By 1970-1975, phosphonoyl substituents had been introduced at almost every position of aldoses, with the exception of the anomeric position of pyranoses and furanoses [ 1,2]. There was no method for the synthesis of glycosyl phosphonates of the type illustrated by glucose-l-phosphonate l and ribose-1 -phosphonate 2, which were of particular interest as (presumably) isopolar, nonisosteric analogues of the biologically important glycosyl-1 -phosphates and as enantiomerically pure a-heteroatom substituted phosphonates [3] of potential use for the "Umpolung" of the reactivity at the anomeric center [4,5]."To whom correspondence should be addressed Paulsen's group had thoroughly investigated the Michaelis-Becker and the Arbuzov reaction of variously protected glycosyl halides and had shown that mainly elimination products are formed, presumably according to an E2C mechanism. Dialkoxyphosphonates of type 3 were obtained when silver dialkyl phosphites and AgC104 or the corresponding halomercury compounds were used. The desired C(1), P bond was formed by treating allylic glycosyl halides with trialkyl phosphites under acidic conditions, to give allylic phosphonates of type 4 (Figure 1). These compounds were hydrogenated and used for the determination of the anomeric effect of the -P03Me2 group (0.56 kcal/Mol) (see [2] and earlier papers of the series). Unfortunately, the olefinic double bond could not be epoxidized or dihydroxylated. The desired glycosyl phosphonates remained inaccessible.Our interest in the chemistry of phosphonic acids [6] and in new carbohydrate derivatives involving the anomeric center led to the questions formulated in Scheme 1 . We expected that the reaction of dialkyl phosphite anions with halonitroalkanes and 1 -nitrosulfones reported by Russell and Hershberger