From the Introduction ... 'In those areas of the world subject to the twin conditions of severe dry seasons and the prevalence of so-called dry thunderstorms (which include large areas of the African continent), bush and grass fires caused by lightning are by no means uncommon. It may be assumed that such occurrences were as frequent in the past as they are now, and that the pattern and structure of modern plant and animal communities has to some extent been molded thereby. Conditions like these existed over large areas of Africa for many centureis, but it was not until the arrival of man that fire assumed an importance comparable to existing climatic factors in shaping the composition, structure, and distribution of plant communities and thereby, the animal populations with which they are inseparably linked....In the present paper I shall attempt to provide a general picture of the effects of fire on West African vegetation by drawing on my own experience in Ghana and on the work of other authors in various West African territories....'From the text (p.159) ... 'Observers are now generally agreed about the gross effects of fire on West African vegetation. It is accepted (a) that fire retards the natural development of vegetation towards a forest or woodland type climax and may hold it more or less permanently in a tree savanna stage; (b) that fierce hot fires occurring late in the dry season are particularly destructive of woody elements, especially seedlings and suckers; (c) that moderate cool fires occurring early in the dry season are less destructive and will allow slow development of woody vegetation towards a woodland type climax; (d) that fire is responsible not only for shaping the physiognomic structure of plant communities but exerts a powerful selective influence on floristic composition as well, fire-tolerant species being encouraged and fire-tender one eliminated; and (e) that fire, properly used, is a powerful tool for shaping vegetation towards chosen objectives. These observations are backed by a few widely scattered but reliable experiments.'