172 THE PLANT WORLD 



In Java a fig studied by Mr. D. G. Faircliild dropped its leaves com- 

 pletely and put out an entire new set within a few days. With this 

 species the reasons for the leaf-fall seemed to be physiological and in- 

 ternal ; no connection with the seasons was apparent. Similar differ- 

 ences exist among temperate trees, the structure of some of which 

 insures the prompt falling of the leaves in autumn, while others may 

 hold the dead and useless leaves till spring. 



Tropical trees may be said, therefore, to furnish a beginning from 

 which the deciduous habit could readily be developed with the encour- 

 agement, first of dry seasons and then of frosts. It is now known that 

 the injuries of frost are not due alone to the direct effects of the cold 

 upon the protoplasm, but to the loss of water, which constitutes so large 

 a percentage of living matter. Experience in deserts or in regions with 

 long dry seasons would thus be an important preliminary to enable 

 trees to exist in temperate regions. The forests of the arid interior of 

 Guatemala and Mexico contain many trees deciduous in the dry season 

 and for these entrance to higher altitudes or latitudes would be rela- 

 tively easy. Such a transition would also be further facilitated by the 

 fact that in this part of the world the dry season comes in winter, while 

 if the rainy season of active growth came in winter, as in some parts of 

 the world, the opportunities for extension into colder localities would 

 be greatly lessened. 



It may be held, therefore, that the existance of deciduous trees is 

 not to be explained through simple selection by frost, but is rather a 

 specialization and extension of the alternations of active growth and 

 rest normally existing in plants. 



In the moist tropics it might be said to be an advantage for a tree 

 to hold its leaves even in a period when it does not use them, since to 

 let them fall would be to encourage the growth of competing vegetation, 

 but even here there may be special causes favoring the deciduous habit. 

 Such an instance was recently noted in the forest-covered region of 

 Eastern Guatemala. Along the Rio Dulce and the Golfete there is a 

 magnificent Bignonaceous tree with peach-blow flowers, borne in enor- 

 mous profusion when the leaves are entirely wanting. The result is a 

 mass of color visible for a distance much greater than would be possible 

 if leaves were intermixed, and thus much better able to attract the atten- 

 tion of humming birds and other insectivorous species which visit the 

 flowers. Cross-fertilization is thus secured from much greater distances 

 than would be provided by short-sighted insects, and a variation in the 

 direction of dropping the leaves would have a great advantage in be- 

 coming widely spread. In such a case natural selection might be said 

 to take place between variations rather than between individual plants. 

 Trees which dropped their leaves might have been no more fit to sur- 



