HEALTH AND POPULATION 559 



data on literacy and infirmities, the latter being divided into 

 blindness, deaf-mutism, and total infirmity. In Tanganyika a 

 census was made by the German administration in 19 13 and by 

 the British in 1921, 1928, and 1931. The responsibility for these 

 lay mainly with the native administrations, but the figures were 

 checked by enumerations of selected villages carried out by the 

 district officers. A division was made into children and adults, 

 the line being drawn at puberty rather than at a definite age. 

 In the results it was assumed that a high proportion of children 

 is a sign of a growing population and suggests a correlation with 

 satisfactory health conditions, particularly diet, since the highest 

 percentage was found among tribes practising both agriculture and 

 animal husbandry. In Kenya no complete census has yet been 

 made of native races. Except for Europeans, the only areas where 

 all races were actually enumerated in 193 1 were Nairobi and Mom- 

 basa, the figures being supplemented by returns of natives resident 

 on European farms. The native population of nearly 3,000,000 is 

 estimated annually by the chief native commissioner on the basis 

 of counts made for hut and poll-tax: married men, single men, 

 women, and children are shown separately, but the counts of 

 men, old women, and children are said to be definitely inaccurate, 

 the last being sometimes estimated as a percentage, say 37 per 

 cent of the total population. The Kenya Land Commission 

 (Kenya 1934) concluded that the native population as a whole 

 is increasing rapidly and will do so during the next twenty years. 

 The censuses of Uganda were made in 191 1, 1921, and 1931. For 

 that of 1 93 1 the whole protectorate was divided into the smallest 

 recognized administrative units, and information for each member 

 was asked on tribe, sex, age, civil condition, occupation, and 

 infirmities. There were five age-classes recognized, under one year, 

 one to seven, seven to eighteen, poll-tax payers, and aged persons. 

 There is little doubt that the data are more correct than for most 

 other territories, but it seems that the demand on the native 

 enumerators was so great that many inaccuracies crept in. 



In Nigeria a decennial census has been made in Lagos since 

 1 87 1 and for the whole country in 191 1, consisting of a rough 

 enumeration, in 1921, when population figures were accom- 

 panied by data on native customs, and in 193 1. In that of 1931 



