Of all the areas of Latin American Spanish, the Central American isthmus is the least studied in terms of dialectological investigations. Early studies' lumped all the Central American republics under a general rubric, which at times also included the Yucatan Peninsula of Mexico, while studies directed at specific nations assumed an exclusively lexicological character.2 In recent years, Central America has fared little better, for the area has remained virtually untouched in linguistic studies. The paucity of linguistic investigations becomes even more critical in the phonological dimension, and only quite recently has this area figured in linguistic analyses of Central American Spanish.3 During the Spanish colonial empire, Central America was culturally isolated from the mainstream of Spanish America, with the exception of the capital of Guatemala. As a result, Central American Spanish constitutes a linguistic microcosm with a certain coherency in its own right as opposed to other main currents of Hispanic dialectology, but with the same tantalizingly emphemeral characteristics which form the spirit of Central American solidarity. In some cases Central American dialects have participated in phonological processes common to most of the rest of Latin America, but the historical factors which have led to the unique cultural configurations have also permitted a number of divergent developments and even some innovations. This article has a double purpose. The first is to offer a concrete description of one facet of Central American phonology, namely the behavior of the phoneme /s/. This choice of operating parameter represents one of the key defining characteristics of Hispanic dialectology, and enables the Central American data to be fitted against the already existing studies of the behavior of /s/ in other Spanish-speaking regions. The second thrust, which in this article will remain subordinated to the first, is to consider the possible diachronic processes which could have given rise to the current configurations, and the theoretical consequences implied by the reconstructions.4 In order to observe the details of the distribution of /s/ in Central America, I collected data on the speech of the capital cities. During periods of field work5 recordings were made in the five Spanish-speaking Central American republics. In the present discussion, the speech patterns of Guatemala City, San Salvador, Tegucigalpa, San Jose and Managua will be compared. As a prelude to the quantitative presentations, we observe that these data are generally representative of the dialect patterns of the respective countries, and that the internal regional differences which exist are mainly differences of degree of application of certain phonological processes. Naturally, considerable sociolinguistic stratification exists, especially in urban areas, where a broader spectrum of sociocultural groups may be found. Guatemalan Spanish phonology is quite homogeneous, except for the small Caribbean costal area and part of the Pacific coastal region near the Salvadoran border. The Salvadoran and Nicaraguan dialects are similarly homogeneous, with few regional differences being discernible in either country. Honduran Spanish is much more varied, since the major transitions between the phonetically conservative Guatemalan dialect and the phonetically advanced Nicaraguan region occur within the borders of Honduras. The capital, Tegucigalpa, provides an intermediate point among the various influences; however, the regional differences within Honduras are quantitative rather than qualitative. Costa Rican Spanish is the most heterogeneous, since the central highlands dialect, including the capital, is radically different from neighboring Nicaragua and Panama. However, the northern provinces which border on Nicaragua exhibit phonetic patterns indistinguishable from those of Nicaragua, while the regions bordering on Panama exhibit Panamanian phonetic characteristics. The dialect of the meseta central therefore occupies a unique position as a phonological oasis in the midst of two dialects which are characterized by a more advanced degree of phonetic reduction. For the present study, ten informants were