•68 PRESIDENTIAL ADDRESS SECTION C. 



consists in interpreting Nature's own experiments. It has certain 

 very definite advantages. While quadrat experiments carried out 

 in one locality may not apply to another locality, observation by 

 Nature's experiments can be made over very wide areas. 



We observe, for instance, that certain grasses colonise bare 

 area*- such as old roads (species of Aristida, Eragrostis, etc.). 

 Tufts of these pioneers are found elsewhere being smothered and 

 killed by the denser growth of species (e.g., Themeda) which do 

 not as a rule themselves act as pioneers. The latter are in turn 

 themselves killed out by still taller species (Andropor/on). Three 

 stages of the succession are at once displayed, and it is unnecessary 

 to await the course of events in certain definitely mapped quadrats, 

 which, to give the same information with equal certainty, would 

 require to be very numerous and to cover a very wide area. 



A walk along the margin of a forest gives just as definite 

 information regarding its history. If the marginal trees are all 

 branched from the base outwards towards the light these trees 

 cannot have grown up in the centre of dense shady forest. The 

 forest in this case is not of the so-called retrogressive type. Its 

 margins are not being eaten into by grass-fires. Further we often 

 find that species characteristic of the scrub zone, which usually 

 surrounds our forest-species — which are light demanders and quite 

 intolerant of shade — may be found just inside the forest margin 

 in a dead or dying condition. They have been overtaken by the 

 forest trees. It is quite obvious that they could not have grown 

 up in the shade from the beginning. Two stages in the succession 

 are here clearly demonstrated, the scrub stage and the forest stage. 

 But these stages may be connected with the grassland stages already 

 analysed for the. tall (Andropor/on) species of grasses are in the 

 same way seen to be killed out by scrub. We thus find displayed 

 five stages of a complete unit succession or "sere." 



By observing the natural regeneration of the tree species on 

 the floor of a forest, by determining the relative ages to the com- 

 moner species, by searching for the causes of the death of certain 

 species, further successional tendencies are brought to light. Not 

 only can the past history of the forest be read, but its future 

 development can be foreseen. 



Cases of plants suppressing other plants are always to be looked 

 for. A dead plant or a dying plant is a flag signal to the ecologist. 

 Often the cause of death may not be suppression by other plants 

 but an attack by insects or fungi. Insect attacks on the dominant 

 species may change the whole course of the plant succession, and a 

 stable type sometimes is changed without any change in climate, 

 as in the case of the falcate yellowwood forests of the Drakensberg. 

 There are many other interesting points in regard to the biotic 

 factors and their influence on the vegetation of Natal. Plant suc- 

 cession is influenced greatly by seed dispersal and the agents con- 

 trolling it, a subject which I dealt with briefly in a paper read 

 before this Association at Stellenbnsch.* The influence of Termites 



* Bews, J. W. "The Plant Succession in the Thornveld." SA. Journ. 

 of Science. 1917. 



