This paper presents dry deposition of major ions on tropical foliage (leaves of Ashok (Polyalthia longifolia) and Cassia (Cassia siamea)) at St. John's, Agra, an urban site of tropical India on nonrainy, nondewy, and nonfoggy days. The deposition flux was higher on Cassia leaf than Ashok leaf probably due to a rougher surface as shown by scanning electron microscopy. Dry deposition of cations varies from 0.46 to 12.16 mg m(-2) day(-1) while anions vary from 0.04 to 3.24 mg m(-2) day(-1). The percentage contribution of alkaline components is greater than that of acidic components, indicating the alkaline nature of dry deposition. Two-way analysis of variance results reveal significant seasonal variation only for K+, SO4(2-), and F-; however, values varied season to season for Na+, Ca2+, Mg2+, Cl-, NO3-, and NH4+ also. The large seasonal variation in deposition flux may be due to meteorological conditions, diameter of particles, and variation in atmospheric level. SO42- and NO3- show significant correlation, indicating their origin from similar sources while significant correlation between Ca2+ and Mg2+ implies their origin from soil. Poor correlation between Ca2+ and SO4(2-), Ca2+ and NO3-, and Mg2+ and SO4(2-) indicates that in addition to soil other sources also contribute to dry deposition. Low dry deposition fluxes of SO2- and NO3- compared to Ca2+ and Mg2+ may be due to low mass medium diameters of SO4(2-) and NO3- and may be due to uptake through the stomatal pores abundant on leaf surfaces. Factor analysis was employed to identify the sources. F-, Cl, SO4(2-), NO3-, and K+ are grouped together in the first factor, indicating their probable contribution from combustion, Ca2+, Mg2+, and NH4+ are grouped in factor II, which may be attributed to road dust and soil, and factor III includes mainly Na+ and F-, probably contributed from brick-kiln industries. Atmospheric concentrations of F-, Cl-, NOs-, SO4(2-), Na+, K+, Ca2+, Mg2+, and NH4+ were found to be 0.38, 2.28, 1.31, 2.74, 0.44, 0.59, 1.21, 1.2, and 2.29 microg m(-3), respectively.