Analysis of geological materials may be the most challenging application for nuclear microprobes. Quantitative multi-element PIXE results are required on thick, compositionally complex, fine-grained materials. Major elements such as calcium and iron dominate the spectra that include numerous X-ray peak overlaps from a wide range of minor and trace elements. A concerted effort is required to optimize sample selection, data acquisition, and data reduction. Geochemical complexities can also make RBS and nuclear reaction analysis difficult. At Los Alamos, we are rebuilding our superconducting-solenoid nuclear microprobe beam line. We have moved to a new 3 MV tandem accelerator, simplified our ion optics, added two-dimensional beam scanning, and are working on a new precision stage and vacuum chamber. Continuing development of our software is aimed at obtaining maximum compositional information from PIXE spectra. In addition to the geochemical work of Cahill (UCD), Cabri (Ottawa), Clayton (Lucas Heights), and Minkin (USGS/Heidelberg), we (and our collaborators) have addressed several meteoritic and terrestrial problems. These range from early solar system condensation, planetary differentiation, and the possibility of “martian” meteorites, to in-situ retorting of oil shale, mid-ocean ridge hydrothermal processes, and the deduction of American indian trade routes through the characterization of the source minerals used in their copper metallurgy. Natural and synthetic standards have demonstrated that PIXE spectra can yield quantitative concentrations for more than 60 elements in a single spectrum and that, in favorable cases and with effort, sub-ppm detection limits are attainable.