Underground blasting in permanently frozen gravel was studied in a 7- by 12-foot drift by running a nonreplicated 2x2x2 factorial experiment on the effects of type of cut, delay, and explosive on the fragmentation subsystem. Comparisons were made between a v-cut round and a burn-cut round, millisecond delays and half-second delays, and 40 percent special gelatin and 60 percent high-density ammonia dynamite. The number of large blocks in the muck pile, loading time, shuttle car operating time lost because of cleanup needs, length of advance per round, and length of muck pile were recorded and statistically analyzed. None of the blast parameters had a significant effect on the length of advance or length of muck pile, although the half-second delays tended to give a shorter muck pile than millisecond delays. The burn cut required less secondary breakage, less loading and cleanup time, and gave better overall fragmentation than the v-cut. The ammonia dynamite shots required less secondary breakage and less loading time than those using special gelatin. The half-second delays gave better overall fragmentation and required less cleanup time, but the millisecond delays gave a better loading rate. None of the differences between delays and explosives were notably significant.