We have determined the major and trace element compositions of 176 individual microtektites/spherules from the Australasian, Ivory Coast, and North American microtektite and clinopyroxene-bearing (cpx) spherule layers. Trace element contents for up to 30 trace elements were determined by instrumental neutron activation analysis (INAA), and major element compositions were determined using energy dispersive X-ray (EDX) analysis in combination with a scanning electron microscope (SEM). In addition, petrographic data were obtained for the cpx spherules using the SEM and EDX. This is the first trace element study of individual Australasian microtektites, and the data revealed the presence of a previously unrecognized group of Australasian microtektites with high contents of Ni (up to 471 ppm). In previous studies the high-Mg (HMg) Australasian microtektites were thought to be related to the HMg Australasian tektites, but our trace element data suggest that the high-Ni (HNi) Australasian microtektites, rather than the high-Mg microtektites, are related to the high-Mg Australasian tektites. We find that Cenozoic microtektites/spherules from a given layer can be distinguished from microtektites/spherules from other layers as a group, but it is not always possible to determine which layer an individual microtektite/spherule came from based only on trace element compositions. The cpx spherules and most of the microtektites have Cr, Co, and Ni contents that are higher than the average contents of these elements in the upper continental crust, suggesting the presence of a meteoritic component. The highest Cr, Co, and Ni contents are found in the cpx spherules (and low-Si cpx-related microtektites). Unetched to slightly etched cpx spherules have Ni/Cr and Ni/Co ratios that generally lie along mixing curves between the average upper continental crust and chondrites. The best fit appears to be with an LL chondrite. The moderately to heavily etched cpx spherules have values that lie off the mixing curves in a direction that suggests Ni loss, probably as a result of solution of a Ni-rich phase (olivine?). The Ni-rich Australasian microtektites also have Ni values that lie close to mixing curves between the average upper continental crust and chondrites. However, both the cpx spherules and HNi Australasian microtektites appear to have Ir (and to a lesser extent Au) contents that are much too low to have Ni/Ir ratios similar to chondritic values. We have no explanation for the low-Ir and -Au contents except to speculate that they may be the result of a complex fractionation process. The Ivory Coast and North American microtektites do not have high enough siderophile element contents to reach any firm conclusions regarding the presence of, or nature of, a meteoritic component in them. Trace element compositions are consistent with derivation of the Cenozoic microtektite/spherule layers from upper continental crust. The normal Australasian microtektites appear to have been derived from a graywacke or lithic arenite with a range in clay and quartz content. The source rock for the high-Mg Australasian microtektites is not known, but the HMg microtektites do not appear to be normal Australasian microtektites that were simply contaminated by meteorites or ultramafic rocks. The average Ivory Coast microtektite composition can be matched with a mixture of target rocks at the Bosumtwi crater. The average composition of the North American microtektites suggests an arkosic source rock, but with graywacke and quartz-rich end members. However, we could not match the composition of the North American microtektites with lithologies in impact breccias recovered from the Chesapeake Bay impact structure that is believed to be the source crater. Likewise, we could not match the composition of the cpx spherules with mixtures of basement rocks and overlying sedimentary deposits (for which compositional data are available) at the Popigai impact crater that may be the source crater for the cpx spherules. This may be because the cpx spherules were derived, in large part, from clastic surface rocks (sandstones and shales) for which no compositional data are available.