National statistics for railway fatalities in England and Wales show a reduction of train crashes and a fluctuating level of deaths, of which an increasing proportion is from suicide. A closer examination of a two-year sample from the South of England revealed a large proportion of probable suicides and a small proportion of pure accidents. The remainder appeared to have medical, mainly psychiatric, contributions to their death, of which alcohol was an important factor in single young men. Rail suicides appear to be younger, the men less often married, the women more often married, and both sexes less often widowed than other suicides. They included more cases of major psychosis and neurosis, but fewer and less severe alcoholics. Characteristic patterns of this method of suicide are described with examples. Hypotheses to explain the choice of method suggest that it is not related to either volume of traffic, or residence in a rail-dense area or in an area with a high suicide rate, or the proximity of a psychiatric hospital.