106 



spikelets are very small and provided with 2 thin, rather long awns; 

 over land they are easily dispersed .by wind. It is quite unknown 

 whether animals also take part in their dispersal. The statement of 

 Docters van Leeuwen 1 ) that this species is spread by 

 birds is apparently not based on actual observation, but seems to be 

 a mere guess, lacking probability. Though in 1 897 this grass may 

 have plentifully occurred on steep banks on Krakatao, it cannot 

 possibly have taken part in the savanna-vegetation proper.- Imperata 

 cylindrica P. B, is the notorious alang alang, a grass spread all over 

 Java from the plains up to an altitude of 2700 m. In favorable 

 localities it multiplies with marvellous rapidity, not only by the growth 

 and ramification of its rather deep-lying and far-creeping rhizomes 

 but also by its fruits, which remain enclosed within the spikelets. 

 These spikelets are clothed with very long white hairs; when ripe 

 they disarticulate from their pedicels and are dispersed far and 

 wide by the wind: one often sees them by thousands floating through 

 the air. It is unknown how far the spikelets can be carried by the 

 wind in a single flight but as to this matter I refer to what I have 

 said on p. I0l on the dispersal of the fruits of Saccharum spontaneum. 

 Impcrata is no kremnophyte; the plant never grows an vertical 

 walls but on plains, ridges and slopes which, without being very dry, 

 yet are not sufficiently humid for a dense growth of trees that would 

 kill this heliophilous grass by their shadow; moreover it is very often 

 found in deserted fields formerly used for agricultural purposes and 

 on spots where forests have been destroyed by fire or by the axe. 

 In such localities it often predominates in countless numbers, it forms 

 those vast fields so vividly described by ) u n g h u h n -). Imperata 

 always grows intermixed with other plants which in a young alang- 

 field may be hidden by the high grass-blades, but continually increase 

 in size and number, wherever the alang-fields are left undisturbed. 

 In the grass-jungle gradually young forest develops, except in too dry 

 localities; the groves of trees at first form small islands which after- 

 wards blend into each other ;i ) and finish by killing out the grass. 



') Handelintjen Eerste Natuurwetenschappeliik Congres (1!)20), p. 62.- Annales 

 lardin Botanique Buitenzorg XXXI (192I), p. 131. 



2 ) lunghulin, ]ava. Second Dutch Edition I, 289; German Edition I, 212. 



") I he natives often wish to preserve the alang-fields which yield old leaves for 

 thatching, young leaves for fodder. In order to renovate the pasture and destroy the 

 young trees they set fire to the alang-alang in the east-monsoon, when many of the 

 old leaves are dead and quite dry. Hie fire Uills the young forest-growth but does 

 not damage the deep-lying rhizomes of the grass which soon afterwards emit young 

 leaves, junghuhn "*) wrongly meant that the alang-alang was set fire to in order 

 to destroy it, just the reverse is aimed at. The fire is the pitiless foe of the young 

 forest, the protecting friend of the grass. 



') lunghuhn, |ava. Second Dutch Edition I, 296; German Edition I, 126- 



