We show that under certain conditions a simple relationship exists between the elastic scattering of a composite halo nucleus and of its core from a stable target. The coupling of the elastic and projectile excitation channels is crucial to the analysis, which is particularly useful when the ratio of the halo to the core mass is small. In the case of 11 Be elastic scattering the cross section relationship is quite well satisfied. For both 11 Be and 19 C our analysis reveals a significant sensitivity of elastic scattering data to the halo size and structure. [S0031-9007(97)03995-1] PACS numbers: 25.60. Bx, 25.70.Bc The existence of a class of light nuclei with a localized central core surrounded by a dilute "halo" of neutron matter is now well established. Evidence for these novel structures has been gained mainly from measurements of total neutron removal [1] and breakup reaction cross sections [2][3][4], particularly at high energy. We investigate to what extent complementary information can be gained from high quality elastic scattering measurements at lower energies.In this Letter we show that, in certain circumstances, the elastic scattering of a halo nucleus from a stable target can give simple direct evidence for the structure of the halo. The theory makes explicit use of the characteristics of halo nuclei, namely their very small neutron separation energy and the large spatial extension of the halo, which in turn result in strong coupling between the halo ground state and low energy excitations. This coupling of the elastic and projectile excitation channels plays a crucial role in the analysis, the results of which cannot be readily understood in terms of optical or folding model approaches. In elastic scattering, the analysis is expected to be particularly useful in systems where the ratio of the halo particle mass to the core mass is small.We consider the scattering of an assumed two-body projectile nucleus, composed of a core of mass m c and a valence particle of mass m y , by a third (target) nucleus of mass m T . It is assumed that these three bodies are spinless and structureless, although these are not essential assumptions. Two key conditions must be met, however, for the subsequent development to be a useful one:(i) The interaction between the core and target (V cT ) should be effectively much stronger than that between the valence and target particles (V yT ).(ii) The relative motion of the core and valence particles is slow compared to the relative motion of the center of mass of the projectile and target, and so can be treated adiabatically, in the spirit to the work of Johnson and Soper [5].Requirement (i) places limitations on the likely regions of applicability of our model. In strong interactiondominated processes, it is likely to be most valid when the number of nucleons in the core far exceeds that of the valence body. For 11 Be elastic scattering, where this ratio is 10:1, the assumption will be seen to be a good starting point for small scattering angles. Requirement (i) could also...