The structure of a new alkaloid, dehydro-deoxynupharidine (I), isolated recently from the roots of Nuphar japonicum DC. was reported1∗ briefly and its detail will be described in this report. The base which was distilled between 110~120°/3 mm. Hg from the methanol extract of the roots of Nuphar japonicum DC. is composed mainly of deoxynupharidine.2~5) The distillate is much unstable than (')~deoxynupharidine (II), b.p3 112~115°, [α]D -112.5°, which was purified through the HCl salt, and it changed its color into brown when kept standing overnight. This fact suggested that the crude base might be contaminated with easily oxidizable bases and, therefore, an investigation of new bases was carried out. The crude deoxynupharidine was once converted into its HCl salt, and the HCl salt of II was attempted to be separated as much as possible by recrystallization. The free base from the mothor liquor was afforded as a crystalline perchlorate, C15H21ON.HClO4, m.p. 159-161°, [α]29/D+130.1°(CHCl3) and hydrobromide, C16H21ON.HBr, m.p. 199.5-201°. The infrared spectrum of the perchlorate in chloroform showed at 877, 1510, and 3128 cm-1 which may be assigned to furan, and also at 1640 cm-1 which may be considered the iminium conjugated with furan ring. Similarly, thé nuclear magnetic resonance spectrum indicated 3 signals (8 8.09, 7.56, and 6.82) which corresponded to one proton each. These data assigned clearly to the furan ring. The free base (I) derived from the salt, was so unstalbe colorless liquid, b.p3 125° (bath temperature) that it changed its color into brown during the distillation. The new base, C15H21ON (I) which gave perchlorate of m.p. 159~161°, had been named dehydro-deoxynupharidine. The perchlorate of I was reduced by sodium borohydride to a colorless oily base, C15H23ON (III), b.p3 125° (bath temperature), [α]D '114.1° (CHCl3), IR cm-1: 2764, 2792 (trans-quinolizidine), 874, 1032, 1505, 3140 (furan), which gave picrate, m.p. 154.5~155.5° and perchlorate, m.p. 203~204.5°. No difference of the infrared spectra and no depression of melting point by admixture indicated that III was in accordance with II. Therefore, I is considered to be composed of having similar C-N-O skeleton to deoxynupharidine (II), though the position of one double bond being remained undecided. The facts that the perchlorate of I prepared the dihydro compound (II) by sodium borohydride reduction, and that it indicated a negative reaction by the active hydrogen measurement of Zerewitinoff, suggested the enamine type formation of I, and also the iminium-type formation of the salt of I. On the other hand, the enamine had been already reduced also to dihydro derivative5) by the reagent. The reduction of I by sodium borohydride was attempted and the production of the expected II was identified. © 1965, The Pharmaceutical Society of Japan. All rights reserved.