From the text ... 'this general policy of minimal interference did not and could not extend to fires in Africa, as they are almost always due to man, and at their current frequency and extent must be considered to be 'unnatural' to the ecosystem. In these circumstances the control of fire is a necessary and very powerful tool for the control or alteration of both the plant and animal populations.... ...In brief summary, the essence of our policy is therefore as follows: 1) Our primary management aim is to maintain the present variety -- variety of habitat and variety of fauna. 2) Leaving the question of fire aside, management interference with these still largely natural and very complex ecosystems should probably be restricted as much as possible. 3) But fire, man-made fire, which is the main modifier of the ecosystem, must be controlled. 4) How best to control it requires research, including research into the most effective methods and means of control, bearing in mind the practical limitations of finance and staff.The history and prehistory of fire in the Serengeti ... is not known with any precision. But there are two safe assumptions. First, that fire has been a factor of primary ecological importance for millenia, that is as long as the Serengeti ecosystem has been substantially in the state it is today; for it is characterised essentially by fire modified vegetation. And, second, that the incidence of fire has increased during the last few decaded due to the rapidly increasing population round its borders. Obviously, therefore, the control of fire ... must lie broadly in the direction of reducing but not eliminating the impact of fire in the Park....any management policy aimed at the total exclusion of fire over most of the Serengeti, except in a few small areas of special value to visitors, is doomed to failure. Because of the inadequacy of the animal impact on the pasture in the woodlands and the consequent dangers of building up a supply of fuel, it is also probably unsound.Uncontrolled late hot burning is the commonest form of fire at present. It needs no encouragement from management.This leaves us only with the early burn as making any sense either from the practical or ecological points of view. ...it is patchy and this is its most valuable characteristic; it produces, and tends to maintain, a fire mosaic. Clearly, therefore, the main management goal in the Serengeti must be to increase the patchiness of fire, to encourage the formation of fire mosaics and to reverse the present trend towards blanket burns.The main immediate task of research must be to determine which areas most require to be given the chance of having a fire mosaic with the minimum demands on money and staff.... In the relatively small Arusha and Manyara Parks the picture is much less gloomy. Here, because of the considerable animal impact and the fact that we are dealing with 50 square miles instead of 5,000 the conversion of fire from a master to a servant, in Dr. John Phillips' words, appears to be entirely practicable. However, in the Tarangire (1,010 square miles) and Ruaha (5,000 square miles) the problems are on a scale similar to those of the Serengeti. In Mikumi, little ecological research has yet been done but, on an entirely superficial level, it would appear that the fire impact is not resulting in marked changes in the habitat. On Kilimanjaro ... there is going to be the very difficult long term problem of maintaining the forest girdle around the mountain at the lower altitudes; this will lie outside the boundaries of the Park.And, lastly, in the Gombe National Park our aim must be to exclude the fires which sweep down from outside over the crest of the escarpment, causing grasslands to encroach into the edge of the beautiful indigenous forest on which the chimpanzees depend....'