Ten-day backward trajectories are used to determine the origins of air parcels arriving at locations of airborne DC-8 chemical measurements during NASA's second Pacific Exploratory Mission in the Tropics B that was conducted during February-April 1999. Chemical data at sites where the trajectories had a common geographical origin and transport history are grouped together, and statistical measures of chemical characteristics are computed. Temporal changes in potential temperature are used to determine whether trajectories experienced a significant convective influence during the 10-day period. Trajectories describing the aged marine Southern Hemispheric category remain over the South Pacific Ocean during the 10-day period, and their corresponding chemical signature indicates very clean air. The category aged marine air in the Northern Hemisphere is found to be somewhat dirtier. Subdividing its trajectories based on the direction from which the air had traveled is found to be important in explaining the various chemical signatures. Similarly, long-range northern hemispheric trajectories passing over Asia are subdivided depending on whether they had followed a mostly zonal path, had originated near the Indian Ocean, or had originated near Central or South America and subsequently experienced a stratospheric influence. Results show that the chemical signatures of these subcategories are different from each other. The chemical signature of the southern hemispheric long-range transport category apparently exhibits the effects of pollution from Australia, southern Africa, and South America. Parcels originating over Central and northern South America are found to contain the strongest pollution signature of all categories, due to biomass burning and other sources. The convective category exhibits enhanced values of nitrogen species, probably due to emissions from lightning associated with the convection. Values of various species, including peroxides and acids, confirm that parcels were influenced by the removal of soluble gas and particle species due to precipitation. Finally, current results are compared with those from the first PEM-Tropics mission that was conducted in the same region during the southern hemispheric dry season (August-October 1996) when extensive biomass burning occurred. Results show that air samples during PEM-Tropics B are considerably cleaner than those of its dry season counterpart. Copyright 2001 by the American Geophysical Union.