Brain weight is often said to be decreased in schizophrenia, but a reduction has only been found in a minority of studies. We have therefore carried out a meta-analysis to answer this basic neuropathological question. Data were identified from literature searches and from contacting researchers in the field who were invited to submit unpublished data. Inclusion criteria were: an operational diagnosis of schizophrenia, or comparison subjects with no neurological or psychiatric history, aged 18 or over, for whom brain weight, age and sex were known. Exclusion criteria were: a history of head injury, epilepsy, substance dependence or leucotomy; neuropathological evidence of neurodegenerative disorder or focal brain lesion. Results were analysed by multilevel modelling. Brain weight was, as expected, related to age and sex (both p<0.0001). After control for these factors, there was an effect of diagnosis, with brains from the 540 schizophrenia subjects being 2% lighter than from the 794 controls (weighted mean difference=24 g [95% confidence interval, 1-47 g]; p=0.04). The difference was similar in male and female patients. There was no correlation with duration of illness. In conclusion, brain weight is slightly but significantly reduced in schizophrenia, consistent in direction and magnitude with MRI volumetric findings. The result encourages a continuing search for the histological and molecular correlates of schizophrenia.