CHAr. 11] PERIODIC PHENOMENA IN THE TROPICS 245 



AiiliiDiii (February to April). Continuation of the rainy season, after an interrup- 

 tion of dry wcatlier in January to February. Tlic leaf-fall begins in March, before 

 the end of the rainy season. 



The greater the precipitations during the rainy season, tlie more do the 

 evergreen trees predominate over those that pcriodicall}- shed their leaves. 

 In the dense forests of constantly humid districts, defoliation occurs only 

 in the gigantic trees, whose crowns rise like cupolas above the general 

 leaf-canopy and are therefore more exposed to drought. Among these 

 periodically leafless giants may be reckoned fig-trees in particular ; also 

 the rasamala, Altingia excelsa, the tallest tree in the dripping forests of 

 Java, is bare of foliage in a short time after the commencement of the dry 

 east wind. 



In all tropical districts with very weak climatic periodicity, there are 

 woody plants tliat sited ilicir leaves at longer or shorter intervals {one to six 

 times a year), without any eonncxion ivith the season of the year, so that 

 trees of one and the same species, under the same external conditions, 

 acquire fresh foliage and shed their leaves at times that do not agree. 

 Thus, for instance, I saw at Singapore trees of flame-of-the-forest, Poinciana 

 recia, erowine together with and without foliage, and I have noticed the 

 same behaviour in Terminalia Catappa at many places. Haberlandt makes 

 a similar statement regarding Palaquium macrophj'llum at Buitenzorg. 

 The period during which such trees remain leafless is usually very short — 

 one to two days, for instance, for Excoecaria Agallocha. Acer niveum, many 

 species of Urostigma ^ 



In the botanic garden at Buitenzorg I have carefully observed Urostigma 

 glabcUum, a gigantic tree which sheds its leaves and produces new foliage 

 about every two months. On December to, 1889, the whole foliage fell 

 during the day while still perfectly green, so that the tree that appeared 

 quite verdant in the morning became bare of leaves by evening. On 

 December 20 the foliage had been almost completely renewed. One of the 

 young shoots which had been formed in the interval and was plucked 

 haphazard was 26 cm. long from its base to the tip of its topmost leaf, the 

 axis alone 12 cm. long, and the blade of the third oldest leaf 13 cm. long; 

 that of the youngest 8 cm. 



Such cases of the loss and restoration of the leaves independently of the 

 season of the year can only be due to internal causes. Frequentlj^ such 

 a leaf-fall is a sign that the tree is preparing to blossom. In Urostigma 

 glabellum and in many other species such a connexion does not exist, or 

 not necessarily so. 



It suggests itself that in such cases the swelling buds draw the transpiration- 

 current to themselves ; but no experiments have been made in relation to this. 



' Koorders and Valeton, op. cit. 



