TECTONITES AND THE PMG TRIANGLE
For the purposes of the proposed classification, the
term 'tectonite' is the most suitable for a metamorphic rock with a mineral fabric developed during deformation. This is in accord with general modern usage which follows Turner & Weiss (1963) who used the term tectonite as a synonym for metamorphic tectonite. Fabrics can be categorized by origin as P-, M- or G-types (as defined here and P, M and G tectonites (each characterized by the one fabric type) can be
nt plotted at the corners of a triangular diagram (M-types). The rest of the triangle can be subdivided into intermediate PM, MG, PG and PMG tectonite fields, according to the the relative Proportions of minerals of contrasling mineral origin in a particular tectonite (Fig. 1).
The essentially monomineralic rocks quartzites, silicates, dunite, provide good examples of P type (Gray 1979) and these rocks often plot as P tectonites.
In P-type fabrics, grain shapes (rib quartz, etc.) may give some indications of the type of the strain, and the LPO will reflect the particular sheet systems operating. Often the SPO and LPO reflec the later stages of deformation, and at moderate grain size,
temperatures, static or dynamic recrystallization obscure the SPO and modify the LPO produced and plastic deformation. If the modifying effects are significant, then the fabric type moves towards recognized
field; this and other effects of recrystallization and, Cataclasis
are discussed further later.
The concept that rotation of inequidimensional mineral grains during a deformation can produce a preferred orientation was clearly enunciated by Sorby (1853). The idea is still widely advocated for sheet silicates in slates (Wood & Oertel 1980) for which a quantitative measure of strain can be made according to the ideas of March (1932). In some slates, selective pressure solution of quartz, particularly along steep limbs of microcrenulations, leads to seams of sheet silicates with a high degree of preferred orientation (Gray 1979). In such slates, mechanical reorientation is partly due to volume loss because of the solution of quartz. Volume loss due to water escape during strain has a similar effect, and Maxwell (1962) argued that sheet silicates become oriented as a direct result of water flow through muddy rocks. Shelley (1975) described grain-size sorting due to water flow in association with cleavage formation in the Ordovician slates of New Zealand, and Jones & Addis (1986) argued that particulate deformation is more widespread than generally recognized. Cataclasis may be an important deformation mechanism in some rocks. In general it can produce M-type
fabrics only if the mineral grains in question are inequidimensional.
In superplasticity, most strain is achieved by grain boundary sliding, though this must be accompanied by plastic deformation or diffusional mass transfer processes to accommodate necessary grain shape changes.
Superplasticity may occur in natural rocks if dynamic recrystallization produces an ultra-fine grain size, and in general the onset of superplasticity is marked by a rapid decline in the strength of a preexisting LPO (Behrmann 1985). However, Gapais & White (1982) found that elongate subgrains of quartz, produced during such a dynamic recrystallization, may be bounded by prism and rhomb planes, and they proposed that rotation during grain-boundary sliding produced M-type fabrics with quartz c-axes close to the extension or shear direction.
G-types Perbaps the most obvious examples of G-type fabrics are those produced during fracturing and precipitation of minerals from solution in the resulting spaces, a process termed 'crack-seal' deformation by Ramsay (1980). The new mineral growths may have an LPO due to growth competition and anisotropies (Cox & Etheridge 1983), as well as a strong SPO with grain lengths parallel to the opening directions. In some cases, the new growth simply enhances an existing preferred
orientation in the host rock.
Many schists are characterized by new mineral growth, often as segregation layers, as well as a general recrystallization, and one might expect G-type fabrics to be typical. This expectation is not reflected in the literature, though in my view growth fabrics are more import ant than is generally recognized. An example from the New Zealand Haast Schists (Shelley 1989) is summarized above. In that example, feldspar, quartz and the sheet silicates are in about equal proportions in the rock, and most fabric elements can be ascribed to diffusion aided anisotropic growth, probably resuIting from a stress-driven combination of solution transfer and crack-seal mechanisms, as advocated, for example, by Sawyer & Robin (1986). Until recently there had been a general lack of information on feldspar preferred orientations in schists, although this situation is now partly rectified. In the greenschist facies (Shelley 1986, 1989), a growth origin of feldspar fabrics is most consistent with the common observation that feldspar is not ductile at low temperatures. However, at higher grades of metamorphism, Olsen & Kohlstedt (1985) and Shaocheng & Mainprice (1988) give evidence for the plastic deformation of feldspar and the production of P-type feldspar fabrics.
Many workers have suggested that a sheet silicate preferred orientation may result from various growth mechanisms (Vernon 1976, Ishii 1988). It maybe generated by growth of new mica at a high angle to earlier
kinked grains, nucleated by the most strongly rotated parts of the original grains (Bell 1978), or, to quote Etheridge ef ai. (1974) "by the interaction of anisotropic growth with either anisotropic fluid movements or rock structure, or with orientation dependent pressure sol
ution". The usual problem is to assess whether crystallization or recrystallization has mimetically adopted an orientation produced by an earlier mechanical alignment. A mineral which may generally develop preferred orientations by growth is hornblende. There is a dearth
of thorough modern descriptions, and the common observation of very elongate euhedral or subhedral hornblendes almost perfectly aligned and closely spaced in linear hornblende schists requires explanation.
Although there is some evidence for plastic deformation of hornblende, especially at high temperatures (Rooney et al. 1975), it is generally thought to be one of the strongest ofminerals (Wenk 1985). Acknowledging this, Nicolas & Poirier (1976) suggest hornblende preferred orientations develop by growth during strain. Since
feldspar is similarly non-ductile at low temperatures, the fabrics of low-grade hornblende-feldspar schists are probably the result of stress-driven solution and growth. Finally, an example of G-type quartz fabrics is given by Gapais & Barbarin (1986) in which quartz grains grew (either by primary grain growth or by'migration recrystallization' with c-axes close to the extension direction at high temperatures during the syntectonic deformation of a granite. There is a transition between this behaviour and the plastic deformation of the quartz at lower temperatures.