METEOROLOGY 355 



Although this memoir, as has already been stated, deals 

 primarily with this one class of storm, an account is also 

 included of the storm-squalls and tornadoes of West Africa — ■ 

 a totally different class of phenomenon. 



The literature dealing with tropical revolving storms is 

 voluminous and scattered, and contains much speculation mixed 

 up with genuine observation in a highly confusing manner. It is 

 for these reasons very satisfactor}'' to have all available informa- 

 tion collected together into a single volume. The bulk of the 

 work consists, therefore, of a comparative study of the cyclones 

 of various parts of the world. This is preceded by an intro- 

 ductory note written by vSir Napier Shaw dealing with the 

 theoretical side of the subject, an account of which it is pro- 

 posed to postpone until after the facts to be explained have 

 been set out. The same applies to a short theoretical discus- 

 sion contributed by Dr. Harold Jeffreys. 



Tifncs and Places of Occurrence of Tropical Cyclones, — The 

 method of grouping adopted in the memoir under discussion 

 has practical merits to recommend it, but an alternative 

 system will be suggested later on. The original is adhered to 

 in the table on p. 356, which shows : 



(i) The month or months of maximum frequency of 

 occurrence (in italics). 



(2) Those months in which the frequency of occurrence 



is at least 50 per cent, of the frequency in the 

 maximum month. 



(3) The average annual frequency. 



The table is so arranged that the months noted down for the 

 Northern Hemisphere correspond as regards season with those 

 vertically beneath for places in the Southern Hemisphere. 



As has already been mentioned, an alternative system of 

 grouping might be adopted. In this alternative system the 

 cyclones which originate to the north of Australia and pass across 

 the western districts, first in a south-westerly direction, and then 

 south-eastwards, are taken to belong to the South Pacific area. 

 It then appears to be an accurate statement of the facts to say 

 that cyclones appear in the summer and autumn seasons near 

 the Equatorial margin of the trade winds on both sides of the 

 Equator wherever those winds extend eastwards across several 

 thousand miles of warm ocean. Three oceans provide the 

 required conditions, namely the Atlantic, Indian, and Pacific 

 Oceans. Of these, however, the Atlantic scarcely satisfies the 

 required conditions south of the Equator, for the Benguela 

 current chills the waters near the African coast, and on the 

 Brazilian coast the requisite stretch of warm ocean is therefore 

 not available. Thus of the six possible regions of cyclonic 

 formation, one is unsuitable for special reasons, the other five 



