560 SCIENCE IN AFRICA 



a general census was carried out in the Northern Provinces with 

 an intensive census in six districts of Katsina Emirate and a num- 

 ber of selected villages elsewhere. For the general census data 

 were given for age, sex, occupation, and religion, while the inten- 

 sive census included also birthplace, infirmities, fertility, numbers 

 of livestock, amount of cultivated or uncultivated land, and yield 

 of crops. In the Southern Provinces unsettled political conditions 

 made a complete census impossible and the figures were based 

 only on compiled estimates. The degree of accuracy of these 

 various censuses was estimated as follows: the probable error was 

 found to range from 2 per cent in the intensive census of the 

 Northern Provinces to 10 per cent in most of the Colony and 

 Southern Provinces, while in the provinces of Onitsha, Owerri, 

 and Calabar the error was up to 15 or 20 per cent. The Gold 

 Coast has had a decennial census since 1891 though the first two 

 were confined to the Colony. It was not till 1931 that educated 

 enumerators were used throughout. In the rural areas data 

 include tribe, colony of origin, infirmity, standard of education 

 (below' or above Standard IV), and age in three groups, under 

 fifteen years, fifteen to forty-five, and over forty- five. For the 

 large tow^ns occupation was added to these data and age was 

 further divided into less than one year, one to five, and five to 

 •fifteen. The results have been fully analysed by A. W. Cardinall 

 (1932). Sierra Leone has had a decennial census in the Colony 

 since 1881 and in the Protectorate since 1901. In 193 1 a complete 

 enumeration was made for non-natives, but Africans were still 

 being estimated by the administrative officers after detailed counts 

 in selected villages. 



In the French colonics a five -yearly census has been made 

 throughout the present century, in which French citizens repre- 

 senting about 2 per cent of the total population are enumerated. 

 For the rest estimates have been based on tax registers and 

 accuracy is not claimed (A.O.F. 1935). There are no data on 

 age-classes, etc. In 1933 a schemxC was put forward for French 

 West Africa by the Governor-General for an enlargement of staff 

 to obtain a proper census within five or six years, but no such 

 steps have been taken in French Equatorial Africa. French 

 Togoland and Cameroons have similar five-yearly censuses, but 



