This review sets out to define the role of surgical biopsy in the diagnosis and prognosis of diffuse parenchymal lung disease, the pros and cons of each procedure and the interpretation of the results in the context of clinical and other investigative data. Surgical lung biopsy remains the investigation with the greatest overall diagnostic sensitivity for diffuse parenchymal lung disease. Rates of diagnosis are approximately 90-95%, compared with approximately 70% for both high-resolution computed tomography (HRCT) and transbronchial biopsy. Although open lung biopsy, and more recently videoassisted thoracoscopic surgery, have good safety records, neither technique has been utilised as frequently as might be expected, with wide regional variation for various reasons. More recently, HRCT used in conjunction with clinical and other investigative modalities, has increased the accuracy of diagnosis for some diseases, including the majority of cases of cryptogenic fibrosing alveolitis (HRCT & bronchoalveolar lavage), Langerhans' cell histiocytosis (HRCT), lymphangioleiomyomatosis (HRCT), and silicosis (HRCT). At present, the vasculitides and the idiopathic interstitial pneumonias, other than cryptogenic fibrosing alveolitis, always require a surgical biopsy. As experience with high resolution computed tomography grows, it is possible that other diseases will be able to be diagnosed without surgical biopsy. These newer modalities of investigation do have appreciable limitations and where sufficient doubt exists about diagnosis, surgical lung biopsy must continue to be utilised.