The variation in the mobile fraction seen in the H-1 n.m.r. free induction decay has been analysed using the Thomson-Gibbs equation to give the hard block lamellae thickness distribution in a poly(ester-urethane) elastomer. Attempts to use H-1 spin diffusion to fix the absolute distance scale were unsuccessful since the results indicated a heterogeneity with not all of the soft segments contributing to a simple two phase hard/soft domain morphology. Selective C-13 n.m.r. experiments confirm the partitioning of the H-1 n.m.r signal at room temperature into rigid and mobile components from the hard and soft segments respectively. Chain extender contributes to both the rigid and mobile H-1 n.m.r. signal.