The behavior of aftershock sequences around the Nevada Test Site (NTS) in the southern Great Basin is characterized as a potential discriminant between explosions and earthquakes. The aftershock model designed by Reasenberg and Jones (1989, 1994) allows for a probabilistic statement of earthquakelike aftershock behavior at any time after the mainshock. We use this model to define two types of aftershock discriminants. The first defines MX, or the minimum magnitude of an aftershock expected within a given duration after the mainshock with probability X. Of the 67 earthquakes with M>4 in the study region, 63 of them produce an aftershock greater than M99 within the first 7 days after a mainshock. This is contrasted with only six of 93 explosions with M>4 that produce an aftershock greater than M99 for the same period. If the aftershock magnitude threshold is lowered and the M90 criteria is used, then no explosions produce an aftershock greater than M90 for durations that end more than 17 days after the mainshock. The other discriminant defines NX, or the minimum cumulative number of aftershocks expected for a given time after the mainshock with probability X. Similar to the aftershock magnitude discriminant, five earthquakes do not produce more aftershocks than N99 within 7 days after the mainshock. However, within the same period, all but one of the explosions produce fewer aftershocks than N99. One explosion is added if the duration is shortened to 2 days after the mainshock. The cumulative number aftershock discriminant is more reliable, especially at short durations, but requires a low magnitude of completeness for the given earthquake catalog. These results at NTS are quite promising and should be evaluated at other nuclear test sites to understand the effects of differences in the geologic setting and nuclear testing practices on its performance.