Our exploration of the Moon and planets began with telescopic studies of their surfaces, continued with orbiting spacecraft and robotic landers, and culminated (or will someday we hope) with manned exploration and sample return. For the Moon and Mars we also have accidental samples provided by impacts on their surfaces, the lunar and martian meteorites. How much would we know about the lunar surface if we only had lunar meteorites, orbital spacecraft, and robotic exploration, and not the Apollo and Luna returned samples? What does this imply for Mars? With martian meteorites and data from Mariner, Viking and the future Pathfinder missions, how much could we learn about Mars?

Remote exploration of the Moon and planets began in antiquity with the unaided eye, picked up speed with the inventions of telescopes and spectroscopy, and reached its stride with modern instruments aboard orbiting spacecraft. Robotic exploration of the Moon included photography by lunar orbiters, surface soil analyses by the Surveyor and Luna landers, geochemical mapping by the Apollo orbiter -and x-ray experiments, and current mineralogical mapping by Clementine reflectance spectrometers. Yet the basis of most of our detailed knowledge about the Moon is the Apollo samples. They provide ground truth for the remote mapping, timescales for lunar processes, and samples from the lunar interior.

Could the lunar meteorites, which are random samples from 5-9 different sites, tell us as much about lunar evolution? The first order conclusion of the Apollo program is that the major processes on the Moon's surface are impact and volcanism. Many of the Apollo samples are breccias formed by impacts. Seven of the nine lunar meteorites are breccias. The Apollo samples showed that the highlands are anorthositic and the mare are basaltic. Four of the lunar meteorites are anorthositic; four are basaltic. With geochemical mapping and imagination, we might have arrived at the magma ocean model for lunar evolution. However, it is the diversity of the Apollo samples that is missing from the lunar meteorite suite. The most notable missing rock types are Mg-suite plutonic rocks, KREEP and evolved rocks, and high-Ti mare basalts. We might have been able to detect these rocks with geochemical or mineralogical mapping, but might not have recognized them without knowing from the Apollo samples that they were there. For example, we would probably have attributed the high K, U, Th in the Imbrium region to granite and not KREEP. Thus the lunar meteorites and remote sensing would provide us with the first order conclusions from the Apollo samples, but not the diversity of rock types. Therefore the complexity of lunar petrogenesis and completeness of our models of lunar crustal evolution would be reduced.

The Moon is the foundation of planetary science and the basis for our interpretation of the other planets. Mars is similar to the Moon in that impact and volcanism are the dominant processes, but Mars' surface has also been affected by wind and water, and hence has much more complex surface geology. Martian meteorites are all young basaltic or ultramafic igneous rocks from the young northern volcanic province. They tell us relatively little about the range of rock types on Mars. Future geochemical or mineralogical mapping of Mars' surface should be able to tell us whether the dominant rock types of the ancient southern highlands are basaltic, anorthositic, granitic or something else, but will not be able to tell us the detailed mineralogy, geochemistry, or age. Without many more martian meteorites or returned samples we will not know the diversity of martian rocks, and therefore, will be limited in our ability to model martian geological evolution. Mars may look simply like a "wet" Moon, and not like the complex planet we expect it to be.