Mortality data from the vial bioassay procedure were used to establish the numbers of replications (based on 10 moths/treatment/rep) required to yield repeatable results at different precision levels. Precision levels for the coefficient of variation of 10, 30, and 50% were tested. For the 1/2 confidence interval, precision levels tested were 0.5, 1.0 and 1.5%. Bioassays, including an untreated check, were conducted on tobacco budworm, Heliothis virescen (F.) males, collected in traps at three locations: 1) the Lower Rio Grande Valley (LRGV), 2) the Brazos Valley (BV), 3) the Uvalde-La Pryor (U-LP) areas during 1988. Results were calculated from 876 bioasssays. At each location increased replications were required as treatment levels and mortality levels decreased in order to provide adequate precision based upon the statistical parameters tested. At 30% C.V., 100, 42 and 18 replicates of 10 moths/bioassay needed to be determined in the LRGV, BV and U-LP for the check. With 1% 1/2 C.I. calculation 35, 14 and 6 replicates at the same locations, respectively. At 1 µg/vial and 0.000625 KG(a.i.)/HA, the minimum dosages tested, only 20% to 50% of these totals needed to be sampled. Also, as mortalities increased (maximum observed 98%), the variance, as percentage of the mean, decreased to a low of 10%. Mortalities of check moths were always the lowest and thus required the greatest number of replication to provide statistical repeatability at the precision levels tested, regardless of the location. For the check from the Brazos Valley, 50 bioassays with a mean percentage mortality of 0.3 showed the same results as 57 and 114 bioassays.