The disproportionate difficulty in obtaining compelling experimental evidence from 4C-radiolabelling that the indole moiety of the otherwise isoprenoid penitrem A is biosynthesized by Penicillium crustosum directly from tryptophan has been explored. [benzene ring-' 4C]Tryptophan added to the broth beneath the mycelial mat of stationary liquid cultures labelled penitrem A with 1.4% incorporation, only threefold more than that determined for [methylene-14C]tryptophan or [U-14C]tyrosine, incorporation of which could only have been indirect. In contrast, the substituted tryptophan-histidine diketopiperazine roquefortine, biosynthesized concurrently with penitrems by this organism, was labelled with compelling efficiency (23.4% incorporation of [benzene ring-' 4C]tryptophan). In submerged culture, Claviceps paspali concurrently biosynthesized an analogous pair of metabolites, 3-hydroxy-3-methylbutenyl paspalinine and lysergic acid a-hydroxyethylamide. This feature enabled experimental demonstratation of [benzene ring-' 4C]tryptophan incorporation to an extent more consistent with direct contribution of the indole moiety of the indole-diterpenoid paspalinine derivative. The same precursor applied to the sporing surface of P. crustosum stationary cultures also provided stronger evidence for a direct biosynthetic role in the formation of penitrem A. In the absence of competition from any other indolic secondary metabolite, a submerged culture of Penicilliumpaxilfi incorporated 5 % of the [benzene ring-14C]tryptophan given during growth into the indole-diterpenoid paxilline. A double-labelling time-course experiment indicated temporal separation of steps in the biosynthesis of roquefortine. The inadequacy of classical precursor techniques for studying biosynthesis of indole-diterpenoids in P. crustosum is discussed. The more homogeneous submerged culture fermentation system is preferred for experimentation.