The reaction product of boric acid and the polysaccharide guaran (the major component of guar gum) has been investigated by 11B NMR spectroscopy. By comparison with the 11B NMR of boric acid and phenylboronic acid complexes of 1,2-diols (HOCMe2CMe2OH, cis-C6H10(OH)2, trans-C6H10(OH)2, o-C6H4(OH)2), 1,3-diols (neol-H2), monosaccharides (L-fucose, mannose and galactose) and disaccharides (cellobiose and sucrose) it is found that the guaran polymer is cross-linked via a borate complex of two 1,2-diols both forming chelate 5-membered ring cycles ([B5(2)]), this contrasts with previous proposals. Based upon steric constraints we propose that preferential cross-linking the guaran polymer occurs via the 3,4-diols of the galactose side chain. The DeltaH and DeltaS for complexation of boric acid to cis- and trans-1,2-cyclohexanediol have been determined, from the temperature dependence of the appropriate equilibrium constants, and used in conjunction with ab initio calculations on model compounds, to understand prior conflicting proposals for guaran-boric acid interactions. 11B NMR derived pH dependent equilibrium constants and ab initio calculations have been used to understand the reasons for the inefficiency of boric acid to cross-link guaran (almost 2 borate ions per 3 monosaccharide repeat units are required for a viscous gel suitable as a fracturing fluid): the most reactive sites on the component saccharides (mannose and galactose) are precluded from reaction by the nature of the guar structure; the comparable acidity (pKa) of the remaining guaran alcohol substituents and the water solvent, results in a competition between cross-linking and borate formation; a significant fraction of the boric acid is ineffective in cross-linking guar due to the modest equilibrium (Keq). In contrast to prior work, we present evidence for the reaction of alcohols with boric acid, rather than the borate anion. Based upon the results obtained for phenylboronic acid, alternative cross-linking agents are proposed.