The number of genes known to be involved in mitochondrial energy production and the elucidation of the function of their individual transcripts is still increasing. Although at this stage it is impossible to predict the number of human genes necessary for mitochondrial biogenesis and maintenance, the total number in humans will most probably exceed the number of mitochondrial genes found in, for example, the budding yeast, which is about 800. Without doubt we have only seen the tip of the iceberg of the clinical spectrum of mitochondrial disorders. Recent findings such as mutations in structural complex II genes in certain tumours emphasize the need to think outside the classical clinical presentation. We propose the consideration of a mitochondrial disorder in every chronic, intermittent or progressive disorder with single system or multisystem involvement, even if lactic acid is normal, and discuss such dilemmas as whether we should 'scrape the barrel' in every patient that are raised by this statement. The characterization of mitochondrial and nuclear DNA mutations in patients with enzymatically established mitochondrial defects has taught us that several of the current clinical and diagnostic assumptions have to be altered or even eliminated. The most challenging future task will be the development of new diagnostic criteria covering the expanding clinical spectrum of mitochondrial disorders.