The pressure signatures of sonic booms measured in the field often show considerable variations from the idealNwaveshape. Peaked and rounded versions of theNare commonly observed. Pierce [J. Acoust. Soc. Amer.44, 1052–1061 (1968)] has proposed that peaking and rounding are the result of refraction and diffraction caused by atmospheric inhomogeneities. To test this explanation, we have carried out a model experiment in the laboratory.Nwaves about 1 cm long were produced by an electric spark, refracted and diffracted by a gas‐filled soap bubble, and received by a very wide‐range condenser microphone. The microphone output was displayed on an oscilloscope. The spark‐microphone distance was held fixed, and measurements were taken with the bubble at various points in between. The bubble acted as a converging acoustic lens when it was filled with argon and as a diverging lens when filled with helium. It was found that the converging lens caused peaking of theNwave, while the diverging lens caused rounding. These results qualitatively support Pierre's theoretical model. The data do not serve as a quantitative test of Pierre's specific numerical examples, however, because the particular conditions he assumed were not reproduced in our experiment. A discussion given of the effect of some of these conditions, such as the order and relative time of arrival of the refracted and diffracted waves, on the waveform. Another factor that had a rather strong influence on the waveform, particularly when a helium‐filled bubble was used, was the finite size of the microphone. The waveforms that would have been observed with a true point probe are deduced.