Ferry's (1984) analysis of the biotite isograd in pelitic schists of the Waterville Formation (south-central Maine, U.S.A.) led to the conclusion that metamorphism had been caused by infiltration of 1-2 rock volumes of aqueous fluid. We find that this important observation is extremely sensitive to the assumed temperature of metamorphism and to corresponding uncertainties in the thermodynamic data and solution assumptions. Uncertainties in the experimental calibrations of the geothermometers are on the order of ± 30°C. In addition, an analysis of the effects of minor substituents on calcite-dolomite and garnet-biotite geothermometers indicates that Ferry's (1980) temperatures are about 40°C too low in the garnet zone. Correcting the phase equilibria to higher temperatures by this latter amount lowers fluid:rock ratios by an order of magnitude to about 0·1:1. Since small uncertainties have such great effects it is clearly important to perform careful uncertainty analyses before using Ferry's approach to draw wide-ranging conclusions about fluid-rock interactions. Given a 40°C upward revision of temperature it is not necessary to invoke pervasive infiltration of aqueous fluid as the major agent of metamorphism. With these low fluid:rock ratios the water source could reasonably be a thin sequence of underlying metapelites and/or the granitic intrusives which accompanied the Acadian orogeny.