This thesis describes the cultivation and extraction of filamentous fungi isolated from extreme environments in the search for new antibiotic compounds. Filamentous fungi are a rich source of medicines including antibiotics, and it is believed that many currently unknown fungal species and bioactive fungal metabolites remain to be discovered. Aspergillus fumigatus and Penicillium nalgiovense strains were isolated from an antibiotic-contaminated riverbed near Hyderabad, India, and soil taken from a penguin’s nest on Paulete Island, Antarctica, respectively. It was anticipated that the extreme conditions within these environments would exert unusual selective pressures on their filamentous fungi, possibly causing the secretion of new bioactive compounds. The cultivation, extraction and analysis of metabolites from the A. fumigatus strain resulted in the isolation of the antimicrobial substance gliotoxin. Subsequent investigations revealed that this strain’s secretion of gliotoxin was increased by as much as 65 % when it was cultivated in the presence of pathogen-associated molecular patterns. These results indicate the existence of a fungal receptor/signaling system for detecting nearby bacteria. The scope for using gliotoxin and the related metabolite bis(methyl)gliotoxin as biomarker metabolites for diagnosing the lethal pulmonary condition invasive aspergillosis was also investigated. Bronchoalveolar lavage fluid from 42 patients with and without possible invasive aspergillosis was extracted and analyzed. The results obtained suggest that gliotoxin and bis(methyl)gliotoxin are not suitable markers for diagnosing invasive aspergillosis. Studies on the P. nalgiovense strain from Antarctica resulted in the isolation of the antifungal agent amphotericin B. The secretion of this compound increased when P. nalgiovense was cultured on a potato-dextrose agar enriched with coconut flakes rather than liquid RPMI 1640 medium. This was the first time amphotericin B was isolated from any organism other than the bacterium Streptomyces nodosus. The results presented in this thesis will be useful in the continuing search for novel bioactive compounds, the diagnosis of fungal infections, and as a source of insight into the interactions between microorganisms. Moreover, they show that even extensively studied fungal genera such as Aspergillus and Penicillium are not completely understood and may produce unexpected or previously unknown bioactive metabolites under appropriate conditions.