Upper mantle peridotites from ophiolitic and alpine-type massifs, as well as those brought up as basalt-borne xenoliths, show a spectrum of structures and microstructures that potentially reflect the mechanical behaviour of the upper mantle during flow. It is uncertain, however, if microstructures are unambiguous in this respect. Based on experimental studies and materials science know-how, we use microstructure in its broadest sense to infer the nature of the microphysical processes once operating in the upper mantle rocks in question. Aside the many uncertainties involved in such extrapolation of experimentally determined mechanical data via the microstructure, there is the even more difficult question if the flow and fine structure of a natural rock, deformed at depth and high temperature, really reflects the flow of interest. Inherent to the fact that these rocks are now found at the Earth’s surface, the microstructures seen in naturally deformed mantle rocks may to a certain extent represent the waning stages of deformation and recrystallization in the mantle, i.e., the microstructure frozen in the rock may not appropriately reflect the main flow. The question thus arises in how far the deformation microstructures in mantle rocks really represent a significant part of the ductile flow history. In this context it seems appropriate to use the concept of a steady-state microstructure, i.e., a microstructure that continuously evolves during flow without changing its principal characteristics such as average grain size, grain shape and grain orientation, and that likely reflects the microphysical processes during flow. These concepts are particularly important at high homologous temperatures where deformation-induced rotation recrystallization (development of new grains from subgrains during dislocation creep) may or may not be balanced by migration recrystallization (recrystallization due to the migration of grain boundaries). It may be envisaged that high-temperature flow in the asthenosphere or lower lithosphere thus leads to seemingly undeformed granular rocks, even protogranular in the sense of commonly used peridotite microstructure classifications, or that the samples recovered preserve more clearly deformed, porphyroclastic microstructures including coarse to fine-grained tectonites. A distinct class of microstructures in peridotites concerns very fine to ultrafine-grained deformation microstructures in sometimes spectacular mylonitic shear zones. The mechanical aspects of the initiation and growth of these structures is surrounded with uncertainties, but there is growing consensus regarding their mechanical significance once these structures have developed: they are mechanically extremely weak due to the fact that flow in such ultrafine-grained rocks will be dominated by grainsize-sensitive diffusion creep mechanisms. Mylonites in upper mantle rocks often overprint earlier flow structures, probably because they tend develop at lower temperatures and higher stresses inherent in the later stages of a mantle fragment’s history in the Earth’s upper mantle, e.g., related to emplacement at crustal levels in rifts, or near the ocean floor in ophiolites.