The prevalence of neural tube malformations at birth is higher in the coalmining valleys of South Wales than anywhere else in England and Wales. In the eastern mining valleys of Glamorgan rates exceeded 1 % of all births in the period 1964-66 (Richards, Roberts, and Lloyd, 1972). The prevalence in the coastal regions of South Wales is only half that in the mining valleys and this is not due to area differences in social class, parity, maternal age, migration or Welshness (Richards, Roberts, and Lloyd, 1972). Recently Renwick (1972) has claimed that potato blight may be responsible for 95 % of all neural tube malformation prevalence, and that this hypothesis is consistent with the high prevalence in the coalmining areas of South Wales because many miners have allotments, and in the absence of cellars or other cool storage places, keep their harvested homegrown potatoes indoors in temperatures which are high throughout the year because of free or low-cost coal, conditions ideal for rapid deterioration.A recent study by Clarke, McKendrick, and Sheppard (1973), based on 83 Liverpool schoolchildren with spina bifida, found no significant differences in the amount of potatoes eaten by their mothers compared with a matched control group. We have available to us the names and addresses of each of the 40,032 infants born in Glamorgan between 1964 and 1966, of whom 26,114 were born into coalmining communities. Also available are the names and addresses of all such infants discovered by their second birthdays to have had neural tube defects. In this paper we have confined our attention to that group of potatoes considered by Renwick to provide the greatest risk to mothers in South Wales (viz., homegrown potatoes) but we have also collected additional information on length and site of storage and allotment status of parents, factors considered by Renwick to be critical in the explanation of the very high prevalence of neural tube defect in the South Wales mining valleys. MATERIAL AND METHODFor the three years beginning 1 January 1964 we had already collected information about all infants born to women resident in a defined area in South Wales and about all the congenital defects identified in that birth population. The scope of the investigation and the methods used are described elsewhere (Richards and Lowe, 1971). During the period of the study 743 singleton infants were ascertained as having neural tube malformation (out of a total birth population of 90,921 singletons), of whom 336 lived in Glamorgan. One control for each Glamoigan case was selected and matched for social class, parity, place of birth, and month of birth; maternal age had not been found to be related significantly to the prevalence of neural tube malformations (Lowe, 1972).A short questionnaire was prepared asking about the amount of homegrown potatoes eaten, length of time potatoes were stored before being eaten, and whether the child's father or grandfather had an allotment. The questions were asked in relation to the time of the infant's birth . Q...