Chromatographic analysis on silicic acid of the so-called "tri-0-tosyls~~crose" which is formed in %yo yield on the reaction of 3 moles of p-toluenes~~lphonyl chloride with 1 mole of sucrose in pyridine a t 0' has shown the substance to contain penta-, tetra-, tri-, and di-0-tosylsucroses in the molar ratios 0.05:0.33:1:1, respectively. Any mono-0-tosylsucrose present in the reaction mixture may have been lost in the isolatio~l of the p r o d~~c t .'l'he cornposition of the product was substantially the same when the reaction was performed a t -18". The chromatogram separated the tritosylates into two subfractions. The major subfraction represented 29% by weight of the original "tri-0-tosyls~~crose" and c o~~l d be converted to lr,2:3,G:3',6'-tria~~hydrosucrose in 77.4% yield. Therefore, the so-called "tri-0-tosylsucrose" can contain only 25-29yo of 11,rJ,6'-tri-0-tosylsucrose. G,G1-Di-0-tosylsucrose crystallized from the di-0-tosylsucrose fraction in 10% over-all yield.Cornpton (1) obtained evidence that prinlary hydroxyl groups can undergo tosylation substantially more rapidly than do the secondary positions in a carbohydrate structure. He obtained yields of 41y0 and 36% of the 6-0-tosyl derivatives on ~nonotosylations of the methyl P-and a-D-glucopyranosides, respectively. These results indicate a difference in reactivity between the primary position and the average secondary position in these compounds in the order of 7. On the other hand, Hockett and Downing (2) found that the hydroxyl groups in 1,2:3,6-di-0-isopropylidene-a-D-glucose and 1,2:3,4-di-0-isopropylidene-a-D-galactose differed in reactivity by a factor of 70. Little is Itnown about the factors responsible for such differences in reactivity. Undoubtedly, non-bonded interactions in the transition state are important and probably ~nainly responsible for the generally greater reactivity of primary positions (3, -1). I t should be noted, for example, t h a t the large difference in reactivity found for the di-0-isopropylidene derivatives of glucose and galactose may be mainly clue to the fact that the 3-hydroxyl group of the glucose derivative is eclipsed with the 4-position. Lemieux and I/IcInnes (3) have recently shown that intramolecular hydrogen bonding also can strongly influence the rate of reaction. Furthermore, little information is available on the relative rates for the tosylatio~l of the starting material and the first product of tosylation. Experience in related fields has shown (6, 7, 8) that this can be a dominant feature of the esterification of polyhydroxy conlpounds with hydrophobic reagents. Thus, when it is considered t h a t the reactivities of both the primary and the secondary hydroxyl groups must vary considerably amongst then~selves, it seems clear that attempts a t preferential tosylations of the primary positions of polyhydroxy conlpounds can be expected to proceed in widely varying yields.Hocltett and Zief reported the preparation (0) of a substance termed "tritosylsucrose" by treatment of 1 ~nole of sucrose with 3 ...