Abstract In June, 1984, a large-scale outbreak of type A botulism occurred in 14 different prefectures in Japan, involving 36 cases and 11 deaths, due to consumption of vacuum-packed ‘karashi-renkon’ (deep-fried mustard-stuffed lotus root). The onset of symptoms ranged from June 9th through July 28th and many batches of ‘karashi-renkon’, having been manufactured by a factory during a period of three weeks from June 5th to June 25th, were found to be contaminated with type A spores. Type A spores (1.0 × 10 4 cfu/g) experimentally inoculated into the mustard-miso stuffing [water activity ( a w) 0.92, pH 5.0, NaCl 5.7%] did not multiply. By adjusting the stuffing to a w 0.98 (pH 5.4) growth of type A was still inhibited. When the stuffing with elevated a w and pH was heat-treated for 60 min at 80°C or autoclaved, it supported growth of Clostridium botulinum type A as indicated by the achievement of toxin levels higher than 10 5 LD 50/g. When mustard-miso was stuffed into the holes of lotus roots and refrigerated overnight or a longer period, both a w and pH of the stuffing increased to 0.98 and 5.3, respectively. The changes were apparently due to dialysis between the stuffing and the lotus root. This refrigeration process was actually performed at the plant, where the incriminated foodstuff was manufactured. The dialyzed mustard-miso stuffing, when heat-treated, supported the growth of type A organisms. Thus, we substantiated the possibility of primary multiplication of C. botulinum type A in mustard-miso stuffing subjected to refrigeration in lotus roots and then heat-treated.