Inorganic gas phase halogen species (e.g. Cl, HOBr, and ClONO 2 ) likely play an important role in the oxidative chemistry of the troposphere. These reactive halogen gases arise when halide salts are oxidized, a process that occurs in sea sea salt aerosols throughout the marine boundary layer. Reactive halogen cycling influences tropospheric ozone levels, and the Cl atom in particular destroys the greenhouse gas methane. Both of these roles may significantly influence global atmospheric chemistry and climate. However, the levels of reactive halogens and their activation mechanisms are not known well enough to quantitatively assess the global importance of these species. This is largely due to a paucity of ambient measurements. This dissertation describes measurements and modeling of reactive halogen species in the marine atmosphere, focusing on reactive chlorine species in the semi-remote atmosphere. Measurements of Cl 2 were made in 2007 at the Cape Verde Atmospheric Observatory, and measurements of Cl 2 and HOCl were made at the same site in 2009. These are the first measurements of HOCl in the troposphere, and the first measurements of Cl 2 far from highly polluted coastal areas. Both HOCl and Cl 2 are important precursors of the highly reactive Cl atom. The measurements show large variability in levels of reactive chlorine (Clx: Cl 2 , HOCl, BrCl, ClONO 2 , etc.). An estimate of the global impacts of Clx based on extrapolating these measurements would therefore have a large uncertainty. The highest HOCl and Cl 2 mixing ratios were associated with aged polluted air from Europe. This is probably due to enhanced chlorine cycling caused by increased aerosol acidity. On these days Cl atoms accounted for an estimated 15% of marine boundary layer methane destruction. Cleaner marine air that arrives at the site has lower levels of reactive chlorine, and on these days Clx has only a minor impact on photochemistry. These findings are generally consistent with model predictions. However, the absolute levels of Clx and relative abundances of HOCl and Cl 2 differ between model and observations. In particular, observed HOCl levels were a factor of 2-5 times higher than predicted by a model simulation. We need to reconcile these differences in order to make accurate estimates of the global impacts of Clx chemistry, both now and in the future. This requires improving our understanding of the mechanisms of Clx activation and cycling.