A novel alkaloid, (-)-corydecumhine (1) was isolated from the cultivated tuber of Corydalis decumbens (Thunh.) Pers. (Papaveraceae) and the structure was determined by the use of 2D NMR. Quirion et al. reported two alkaloids; (+)-decumbensine and (+)-epi-a-decumbensine from Corydalis decumbens' and their structures were proposed as 2 and 3, respectively. In 1989, Rozwadoska er al. synthesized a-hydroxyhenzyltetrahydroisoquinolines (2) and (3) but the spectral data of (+)-decumbensine and (+)-epi-a-decumbensine were found to be quite different than those of 2 and 3, so they pointed out that (+)-decumbensine and (+)-epi-a-decumbensine were not the a-hydroxybenzyltetrahydroisoquinolines (2) and (3). In addition, they also suggested that (+)-epi-decumbensine was probably the (+)-corytensine (4)' based on the 'H-NMR data, in spite of different molecular formula. Also in 1990, Gawley et al. synthesized (-) egenine (5) and (-)-co~ytensine while the (+)-egenine (6) was previously isolated from Fumaria vailantii ' and the spectral data of egenine and (+)decumbensine were found to be the same. Therefore it was concluded that (+)-decumbensine and (+)-egenine (6) were same compounds, so (+)-decumbensine was nor a new alkaloid. Latter, Quirion et al. also realized that their proposed structures for (+)-decumbensine and (+)-epi-adecumbensine were not correct. On this series of the alkaloids, Wu et al. also incorrectly suggested that (+) corytensine is the C.I' epimer of (+)-egenine (6). As we were comparing the alkaloidal constituents of wild and cultivated C. decumbens tubers in order to evaluate the quality of the crude drugs by hplc method, we came across one of the alkaloidal constituents showing a close similarity with compounds discussed above. We studied the structure of 1 by 2D nmr experiments such as 'H-~H COSY, 'H-¹³C COSY, 'H-¹³C long-range COSY in detail and other spectroscopic methods, which revealed that compound (1) is a new and the first type in this series of alkaloids due to presence of methylenedioxy group at 4' and 5' positions instead of common positions at 3' and 4'. We wish to report the structure elucidation of 1 in the present communication.