Summary. In order to make comparisons with the findings in a high frequency area, South Wales, with those in a low frequency area, south-east England, a birth frequency and family study was made of all births with neural tube malformations (spina bifida cystica, encephalocele, anencephaly, and iniencephaly) in 32 of the 33 London Boroughs over a 3-year period from 1 April 1965 to 31 March 1968. The births were ascertained through local authority registers, stillbirth and infant death certificates, and hospital records. The frequencies found were 1-54 for spina bifida (including encephalocele) and 1 41 for anencephaly (including iniencephaly). This was less than four tenths of the South Wales frequency. Evidence of an excess of winter births was found for both types of malformation, with a peak for conceptions in February, March, and April.The parents of 870 of the original 1209 index patients were traced and visited for the family survey. The usual social class effect was seen, a deficit of fathers in social class I and II. The birth order distribution of legitimately born patients standardized for maternal age showed only a small excess of firstborn and a deficit rather than an excess of lateborn. For maternal age, however, standardized for birth order, there was an excess of patients born to mothers under 20 and over 35 years of age. The whole family study sample showed a striking excess of patients born to parents from India and Pakistan compared to parents born in the West Indies in relation to households of immigrant parents in the 1966 sample Census. A small sample of 164 patients with matched controls had more parents born in Ireland and India and Pakistan and fewer born in south-east England and the West Indies than the controls.The proportions ofsibs affected with spina bifida and anencephaly were 3 42% for spina bifida index patients and 5 44% for anencephaly. For patients born after the index patient the proportions were 5-17 and 4-17%, respectively. The overall risk to sibs was lower than that shown in the South Wales survey, but substantially higher relative to the population birth frequency. The risk to sibs was not apparently influenced by father's social class but there was an indication of an effect of mother's father's social class, with a lower risk where mother had grown up in a class I, II, or IIIa home. There was no apparent influence of grandparental birth place. There was no apparent effect ofa relative affected other than a sib. Among cousins a significant increase over the population birth frequency was seen only in mother's sisters' children. The findings, like those of earlier surveys, suggest a multifactorial aetiology of the neural tube malformations, depending both on genetic predisposition and environmental triggers.