Cellulosic microfibrils in plant cell walls are largely ensheathed and probably tethered by hydrogen-bonded hemicelluloses. Ensheathing may vary developmentally as hemicelluloses are peeled to enable cell expansion. We characterised a simple method to quantify ensheathed versus naked cellulosic surfaces based on the ability to adsorb a radiolabelled 'cellulose-complementary oligosaccharide', [3H]cellopentaitol. Filter-paper (cellulose) adsorbed 40% and >80% of aqueous 5nM [3H]cellopentaitol within ~1h and ~20h respectively. When [3H]cellopentaitol was rapidly dried onto filter-paper, ~50% of it was desorbable by water, whereas after ~1d annealing in aqueous medium the adsorption became too strong to be reversible in water. 'Strongly' adsorbed [3H]cellopentaitol was, however, ~98% desorbed by 6M NaOH, ~50% by 0.2M cellobiose, and ~30% by 8M urea, indicating a role for hydrogen-bonding reinforced by complementarity of shape. Gradual adsorption was promoted by kosmotropes (1.4M Na2SO4 or 30% methanol), and inhibited by chaotropes (8M urea), supporting a role for hydrogen-bonding. [3H]Cellopentaitol adsorption was strongly competed by non-radioactive cello-oligosaccharides (Cell2-6), the IC50 (half-inhibitory concentration) being highly size-dependent: Cell2, ~70 mM; Cell3, ~7 mM; and Cell4-6, ~0.05 mM. Malto-oligosaccharides (400mM) had no effect, confirming the role of complementarity. The quantity of adsorbed [3H]cellopentaitol was proportional to mass of cellulose. Of seven cottons tested, wild-type Gossypium arboreum fibres were least capable of adsorbing [3H]cellopentaitol, indicating ensheathment of their microfibrillar surfaces, confirmed by their resistance to cellulase digestion, and potentially attributable to a high glucuronoarabinoxylan content. In conclusion, [3H]cellopentaitol adsorption is a simple, sensitive and quantitative way of titrating 'naked' cellulose surfaces.