The theory of stress corrosion for slow crack propagation is reviewed in the light of classical Griffith theory of fracture. Experimental data for stress corrosion cracking for glasses, ceramics, and metals are reviewed. We suggest that stress corrosion cracking plays an important role in the intrusion of magmas and in the transport of magmas upward through the lithosphere. It is shown that the effect of decreasing temperature (at progressively shallower levels along the geotherm) would be to decrease the crack velocity by several orders of magnitude if other factors were equal. We also propose that stress corrosion may be an important process in time-dependent earthquake phenomena such as premonitory behavior and earthquake aftershocks. We suggest that slow cracking in the earth is not seismically detectable but may nevertheless precede the terminal (catastrophic) phase of the fracture that is discerned as an earthquake. The seismically quiet periods before some earthquakes and the seismically quiet regions beneath some volcanoes may in fact be regimes of slow crack propagation. Slow crack propagation in a lithospheric plate may provide access routes for magmas which give rise to prominent linear volcanic chains.