To effectively monitor a Comprehensive Test Ban Treaty, seismologists must be able to confidently detect and identify low-yield explosions. This requires the use of short-period regional phases, which can be extremely complicated. The Non-Proliferation Experiment (NPE) was a low-yield chemical explosion detonated at the Nevada Test Site (NTS) and recorded at more than 50 broadband seismic stations located throughout the western United States. These data were used to investigate the development of Lg and Pn, two seismic phases used in regional discriminants. The frequency-dependent attenuation for Lg recorded at 43 stations is described by the relation QLg (vertical) = 238 f1.28. The amplitude decay for Pn recorded at 38 stations is proportional to Δ−(1.29 + 0.05f) for the frequency window of 1 to 6 Hz. After removing the effects of distance and attenuation, we found the values of the spectral ratio Lg (4-6)/(2-4) to group according to the geologic terrain associated with the path traveled. Stations located within the Basin and Range had lower ratio values than stations located outside the Basin and Range. However, for the spectral ratio Pn (2-4)/(4-6), geologic terrain had an indirect effect. Pn is affected by Moho structure associated with the backazimuth, causing an azimuthal dependence. Of course, Moho structure can be a function of geologic terrain. Furthermore, the Pn arrival may be large or small, independent of azimuth. The values of the discriminant phase ratio Pn (1-2)/Lg (2-4) have an order of magnitude more scatter than Lg/Lg or Pn/Pn. Nonetheless, the values are both a function of geologic terrain (the Lg contribution) as well as Moho structure (Pn contribution).