Vinylation of 2,3 :4,5-di-0-isopropylidene-aldel~ydo-D-arabnoe gave a 1.3:1 mixture of 1,2:3,4-di-0-isopropylidene-6-heptene-L-gulo (and D-manno)-1,2,3,4,5-pentol (la and 20). Successive ozonolysis and hydrolysis of the mixture of l a and 2a gave D-glucose and D-mannose in a ratio of 1.4:1. Vinylation of 2,4-0-ethylidene-aldel~ydo-D-erythrose gave a 3.9:1 mixture of 1,3-0-ethylidene-5-hexene-L-ribo (and ~-lyxo)-1,2,3,4-tetrol (3a and 4a), which was degraded to form D-ribose and barabinose in a ratio of 4:l. The configuration of each of the C-vinyl compounds la, 2a, 3a, and 4a was assigned by degradation to the corresponding aldose.Canadian Journal of Chemistry. 46, 3679 (1968) A report from this laboratory described a new method of lengthening aldose carbon chains, which involved vinylation of an aldehydo sugar, ozonolysis of the product, and removal of hydroxyl protecting groups (I). Treatment of 2,3 :4,5-di-0-isopropylidene-aldeIzydo-D-arabinose with vinylmagnesium chloride gave a product which was resolved by gas-liquid partition chromatography (g.1.p.c.) into two components, A and B, in the ratio of 1.3 :l. The data obtained from elemental analysis, quantitative hydrogenation, and infrared spectroscopy were consistent with a di-O-isopropylidene-6-heptene-1,2,3,4,5-pent01 structure. The formation of monoacetates with acetic anhydride indicated the presence of a single hydroxyl group. Ozonolysis and deketalization of pure component A gave D-glucose' with a small proportion of an arabinose. Component A (fast moving on g.1.p.c.) was therefore 1,2:3,4-di-0-isopropylidene-6-heptene-L-gulo-1,2,3,4,5-pent01 (la). Similar degradation of component B (slow moving on g.1.p.c.) gave D-mannose,' with some arabinose. Hence component B was the D-manno epimer (2a).Similar methods were used to verify the structures of the 3-acetates (lb and 2b), which 'Degradations of mixed epimers on a preparative scale gave only D-sugars. Aldoses formed for configurational assignment were therefore assumed to be D-enantiomers, although they were identified only by chromatography.were resolved by gas-liquid partition chromatography (g.1.p.c.) into a fast-flowing component, C, and a slow-flowing component D. Since components C and D were degraded to D-glucose' and D-mannose,' respectively, C was the L-gulo epimer (lb) and D was the D-manno epimer (26).Preliminary experiments in which l a and 2a were ozonized in ethyl acetate, and then hydrogenated and deketalized, resulted in the formation of appreciable quantities of arabinose. Efforts were therefore directed to reducing the formation of this by-product. Since conditions which might lead to overozonation were avoided, arabinose formation appeared to be an example of "abnormal ozonolysis" ofan allyliccompound, involving the rearrangement of an alkene-ozone reaction intermediate (2). In this connection Young et al. (3) observed that allylic acetates were less prone to "abnormal ozonolysis" than their parent alcohols, while Milas and Nolan (4) claimed that some "abnormal ozonolyses" could ...