It is shown that the reaction of acrolein with bromomalonate in the presence of an alkali alkoxide, leading to a 2-formylcyclopropane-I,l-dicarboxylate, may be applied to the steroid series; thus, a 20-methylene 21-aldehyde gives in excellent yield the.corresponding cyclopropane derivative. The possibility of converting such systems to a-pyrones was investigated in the case of the acrolein adduct, diethyl 2-formylcyclopropane-1. I-dicarboxylate. While its cyclopropane ring could not be isomerized to a n olefin with aluminum chloride or by reactions catalyzed by transition metals and while photolysis led t o a dihydrofuran derivative, pyrolysis in high vacuum at 550-575" afforded directly a n a-pyrone, ethyl coumalate. Evidence is put forward for the intermediacy of 3-ethoxycarbonyl-a-pyrone in this reaction.La reaction de I'acroleine avec un bromomalonate en presence d'un alcoolate alcalin, qui conduit a un formyl-2 cyclopropanedicarboxylate-l,l, peut &tre appliqute a la serie steroi'de, un 21-aldehyde 20-mCthyl6nique donnant en excellent rendement le derive cyclopropanique correspondant. Nous avons CtudiC la possibilitk de convertir de tels systemes en a-pyrones dans le cas du formyl-2 cyclopropanedicarboxylate-l,l de ditthyle, derive de I'acrolCine. Tandis que son cycle a trois chainons ne put &tre isorntrise en olefine avec le chlorure d'aluminium ou par des reactions catalyskes par des metaux de transition et tandis que sa photolyse conduisit a un derive dihydrofurannique, sa pyrolyse, a 550-575" et sous haut vide, donna directement une a-pyrone, le c o~~m a l a t e d'ethyle. Nous suggkrons I'a-pyrone 3-tthoxycarbonylke comme intermediaire dans cette reaction.Can. J. Cheni.. 51. 3163 (19731 In 1969 our laboratory reported one of the way to bufadienolides (and generally to cifirst syntheses of bufadienolides, a class of pyrones) suffers only from the relatively low steroids comprising highly active cardiotonics, yield of the dehydrogenation step which, in our which are characterized by an a-pyrone ring experiments, did not exceed 35% of pure prodwhose position 5 is linked to the 17P-position of uct; Pettit et al. (3) report yields of 43-60% of the steroid nucleus. The salient feature of our crude product when the dehydrogenation is synthesis (I, cf. also 2) (compare Scheme 1, pathway a) consisted in a Michael addition of diethyl malonate to a 20-methylene 21-aldehyde, the hydrolysis product of the resulting adduct being subsequently subjected to enol-lactonization with concomitant decarboxylation, and the still required double bond, in conjugation with the carbonyl function of the lactone, being then introduced by a variety of methods. This pathcarried out with sulfur.The attempt to "prefigure" the desired double bond in the form of a suitable substituent of the malonate employed (cf. Scheme 1, pathway b) is obviously problematic. The olefin-forming elimination from the addition product, involving a leaving group (X in adduct iii of Scheme I) which is ci and not P to the ester groups, could hardly be ex~ec...