This study evaluated the necessity and the contribution of solid media when used in conjunction with radiometric Middlebrook 7H12 (BACTEC 12B; Becton Dickinson, Towson, Md.) medium for recovery and complete identification of mycobacteria. Each of 1,184 digested, decontaminated respiratory specimens was inoculated into one BACTEC 12B vial, one 7H11 plate, and two Lowenstein-Jensen (LJ) slants. When the 12B vial was smear positive for acid-fast bacilli, the organisms were subcultured onto LJ slants and the BACTEC p-nitro-alpha-acetylamino-beta-hydroxypropiophenone (NAP) test system was inoculated with the organisms. Niacin tests were performed by using the growth from the original LJ slants and organisms from the LJ slants subcultured from 12B or 7H11 medium. The times to achieve definitive NAP and niacin test results were recorded. Recovery of all 143 isolates found in this study could not be achieved with a single medium. Among the three media, the highest percentage (92.8%) of Mycobacterium tuberculosis isolate recovered was with BACTEC 12B. The use of either 7H11 medium or LJ slants along with a 12B vial increased by 4 to 6% the total percentage of M. tuberculosis organisms that were isolated. Isolation of the M. tuberculosis complex and NAP differentiation in 12B medium were completed in an average of 17 days. On average, isolation and definitive niacin test results for M. tuberculosis cultures were obtained in 39.3 days by a conventional procedure and in 36.3 days when 12B subcultures were used. These results support the conclusion that LJ slants contribute 4 to 6% increased recovery of M. tuberculosis when used in conjunction with 12B medium. Additionally, a subculture onto LJ slants from 12B medium yielded sufficient growth for niacin testing earlier than an original LJ slant did.