An experimental study of sound propagation in the nocturnal boundary layer was undertaken near Wichita, Kansas in October 1999, in conjunction with CASES-99. The propagation experiment involved a series of five 6-m towers placed at distances between 300 and 1200 m from a loudspeaker. Each tower had microphones at heights of 0.5, 1, 2, and 3 m. Data runs consisted of broadcasting a 50-Hz square wave for a 2-hr block of time. The night of October 17 has been analyzed particularly closely. The evolution of the received acoustic signal energy was determined on this night by a complicated combination of gradual strengthening of the nocturnal inversion, two strong discrete events observed at the main tower (likely a density current and solitary wave), and signal variations attributable to changing interference patterns in the propagating acoustic modes. The very dramatic sensitivity of the sound waves to changes in nocturnal boundary layer structure suggests possibilities for remote sensing, but only if the important propagation mechanisms can be identified more clearly.