Tropospheric wave disturbances in the tropical western Pacific have been subjected to intensive analysis during three separate seasons. These all show evidence of spectral peaks in the 4‐ to 5‐day range and at periods longer than 10 days. In this review I attempt to synthesize the results for these three seasons, together with the analysis of a fourth season, into a tentative description of the various types of wave disturbances in this region. The conclusions are outlined below.It appears that there are two types of tropical disturbances that contribute to the 4‐ to 5‐day peaks: synoptic‐scale, westward‐propagating waves with phase speeds on the order of 6‐7 degrees per day, and planetary‐scale equatorial waves that extend into the lower stratosphere. There are also one or more wave types associated with the low‐frequency spectral peaks.All the tropical waves exhibit a similar vertical structure, with a phase reversal in the wind field between upper and lower troposphere. There are warm temperature anomalies at 300 mb, above the low‐level troughs, in qualitative agreement with the thermal‐wind equation. These warm anomalies coincide with the regions of maximum rising motion. As distinguished from tropical waves, the subtropical disturbances do not exhibit a phase reversal in the wind field between upper and lower tropospheric levels. The troughs of these waves appear to be colder than their surroundings at 300 mb.The available potential energy the tropical waves derive from condensation heating is immediately converted into kinetic energy by the synoptic‐scale divergent wind field. This appears to be the dominant energy source for these waves. The leakage of wave energy upward through the tropopause appears to be at least an order of magnitude smaller than the rate of energy generation.The spectral results indicate that most of the synoptic‐scale mass convergence into regions of disturbed weather takes place above the subcloud layer. Although the vertical distribution of latent heat release displays the same gross features in different regions, a simple linear proportionality between boundary‐layer convergence and condensation heating (as is usually assumed in the CISK hypothesi