Using crystals grown into vugs from oriented lunar rocks as directional detectors of cosmic dust particles and solar flare nuclei, I have measured the angular distribution of the flux of micrometeoroids of mass ~3¿10-17 to ~8¿10-14 g. Hypervelocity impact craters of diameter 500 ¿ to 10 μm and tracks from solar flare nuclei of energy ~100 keV/amu to ~20 MeV/amu were observed in crystals from rock 71055 facing lunar south and in crystals from rock 74255 facing lunar east. The vug geometry restricts the region of space viewed by the crystal detectors such that only particles streaming nearly normal to the ecliptic could enter the vug from 71055, while only particles confined to the ecliptic could enter the vug from 74255. By comparing the size-frequency distributions and circularity distributions of craters observed on these orthogonal crystals, I have found that dust grains both in ecliptic orbits and in orbits inclined to the ecliptic have virtually identical mass-frequency distributions and similar shapes. I have evaluated the micrometeoroid fluxes after determining the exposure ages of the individual vug crystals by measuring the solar flare track density gradients in each crystal. The flux of particle of mass >4¿10-15 g confined to the ecliptic is ~1.1¿10-2/m2 s 2&pgr; sr. This value is in good agreement with satellite measurements, suggesting that the micrometeoroid flux has been relatively constant over the past 35,000 yr. The flux of particles producing microcrater on a lunar surface facing south is lower than the ecliptic flux by a factor of ~7. This experiment was unable to distinguish between interstellar dust or interplanetary particles in highly inclined orbits as the source of craters observed on the south-facing surface. The solar flare energy spectra observed in vug crystals from both 71055 and 74255 deviate from the long-term spectrum determined by Hutcheon et al. (1974) at energies below ~10 MeV/amu. I conlude that the shallow energy spectrum cannot be a result of erosion and suggest that the experimental data from which the long-term low-energy Fe spectrum has been derived must be reexamined carefully.