METEOROLOGY IO5 



three-hourly. Data from each of the principal stations is trans- 

 mitted twice daily to the central office at Dakar, where daily 

 weather maps are prepared and forecasts are broadcast. This 

 system of weather forecasting cannot attain its full significance, 

 however, until corresponding methods are adopted by the sur- 

 rounding territories. There are, in addition, fifty-eight first class 

 stations, where ground observations are made twice daily, and 141 

 second class stations for rainfall. Most of the rainfall records were 

 started in 1 900-1 905, but a few records on the Senegal coast go 

 back for fifty years. The service in Equatorial Africa, which is 

 organized on the same lines, was fully established only in 1935. 



The most important publication on this region is the large 

 volume by Hubert (1926) on West Africa, containing all data 

 available at that time, and profusely illustrated with maps and 

 graphs. Many of the results since 1931 are published in the Annales 

 de Physique du Globe de la France d' Outre Mer, edited by Hubert and 

 published bimensually. Welter (1930) has compiled a useful list 

 of eighty-five papers and books, of which thirty-five are by Hubert, 

 on the meteorology of French West Africa, while the most recent 

 publication on the subject is by Foissy and others (1937). 



Rousseau's paper of 193 1 on the rainfall of Senegal is particularly 

 interesting. The distribution of rainfall shows a marked latitudinal 

 or solar control, the average annual quantity ranging from sixty 

 inches in the south to twelve inches in the north. North of Gape 

 Verde cold water keeps the coast arid. The rainy season or 

 'hivernage' occurs between June and October. The rains follow 

 the two passages of the vertical sun in May and August but are 

 accentuated by a coastal monsoon accompanied by tornados which 

 make August the wettest month. The rainfall is much more 

 uncertain from year to year in the dry north than in the wet south. 

 Welter (1931) has analysed the rainfall of Dakar from records at 

 the railway station from 1887 and at the hospital from 1903. These 

 show a pronounced eleven-year periodicity over four cycles, coin- 

 ciding with the periodicity in sunspot numbers. In years of greatest 

 solar activity the annual rainfall amounted to 650-700 millimetres, 

 and in years of minimal solar activity it was reduced to 350-450 

 millimetres. Another long series of records at St. Louis has been 

 analysed by Constantin (1930). Apart from the eleven-year perio- 



