Background/Aims: The accessory pancreatic duct (APD) exhibits several appearances on pancreatography, and sometimes is developmentally obliterated near the duodenum. We evaluated patency of the APD by dye-injection ERP to determine whether patency was related to duct course and shape, and size and location of the minor duodenal papilla (MIP). Methods: After routine ERP, we injected 2-3 ml of contrast material containing indigo carmine into the pancreatic duct via an endoscopic catheter and observed its secretion from the MIP endoscopically in 233 patients. Size of the MIP and distance from the orifice of the major duodenal papilla (MAP) to the apex of the MIP were measured endoscopically with measuring forceps. Maximum diameter of the APD and the length of the main pancreatic duct from its orifice to the junction with the APD were measured. Results: Patency of the APD was 48.5% (113/233). Size of the MIP varied considerably from 3 to 6 mm, but showed no correlation with patency. Six of 13 patients with chronic pancreatitis had the MIP larger than 6 mm, but only one was patent. The MIP was patent in 9 out of 10 patients (90%) when it was within 1.5 cm from the MAP. Frequency of a patent MIP was 17 out of 35 (48%) when it existed 1.5 to 2.0 cm from the MAP, and 33 out of 65 (51%) when the distance was more than 2.0 cm. The caliber of a patent APD was 1.6 +/- 0.6 mm, which was significantly larger than the caliber (1.1+/-0.4 mm) of a nonpatent APD (p<0.01). The length of the MPD from its orifice to the junction with a patent APD was 32.8+/-12.6 mm, which was significantly longer than the length to junction with nonpatent APD (22.6+/-8.2 mm)(p<0.01). The APD was classified according to duct course into long type which formed a straight line and joined the MPD at the neck portion of the pancreas, and short type which joined the MPD near the MAP and sometimes ran a descending course. Patency in the long type (75.5%) was significantly grater than in the short type (34.0%)(p<0.01). In the terminal shape of the APD, patency was most frequently observed in the spindle type (91.7%) and cudgel type (89.9%), and less frequent in the branch type (5.9%). In cases where the terminal APD had a cudgel configuration, the MIP was larger than in cases where the duct had a stick shape. Conclusions: Patency of the APD might be dependent on duct caliber, course, and terminal shape of the duct, and location of the MIP.