Banded fibrous veins are often pointed to as evidence for episodic crack opening driven by oscillations in fluid pressure or bulk strain. Advances in understanding the geochemistry of precipitation, data on veins, and experiments suggest that pressure due to growing crystals may be an alternate explanation for many observations on these types of veins. We propose that some veins originate at sites of precipitation and then propagate due to the pressure exerted by the crystal growth. As materials precipitate, the vein walls are pushed apart. The resulting veins have shapes typical for mode I cracks except that, mechanically, crystallization pressure replaces the role of internal fluid pressure in their propagation. A nonzero remote differential stress serves to align veins. Mechanical and geochemical considerations suggest that this process will be most important in fine-grained rocks such as greenschist-grade pelites where diffusion from sites of dissolution rather than advection is the dominant mass-transport process. Veins may owe their orientation to tectonism, but their initiation and growth are due to processes that supersaturate the pore fluid. Veins formed by this mechanism involve cracking to the extent that precipitation forces an original flaw to extend during precipitation. Cyclic quartz-mica bands may indicate geochemical self-organization at the vein wall driven by pressure-solution-enhanced supersaturation in the pore fluid and nonlinear precipitation kinetics at the vein wall.