There is increasing evidence that inhomogeneous distributions of charge and spinÐso-called`striped phases'Ðplay an important role in determining the properties of the high-temperature superconductors. For example, recent neutron-scattering measurements on the YBa 2 Cu 3 O 7-x family of materials show both spin and charge¯uctuations that are consistent with the stripedphase picture. But the¯uctuations associated with a striped phase are expected to be one-dimensional, whereas the magnetic uctuations observed to date appear to display two-dimensional symmetry. We show here that this apparent two-dimensionality results from measurements on twinned crystals, and that similar measurements on substantially detwinned crystals of YBa 2 Cu 3 O 6.6 reveal the one-dimensional character of the magnetic¯uctuations, thus greatly strengthening the striped-phase interpretation. Moreover, our results also suggest that superconductivity originates in charge stripes that extend along the b crystal axis, where the super¯uid density is found to be substantially larger than for the a direction.Striped phases are inhomogeneous distributions of charge and spin that have been suggested to account for many of the unusual properties of the high-T c copper oxide superconductors 1±7 (T c is the superconducting transition temperature). In the simplest picture, the charge and spin can be thought of as being con®ned to separate linear regions in the crystal and thus resembling stripes. We expect static striped phases for the YBa 2 Cu 3 O 7-x materials not to coexist with superconductivity, except perhaps for materials with low oxygen contents and low values of T c . However, neutronscattering measurements have shown results for¯uctuations of both spin 8,9 and charge 10 that support the existence of a dynamic striped phase in YBa 2 Cu 3 O 7-x materials that have high T c values. Nevertheless, a dif®culty for the dynamic striped-phase picture is that, for both YBa 2 Cu 3 O 7-x (refs 8, 9) and La 2-x Sr x CuO 4 (refs 11, 12) copper oxide superconductors, the magnetic¯uctuations stemming from the spins have displayed a four-fold pattern at incommensurate points around the magnetic (1/2, 1/2) reciprocal lattice position, as shown in Fig. 1A. The interpretation of these measurements has been unclear, with both striped phases 1±7 and nested Fermi surfaces 13±15 being possible explanations. Because the striped phase is expected to propagate along a single direction, only a single set of satellites around the antiferromagnetic position should be observed. Tranquada 7,16 has suggested that the four-fold symmetry might arise from stripes that alternate in direction as the planes are stacked along the c axis.The problem in interpreting the four-fold pattern stems from the fact that the crystals used in the neutron experiments are twinned, so that along a given a or b direction, domains with lattice spacing a or b exist in equal proportion. This makes it impossible to distinguish whether satellites originate from the a* or b* direction in the reciprocal ...