dalton (Peak II) were obtained. The Sephadex G-50 size-exclusion chromatography and the RP-high-performance liquid chromatographic (HPLC) analyses of the crude extracts from a normal pancreas clearly demonstrated the weight distribution and differences between the exocrine pancreas (containing primarily the major digestive enzymes) and the endocrine pancreas (containing insulin, glucagon, etc.). RP-HPLC analyses of crude extracts from various normal pancreatic glands resulted in very similar UV profiles, whereas those from a number of individual diabetic glands differed. Chromatograms of acetic acid extracts from normal pancreata were similar when analysed before or after lyophilization, whereas lyophilization of acetic acid extracts of diabetic glands resulted in severely obscurred chromatograms. RP-HPLC analyses clearly demonstrated several differences between the diabetic and the normal pancreas. In the crude extracts, the extractable proteins from the diabetic pancreas were shifted towards lower molecular weight and/or hydrophobicity. Further, a peak co-eluting with authentic, human insulin could be demonstrated in the raw extract and in the peak II material from the normal pancreas, whereas virtually no mass signal was seen in the UV-profiles of similar materials from the diabetic gland. This finding was further verified by insulin radioimmunoassay (RIA) performed on the isolated fractions after RP-HPLC of a crude extract from a normal and a diabetic pancreas. The insulin content in the diabetic pancreas was found to be ca. 1% of that in the normal pancreas. When authentic glucagon was added to crude extracts from a diabetic pancreas, a single component was found after immediate analysis, but after several hours at room temperature the glucagon was found to be degraded. Added insulin was stable under these conditions. Similar RP analyses were performed on a silica C4 column eluted with an acetonitrile gradient in trifuoroacetic acid. Although the chromatograms were more complex than those obtained with acetic acid as mobile phase, they were comparable in outline, and the differences between the normal and the diabetic glands could again be demonstrated. One of the reasons for the differences between the normal and the diabetic pancreas may be that the normal glands were removed and frozen immediately after death, whereas the diabetic glands were obtained after an ischaemia time of 6 h or more, leaving considerable time for post mortem enzymatic degradation. The RP-HPLC system is very suitable for these analyses. The acetic acid extracts can be applied directly to the RP column and, after lyophilization, the isolated fractions may be re-analysed [RIA or enzyme-linked immunosorbent assay (ELISA)], or subjected to a second (alternative) RP-HPLC. After such double RP-HPLC purification, almost all major individual components in the peak I and II material from a normal pancreas were sufficiently pure for microsequencing, and identification (based on 30 steps in the gas-phase sequence and comparing the sequence information obtained with databases) again reflected the distribution of the principal components from the exocrine and the endocrine pancreas in the peak I and II materials.