The amount of thermal motion occurring in individual polymers of a composite material like a plant cell wall can be determined using NMR relaxation methods. This methodology can be used to indicate of the rigidity of each polymer and its contribution to the rigidity of the cell wall as a whole. We have applied this approach to onion cell walls. The proton T-2 is used to identify individual polymer chains differing in rigidity whereas the proton T-1 rho discriminates between spatial locations within which the average amount of motion differs. Through a T-2-based spectral editing procedure we reconstructed sub-spectra corresponding to a mobile, largely pectic fraction of the cell wall and a rigid microfibrillar fraction which also included a pectic component. There was a third extremely mobile pectic component consisting of a of beta(1,4)-linked galactan and highly esterified galacturonan. These highly mobile, hydrated polymers are not represented in a CP-MAS C-13 spectrum obtained under normal conditions. We found, however, that by a combination of a long-contact experiment and a delayed-contact experiment we could reconstruct a C-13 spectrum of the cell-wall components that are normally too mobile to be visible. Through a T-1 rho-based spectral editing procedure we found some pectic material spatially located near cellulose. This included some 'eggbox' pectin. Pectic material was also located more than 2nm away from cellulose. These results show that, within a single cell wall, pectic polymers are very heterogeneous in rigidity as well as in composition, and are not distributed uniformly within the cell wall structure.