Abstract. The saturated hydrocarbons propane and the butane isomers are both indirect greenhouse gases and key species in liquefied petroleum gas (LPG). Leakage of LPG and its component alkanes/alkenes is now thought to explain a significant fraction of the volatile organic burden and oxidative potential in the basin which confines Mexico City. Propane and the butanes, however, are stable enough to escape from the basin. The gas chromatographic measurements which have drawn attention to their sources within the urban area are used here to estimate rates of ventilation into the free troposphere. The calculations are centered on several well studied February/March pollution episodes. Carbon monoxide observations and emissions data are first exploited to provide a rough time constant for the removal of typical inert pollutant species from the valley. The timescale obtained is validated through an examination of meteorological simulations of three-dimensional flow. Heuristic arguments and transport modeling establish that propane and the butanes are distributed through the basin in a manner analogous to CO despite differing emissions functions. Ventilation rates and mass loadings yield outbound fluxes in a box model type computation. Estimated in this fashion, escape from the Valley of Mexico constitutes of the order of half of 1% of the northern hemispheric inputs for both propane and n-butane. Uncertainties in the calculations are detailed and include factors such as flow into the basin via surface winds and the size of the polluted regime. General quantification of the global propane and butane emissions from large cities will entail studies of this type in a variety of locales. In the discussion section the arguments connecting carbon monoxide and alkane fluxes from the valley are formalized through modeling. Simple one-dimensional simulations of the ventilation process are constructed, and limited application is made of some three-dimensional photochemistry/transport codes at our disposal. We deal qualitatively with the potential for surface winds to import the tracers of interest. Other major uncertainties in the analysis are then listed. Sources of error include our definition of the basin and urban periphery, the average height of the low nocturnal inversion, and alkane distributions. We conclude with some comments on the global ramifications of our findings. By our calculations, Mexico City in and of itself constitutes a significant source of the C3 and C4 alkanes to the terrestrial atmosphere. Recent sampling suggests that many other nondomestic (non-United States) urban areas contribute propane and the butanes through LPG leakage. We note that megacities with pollutant characteristics akin to those of Mexico are developing rapidly worldwide [United Nations, 1992;Kretzschmar, 1993Kretzschmar, , 1994Goldemberg, 1995]. To better quantify their role in the global tropospheric ozone system, measurement of LPG component signatures will be required in many urban settings. We also outline relationships between ...