A 30-year record of subjectively-analyzed daily temperature and geopotential height data from the lower and middle stratosphere (50, 30, and 10 hPa) is used to derive a climatology of planetary waves, with zonal wavenumbers s=1, 2, and 3, of temperature T and geopotential height Z. This should be useful as (1) a reference data set for intercomparison with shorter records and validation of general circulation models and (2) for fundamental understanding of the nature of stratospheric wave disturbances. First, seasonal (December, January and February) statistics are presented. The steady, standing waves are strongest at 60¿--70 ¿N and generally lead to northward eddy heat and momentum fluxes consistent with vertically propagating, dissipated disturbances, in broad agreement with previous studies based on shorter data sets. At 30 hPa the stationary eddies provide about 50--60% of the net northward heat flux, but the transients contribute a similar proportion to the net northward momentum flux. Statistical analysis of the transient wave amplitudes reveals unimodal distributions, which are skew for higher wavenumbers. Transient variations in the eddy fluxes are shown to be mainly due to amplitude variations rather than changes in the barotropic or baroclinic structure of the eddies. It is shown that changes in the geographical location of the waves contribute a considerable amount to the transience, so that a composite mean based on the wave amplitude maximum contributes a much larger fraction of the net eddy fluxes than the stationary waves. The importance of the long data record for these statistical distributions is emphasized by examining a 10-year subsample. Finally, the occurrence of ''preferred flow regimes'' through the 30-year data record is examined; neither trends nor definite links with possible external forcing mechanisms are found, despite considerable low-frequency (biennial to decadal) variability. ¿ American Geophysical Union 1996