No abstract

Thus, these four cancers accounted for 75-1% of the total number of cases.Brief accounts of the spatial and temporal distribution of primary liver cancer and oesophageal cancer were given by Harington and McGlashan (1973a, b).In the present investigation, geographical definition of the patterns of distribution of the four major sites of cancer was sought in the belief that statistically significant gradients, together with changes in the cancer rates in the course of time, could provide a basis for future research programmes directed at aetological implications.METHOD OF SURVEY AND BASIC DATA (i) Areas of recruitment of the population at risk.-The health problems and environmental background of African workers in the South African gold mines were described by

A survey has been carried out covering the years [1964][1965][1966] to obtain the cancer incidence for Africans and Indians of the metropolitan area of Durban, Natal.Tables of comparison of the standardized cancer morbidity rates between Durban Africans, Durban Indians, Durban Hindus, Durban Moslems, Johannesburg Africans (1954) and English resident in England (four regions) are given.All rates used in this paper are standardized against the African Standard Population (Doll et al., 1966).The overall cancer incidence in Durban Africans of both sexes and Durban Indian females is as high as for most westernized countries. Indian males have a low overall cancer incidence which is not readily explained.The commonest male African cancers are at the sites of oesophagus, lung and liver; oesophageal and lung cancer rates show a tendency to rise.The commonest female cancer in both Africans and Indians is cancer of the cervix uteri. In addition, Indian females appear to be prone to cancers of the whole alimentary tract. This tendency is even more pronounced among Moslem women, who also are much more subject to breast cancers than Hindu women.

Summary.-A study of the smoking and drinking habits of 196 oesophageal cancer cases and 1064 control patients was made. All subjects were African males aged 35 years or more, drawn from a mainly urbanized population.It was found that tobacco smoking was prevalent and that pipe tobacco (used in pipes or in hand rolled cigarettes) was used more frequently than has been found in westernized countries. The drinking of alcohol was also a prevalent habit. Tribal affiliations were examined and all three of these factors showed differences between cases and controls. Further analysis of smoking and drinking together showed that only smoking had a positive association with oesophageal cancer, and this was also true after tribal adjustment had been made. A comparable analysis of data on Durban African males yielded similar findings. It was concluded that tobacco smoking was a powerful oesophageal insult but the authors were not able to show that alcohol was important in the development of oesophageal cancer in these people. Cigarette tobacco does not appear to be a significant oesophageal insult but pipe tobacco does, and the use of both these types of tobacco together may have a synergistic effect. Tribal affiliation has bearing on the smoking pattern.THE RISE in incidence of oesophageal cancer in southern Africa over the last 40 years has been well documented (Burrell, 1957(Burrell, , 1962Higginson and Oettle, 1960;Skinner, 1967;Oettle, 1967) and a current high incidence in African males has been found in Durban (Schonland and Bradshaw, 1968) and Johannesburg (Robertson, Harington and Bradshaw, 1971).Tobacco and alcohol consumption have been considered as possible aetiological agents by many workers (Clemmesen, 1965) and it is these two habits, taken separately and together, which have been investigated in this study.

SUMMARY.-The incidence of cancer among the African workers on the gold mines of South Africa has been studied for the period 1964-68. Considering the degree of selection to which they are subjected, the crude cancer rate was unexpectedly high.The most common cancers were those of the liver, the oesophagus, the respiratory system and the bladder. locally at the recruiting centre, the second on arrival at the transit station and the third at the mine to which the recruits have been assigned-these latter two both including X-ray of chest. In view of these repeated examinations the cancer incidence would be expected to belessthanin the normal populations from which they are drawn, as prospective miners with any obviously active disease are rejected at the first or second examinations.After a training of several months, the recruits contract to work for various terms of service which can be extended or shortened at the miners wish. They normally serve about 12-15 months, after which they return to their home clis-* Deceased, December 4, 1970. t Reprints from Dr. J. S. Harington.