Migmatite formation at subsolidus conditions–an alternative to anatexis

This contribution discusses the formation of stromatic high‐grade migmatites. Volume considerations require that most of the leucosome material is not added from outside the system. A segregation mechanism is necessary except in those cases where the protolith of the migmatite already had a banded structure. Although partial melting is most often advocated to provide the segregation mechanism, several arguments can be raised against high degrees of melting: mineral compositions and even zoning patterns are similar in both mesosomes and leucosomes; sufficient degrees of melting at reasonable temperatures require more than the available amounts of water; the leucosomes do not always approximate to a minimum melt composition; high degrees of melting cannot occur without an appreciable volume increase; etc. Diffusion works as a segregation mechanism at low temperatures. As diffusion rates increase exponentially with temperature diffusion must become still more important as a segregation mechanism at high temperatures. A model is suggested based on the diffusion of components in response to the gradient δσ/δx, where σ= 1/3∑ 3 i =1 σ i is the mean pressure. In homogeneously strained rocks, σ 3 is larger in rock parts rich in incompetent phases than in rock parts depleted in incompetent phases. Accordingly, mechanically competent but chemically incompetent high‐volume phases like quartz and feldspars stressed in micadominated parts of a rock (high σσ) migrate to parts of the rock that are depleted in mica (low σ¯). It is suggested that hornblende occurring in many leucosomes may be premigmatitic or early syn‐migmatitic and due to its mechanical competency it initiates the segregation. Diffusion occurs along grain boundaries and is enhanced by small amounts of ‘intergranular fluid’;. At the best, semiquantitative estimates of diffusion rates and distances indicate that the process should work over geological times.