As I have already reported (Trans. IAU XVA, 9, 1973), the Working Group on Units and Time-scales was not able to reach any firm conclusions on whether the concepts of the astronomical unit and ephemeris time should be retained or replaced by the SI unit of length (metre) and atomic time. This disagreement was not unexpected as the membership was deliberately chosen to cover a wide spectrum of opinion. The following comments and suggestions represent a middle-of-the-road approach that is intended to meet the requirements of most astronomers and others who use the astronomical data. It has been suggested to me that this approach is not appropriate for use in those precise applications where relativistic concepts are necessary, and that therefore we should abandon completely the use of astronomical units of mass, length and time. There are, however, many applications where Newtonian concepts are perfectly adequate, and astronomical units are widely in use because of their general suitability and convenience. We should therefore continue to define a system of astronomical units, but should state quite clearly their relationships to SI units. An individual may use either system according to the circumstances, and standard conversion factors will be available for use by those who prefer the other system.