The report¹ that Meerwein–Ponndorf reduction of gedunin (1a) gives 3α-hydroxy-3-deoxogedunin (2a) is wrong; the product is the 3β-epimer (2b). 3α-Acetoxy-7-deacetoxy-3-deoxo-7-oxogedunin (3c) was prepared by a stereospecific synthesis from cedrolide (1b), 7-deacetoxy-7-oxogedunin; osmium tetroxide oxidation then gave the 1,2α-glycol, isolated as the acetate (4a). Similarly, oxidation of the 3β-alcohol (3b) also gave the α-oriented glycol, as anticipated from the steric hindrance of the β-face of the molecule.

Oxidation of the allylic alcohols (2b) or (3b), or of the allylic acetate (2d) with perbenzoic acid also takes place from the α-face of the molecule, giving the corresponding α-oxides (5a–c). The oxidation is pH sensitive; with a benzoate buffer the 7-oxo group in alcohol (3b) undergoes Baeyer–Villiger oxidation, giving the ε-lactone isolated as the acetate (6a). Opening of the oxide ring in the 3β,7α-diacetoxy oxide (5c) is complex. Identification of the products shows that the reaction involves participation by both the acetate groups. In contrast the 3β-acetoxy-7-oxo compound (5d) and the ε-lactone (6a) give the products of ring opening with assistance from the neighbouring acetate group, while the 3β-hydroxy compound (5b) gives the simple bromohydrin.