EFFECTS OF SHADE. 51 



practicable if the weeds and undergrowth cut down in the dry season 

 could be left spread over the ground. This would do more to con- 

 serve the moisture of the soil than the same vegetation alive, but the 

 danger of tire will in most localities forbid the use of this method of 

 culture. 



If the present question could be settled by deciding whether or not 

 Castilla needs to be protected from the sun, it would be easy to estab- 

 lish the negative view; but with shade recognized as a means of 

 influencing natural conditions of soil or climate it becomes evident 

 that each planter will need to use his best judgment in determining 

 what local conditions require. In Costa Rica Koschny advises the 

 thinning of the forest b} T the removal of two or three trees out of 

 every live. At La Zacualpa more are cut out (PL XV). Some of 

 the planters on the Isthmus of Tehuantepec practice clean culture. 

 No general principles will determine which is best, because no one 

 method is applicable everywhere. 



RELATIVE COST OF SHADE CULTURE. 



It must be remembered, in addition, that the planter finds himself 

 compelled to decide not what will be the best for the rubber trees, 

 but what is the best he can afford to do for them. Is it, for example, 

 good policy to use labor and capital in keeping a tract of planted land 

 clean, or will more be gained ultimately if one contents himself with 

 somewhat slower growth and improves the opportunity 7 of planting 

 additional tracts with trees that can also be growing? Careful com- 

 parative experiments might be necessary for an answer, and this might 

 differ for different localities. 



EFFECT OF SHADE ON FORM OP TREE. 



There are great and persistent differences of shape or u habit" 

 among trees. The Lombard v poplar and the weeping willow are not 

 distant relatives. It is a general fact, however, that forest trees are 

 taller and more slender than those of the same species grown in the 

 open. The low spreading habit, which is desired and encouraged 

 among fruit trees, is not desirable in rubber-producing species, where 

 a large expanse of trunk is needed to supply the milk and to give 

 opportunit}- for tapping- without the necessity of wounding the same 

 place too often. Castilla trees growing alone in the open often send 

 out permanent branches 8 or 10 feet from the ground, while those in 

 the forest may have from 20 to 10 feet of smooth trunk before the 

 permanent branches are reached. Open-grown trees may have large 

 spreading branches, while in the forest or under close planting the 

 main axis of the tree continues to grow upward and the lateral 

 branches are relatively small. 



The problems of rubber culture ma} T prove in this respect to be 

 directly opposite to those of coffee, where the formation of much wood 



