Detailed electron microprobe analyses of phyllosilicates in crenulated phyllites from south-eastern Vermont show that grain-scale zoning is common, and sympathetic zoning in adjacent minerals is nearly universal. We interpret this to reflect a pressure-solution mechanism for cleavage development, where precipitation from a very small fluid reservoir fractionated that fluid. Multiple analyses along single muscovite, biotite and chlorite grains (30–200 μm in length) show zoning patterns indicating Tschermakitic substitutions in muscovite and both Tschermakitic and di/trioctahedral substitutions in biotite and chlorite. Using cross-cutting relationships and mineral chemistry it is shown that these patterns persist in cleavages produced at metamorphic conditions of chlorite-grade, chlorite-grade overprinted by biotite-grade and biotite-grade. Zoning patterns are comparable in all three settings, requiring a similar cleavage-forming mechanism independent of metamorphic grade. Moreover, the use of ⁴⁰Ar/³⁹Ar geochronology demonstrates this is true regardless of age. Furthermore, samples with chlorite-grade cleavages overprinted by biotite porphyroblasts suggest the closure temperatures for the diffusion of Al, Si, Mg and Fe ions are greater than the temperature of the biotite isograd (>∼400 °C). Parallel and smoothly fanning tie lines produced by coexisting muscovite–chlorite, and muscovite–biotite pairs on compositional diagrams demonstrate effectively instantaneous chemical equilibrium and probably indicate simultaneous crystallization.

These results do not support theories suggesting cleavages form in fluid-dominated systems. If crenulation cleavages formed in systems in which the chemical potentials of all major components are fixed by an external reservoir, then the compositions of individual grains defining these cleavages would be uniform. On the contrary, the fine-scale chemical zoning observed probably reflects a grain-scale process consistent with a pressure-solution mechanism in which the aqueous activities of major components are defined by local dissolution and precipitation. Thus the role of fluids was probably limited to one of catalysing pressure-solution and fluids apparently did not drive cleavage development.