PREFACE 



Investigations made by the United States Department of Agri- 

 culture during the last ten years have shown that the Central Ameri- 

 can rubber tree is not. as generally accepted, a homogeneous species, 

 but that it consists of several well characterized forms which differ 

 materially in their geographic distribution and climatic correlation. 

 Hitherto rubber planters have made no distinction between these 

 forms, and many of the unsuccessful attempts at the cultivation of 

 Castilla as a rubber-producing tree may be traced to the want of 

 discrimination in the choice of the form from which the seeds were 

 taken. 



In the present paper, by Mr. Henry Pittier. an attempt has been 

 made to describe systematically these various forms, along the lines 

 indicated in a former paper by Mr. (). F. Cook. Although Mr. Pittier 

 has brought to (his work a long practical experience in the field, and 

 has had an abundant supply of material for study, our knowledge of 

 the various forms will still be far from complete. Their number may 

 be reduced or increased and their relative status altered. Neverthe- 

 less, the present paper will remain as the first attempt at a thorough 

 study of the group. 



In recent years new species have been discovered in South America, 

 and the areal limits of the genus have thus been extended far beyond 

 the Isthmus to Bolivia and eastern Brazil. It is almost certain that 

 other species, hitherto hidden in (he unexplored forests of South 

 America, will come to fill the gaps in the distribution of the genus. 



Frederick V, Coville, 

 Curator of the United States National Herbarium. 



