'Tis a land that still with potent charm And wondrous, lasting spell With mighty thrall enchaineth all Who long within it dwell 'Tis a land where the Pale Destroyer waits And watches eagerly; 'Tis, in truth, but a breath from life to death, In the Land of the Cocoanut Tree. - James Stanley Gilbert, Panama Patchwork, 1908 During the late nineteenth and early twentieth centuries, as tropical Central America came under increasing U.S. influence, U.S. policymakers, businesspeople, missionaries, and bureaucrats began to transform the region to meet their needs.(1) They built railways, led military invasions, established banana and coffee plantations, and eventually dug a canal across Panama. Their published accounts and artistic renderings of Central America drew on more generalized, archetypal ideas in the art, history, literature, and photographs of tropics around the world to form a specific discourse about the Central American tropics.(2) Two opposing narratives constituted this discourse: positive ones about Edenic paradises, fertile soil, and exotic beauty; and negative ones about moral laxity, dangerous landscapes, disease, and the threatening abundance of the jungle. These varied ways of seeing Central America revealed themselves in more than just semantic representations; they influenced U.S. actions and policies in the tropics. These contradictory narratives were used to legitimate imperialist intervention and actions in the Panama Canal Zone in the early twentieth century. THE DISCOURSE OF THE TROPICS Lines of latitude have long been used to demarcate tropical regions. Aristotle, for example, separated the world horizontally into frigid, temperate, and torrid (tropical) zones. Today the tropics are depicted as the region that lies between 23 [degrees] 30 [minutes] north latitude and 23 [degrees] 30 [minutes] south latitude - the Tropics of Cancer and Capricorn. Alternatively, the tropics have been defined using temperature and precipitation isolines. These have varied somewhat, from Ellen Churchill Semple's (1911) inclusion of areas within the 20 [degrees] mean annual isotherm to Isaiah Bowman's (1937, 381) use of the 25 [degrees] mean annual isotherm. A physical geography textbook today designates them as lands within the 18 [degrees] mean annual isotherm (Strahler and Strahler 1996, 165). Agreement on the climatic character of the region - the heat and humidity associated with the tropical lowlands - is more general. In fact, highland areas such as the Meseta Central of Costa Rica were excluded from the tropical discourse because, as one traveler stated, they are as cool and healthful as the coastal plains are torrid and fever-infested (Putnam 1913, 10). The interpretation of tropical heat and humidity has shifted over the past two centuries, however. At various times North Americans have imagined the Central American lowland tropics as distant paradises or fever coasts (in the nineteenth century), as banana republics (in the early to mid-twentieth century), as places for revolution (in the 1970s and 1980s), and as sites for romantic ecotravels (in the 1990s). In this article I use the word tropics to refer to the lowland tropics, and I deal exclusively with the late nineteenth and early twentieth centuries. In this geographical and historical context, the character of tropical climates was frequently expressed in subjective terms, even when placed within a so-called scientific framework. Turn-of-the-century U.S. natural science textbooks typically included a classification of the tropical flora, fauna, temperatures, and diseases in Central America. Such descriptions, however, were frequently intertwined with an author's opinions concerning heat, disease, dark-skinned peoples, hot or spicy foods, exotic fruit, fecund vegetation, and economic underdevelopment. For example, in his 1918 Handbook of Commercial Geography, Geo. G. Chisholm describes the specific rainfall amounts, humidity, and temperature that are characteristic of the tropics while noting the excessive heat and irksome humidity (p. …