There are two ways to find the cause of primary immunodeficiency diseases. One approach is to start with the immune defect and work backwards, using in vitro techniques to define where the primary abnormality lies in the immune response. Immunologists favour this approach for obvious reasons; and it is not without virtue since, for example, it has shown that the defect in most cases of primary hypogammaglobulinaemia is due to a failure of B lymphocyte maturation. The alternative approach is to screen empirically for defects in biochemical pathways in the hope of finding a clue which will eventually lead to the underlying disorder. This is a sensible approach in diseases which are clearly inherited (eg X-linked hypogammaglobulinaemia) but is less attractive in a disease such as late onset hypogammaglobulinaemia which is not obviously inherited. In practice, such procedures involve screening the urine for abnormalities in the quality or quantity of excreted compounds. Another way is to screen for abnormalities in organelle integrity by measuring the activity of various enzymes in subcellular fractions. In reality, the clue to the metabolic defect is usually discovered by accident, the prime example in our field being the discovery of adenosine deaminase (ADA) deficiency.