SINCE the discovery of superconductivity in the layered copper a oxide materials(1), a number of microscopic models have been proposed. To know which of these should be considered further, it is important to determine empirically the symmetry of the superconducting order parameter (the wavefunction of the superconducting condensate), For some time there has been conflicting experimental evidence as to whether the superconducting condensate has s- or d-wave symmetry(2); these terms, strictly correct only in tetragonal symmetry, are commonly used to denote whether the superconducting gap is finite in an directions at zero temperature or contains nodes. Tunnelling data along the c axis of these quasi-tetragonal copper oxides, perpendicular to the CuO2 planes, clearly show an s-wave character(3,4). Conversely, tunnelling along the a or b axis of YBa2Cu3O7-delta has indicated a d-wave character of the wavefunction(5-7), with one exception(8). Here I propose that these and other apparently conflicting results can be explained in a consistent way if there exist in the copper oxide superconductors two condensates, with different symmetry but the same transition temperature-in other words, if there are two kinds of superconducting gap.