The original 1969 definition of SIDS as "unexpected by history" and "unexplained after thorough postmortem examination" is under review in the light of two decades of experience. Suggested modifications include restricting the age to less than 1 year, stipulating that the necropsy includes appropriate histology and laboratory tests, and requiring a review of the clinical history and examination of the death scene. The use of a protocol is recommended both by professional and parent groups. Although the diagnosis of SIDS is to some extent one of exclusion, there are several typical findings which are of value in diagnosis and suggest new avenues for research. External examination is important to exclude trauma and signs of suffocation. A recent study has confirmed that petechiae on the face are rare in SIDS and if found raise the question of deliberate or accidental suffocation. Frothy fluid escaping from the nose and mouth is seen in about half of infants who die from SIDS. Postmortem hypostatic staining as an indicator of position has assumed increased importance since prone sleeping has been shown to be a major risk factor for SIDS. Evidence of sweat in clothing suggests overwrapping. Internal examination shows subserosal petechial haemorrhages in the thymus in most cases. These may be related to age and are commoner in babies dying of SIDS than in controls. Relative sparing of the cervical extension of the thymus is strong evidence for negative intrathoracic pressure, perhaps due to upper or lower airway obstruction. Other typical findings are liquid heart blood, prominent lymph nodes, and an empty bladder (which frustrates some biochemical tests in about half of cases). The lungs are usually well inflated, arguing against surfactant deficiency as a significant cause of SIDS. Microscopic evidence of pulmonary oedema and congestion is found in infant deaths for many reasons and is not discriminatory for SIDS. Minor inflammation and infection of the respiratory tract is common in SIDS and may be important by contributing to overheating, apnoea, or sensitisation to bacterial toxins. Mild fatty change in the liver is very common in infant deaths. Panlobular microvesicular fatty change is rare and may require special stains for its recognition. It indicates the necessity of searching for inherited biochemical disorders. Although these are rare in true SIDS, they are an important cause of unexpected death in infancy. Of Naeye's "tissue markers of hypoxia'', extramedullary haemopoiesis in the liver and brainstem gliosis have been confirmed. Persistence of fetal haemoglobin and raised hypoxanthine values in vitreous humour are further pointers to periods of premortem hypoxia. Painstaking neuropathology has shown delayed myelination and maturation of dendritic spines. Changes in the brain may explain the link between antenatal factors such as smoking and SIDS. A second cot death in a family requires expert examination. Minor injuries or unexplained apnoeic spells may be important retrospective clues to non-accidental injury. Investigations mus exclude inherited disorders before the death is ascribed to SIDS. Parents demand that the pathologist takes care of their baby before, during, and after the necropsy, carries out the procedure to a high standard, checks reconstruction of the body, facilitates access, and is responsible for communicating the results of the examination. The "SIDS postmortem" presents both a practical and an intellectual challenge.