Corialstonine is a novel quinoline alkaloid isolated from Alstonia coriacea. Its structure illustrates a possible new transformation of indoles into quinolines. The propensity of some indoles to transform themselves into quinolines is exemplified by the Cinchona or Cryptotheca alkaloids and to a lesser extent by lanceomigine from Alstonia and Huntaria species. We wish to describe a novel quinoline alkaloid, corialstonine 1, which also illustrates an alternative means of transposition of indoles into quinolines. Alstonia coriacea is a shrub from New Caledonia whose stem bark main alkaloid is norcquaternine. It is accompanied by corialstonine, an amorphous base which is slightly fluorescent on TLC ([α]D +102°, CHCl3, C=1). Corialstonine does not react with the Ce(IV) spray but stains orange after Dragendorff pulverization. Its UV spectrum does not resemble those of the usual indole alkaloids; two maxima appear at 317 and 330 nm, which are shifted to 354 nm upon acidification. The MS of 1 displays a molecular ion at m/z 410, analyzed for C23H26N2O5 (410.176; calc: 410.184); it is accompanied by an m/z 424 probably due to transmethylation. 1H NMR and 13C NMR spectra for 1 were obtained at 300 and 75 MHz respectively. They were uneventfully interpreted by means of homonuclear (1H-1H) and heteronuclear (1H-13C) correlated spectroscopy. It is thus found an isolated eight-carbon system corresponding to the core of all type-I indole alkaloids and a quinoline. Other systems are a methoxycarbonyl unit, a quaternary carbon and an isolated methylene with an unusually low geminal coupling constant (J=6.2 Hz); worthy of note is the absence of "tryptamine-like" CH2-CH2 protons. Structure 1 was finally deduced from biogenetic considerations and from observation of long-range 1H-1H and 1H-13C couplings. Of particular value are the correlations across quaternary carbons such as: C-2 with H-6 (3J); H-6 with H-3 (5J); C-4 with H-5'; C-14 with H-14. These experiments have also allowed complete assignment of both proton and carbon spectra including methoxyls and quaternary carbons. Configurations of C-15 and of the 19-20 double bond are assumed to be the "biogenetic" configurations; stereochemistry of the methoxycarbonyl group has not been determined. From a biosynthetic standpoint, 1 may arise from a precursor via keto-aldehyde 5 resulting from opening of the carbinolamine ether, ring closure of a six-membered ring and 1,2-carbon shift whose driving force would be aromatization of the quinoline system. Natural or artificial origin of C-5' is still questionable. Other pathways originating from an earlier biosynthetic intermediate such as geissoschirine may also be envisioned; in this latter case, it is possible to propose mechanisms in which C-5' would be C-17 of the precursors.