high angle to cleavage. The fibre overgrowths are measurements but tends to be averaged out at the scale considered to record extensional strains that of the thin section accumulated during SMt deformation In XY and XZ As discussed in Ring(1996)and Feehan and sections, fibres bundles are typically straight and Brandon(1999), the formation of a SMT fabric unidirectional(Fig 4). The unidirectional geometry requires the accommodation of small motions on the indicates that strains in the y and Z directions are selvage surfaces to account for differential motion of contractional. In contrast, some studies(e. g, Ring and adjacent grains. We see no textural evidence for Brandon, 1999)have recognized multidirectional thoroughgoing slip surfaces or for oblique shearing fibres that point in all directions in the Xr plane between grains The deformation associated with Smt indicating extension in both X and Y processes is accommodated solely by shortening In XZ sections, the fibre bundles generally lie across the selvages and extension in the fibre subparallel to the trace of cleavage(Fig. 4a and b) direction Individual fibre bundles typically have a tapered Textural evidence suggests that the sandstones had geometry, with fibres converging away from the host little porosity at the start of sMT deformation. all of grain. This tapered geometry is recorded in the the space between grains is presently occupied by distribution of fibre directions, which commonly vary selvages or directed fibre overgrowth. Dissolution by as much as +15 around the average direction along selvage surfaces would quickly remove an The taper geometry has been explained as resulting initial porosity. Transient porosity might have existed from dissolution between the fibres to accommodate along the incoherent surfaces that separated the fibre shortening in the Y and Z directions(semi-deformable overgrowths from their host grains. However, the antitaxial fibre model of Ring and Brandon, 1999) porosity along this surface would have been small Extension parallel to X is accommodated solely by given that displacement-controlled fibre overgrowths growth of new fibres. The fibres are inferred to accrete only form when crack apertures are small, on the scale at the grain boundary, so that the amount of shortening of microns or less (Urai et al., 1991; Fisher and across the fibres is largest at the end of the bundles Brantley, 1992). We can think of no other textural This explanation accounts for the observation that the features that might indicate significant porosity during degree of tapering seems to increase with the amount Smt deformation We suggest that mechanical of shortening in the section for instance. fibre compaction had already removed much of the primary bundles appear more tapered in XZ sections than Xr porosity before the onset of smT deformation. This sections because shortening is greater in Z than in Y. result would be expected for a poorly sorted sediment We assume that the fibres track the incremental X where grains of different sizes could be compacted direction during the deformation history of the rock, into a tightly packed aggregate whereas cleavage records the Xr plane for the total Our observations indicate that fibre overgrowth SMT strain. Thus, parallelism between fibres and was the sole mechanism of precipitation during SMT cleavage indicates a coaxial deformation(Feehan and deformation. Alternative possibilities include: (1) Brandon, 1999; Ring and Brandon, 1999). As noted mass transfer associated with metamorphic above, most of our samples( 80%)have fibres that recrystallization of the original detrital grains, (2)the parallel the trace of cleavage. However, some samples precipitation of syntaxial overgrowths on existing (20%)have an average fibre orientation in the xz detrital grains(e.g, the overgrowth of quartz on section that is oblique to the trace of cleavage(e. g detrital quartz grains), and (3)the formation of a fine Fig. 4d), indicating a weakly non-coaxial deformation grained"matrix "around the grains In a few samples, we found individual fibres with an The following evidence indicates that the obliquity of up to 20-30 to cleavage. Nonetheless, the contribution of these other processes was minor.(1) average angle between fibres and cleavage in these The original grain boundaries of the detrital grains are samples is less than 8. We return to this topic below ell preserved and the grains themselves show little when we make specific estimates of the degree of non evidence of internal recrystallization. (2) There is coaxial evidence of alkali mass transfer. such as albitization of In some thin sections, we observed weakly curved plagioclase, but this exchange appears to have fibre bundles around large quartz or feldspar grains occurred by in-situ transfer. For instance, albitized (Fig. 4c and d). This texture appears to record plagioclase grains retain their detrial shapes, which no heterogeneous deformation around the largest grains sign of recrystallization. (3) The fibre overgrowths This localized deformation is included in our appear compositionally uniform within a thin section, which suggests that they grew from a common5 high angle to cleavage. The fibre overgrowths are considered to record extensional strains that accumulated during SMT deformation. In XY and XZ sections, fibres bundles are typically straight and unidirectional (Fig. 4). The unidirectional geometry indicates that strains in the Y and Z directions are contractional. In contrast, some studies (e.g., Ring and Brandon, 1999) have recognized multidirectional fibres that point in all directions in the XY plane, indicating extension in both X and Y. In XZ sections, the fibre bundles generally lie subparallel to the trace of cleavage (Fig. 4a and b). Individual fibre bundles typically have a tapered geometry, with fibres converging away from the host grain. This tapered geometry is recorded in the distribution of fibre directions, which commonly vary by as much as ±15° around the average direction. The taper geometry has been explained as resulting from dissolution between the fibres to accommodate shortening in the Y and Z directions (semi-deformable antitaxial fibre model of Ring and Brandon, 1999). Extension parallel to X is accommodated solely by growth of new fibres. The fibres are inferred to accrete at the grain boundary, so that the amount of shortening across the fibres is largest at the end of the bundles. This explanation accounts for the observation that the degree of tapering seems to increase with the amount of shortening in the section. For instance, fibre bundles appear more tapered in XZ sections than XY sections because shortening is greater in Z than in Y. We assume that the fibres track the incremental X direction during the deformation history of the rock, whereas cleavage records the XY plane for the total SMT strain. Thus, parallelism between fibres and cleavage indicates a coaxial deformation (Feehan and Brandon, 1999; Ring and Brandon, 1999). As noted above, most of our samples (80%) have fibres that parallel the trace of cleavage. However, some samples (20%) have an average fibre orientation in the XZ section that is oblique to the trace of cleavage (e.g. Fig. 4d), indicating a weakly non-coaxial deformation. In a few samples, we found individual fibres with an obliquity of up to 20-30º to cleavage. Nonetheless, the average angle between fibres and cleavage in these samples is less than 8°. We return to this topic below when we make specific estimates of the degree of non￾coaxiality. In some thin sections, we observed weakly curved fibre bundles around large quartz or feldspar grains (Fig. 4c and d). This texture appears to record heterogeneous deformation around the largest grains. This localized deformation is included in our measurements but tends to be averaged out at the scale of the thin section. As discussed in Ring (1996) and Feehan and Brandon (1999), the formation of a SMT fabric requires the accommodation of small motions on the selvage surfaces to account for differential motion of adjacent grains. We see no textural evidence for thoroughgoing slip surfaces or for oblique shearing between grains. The deformation associated with SMT processes is accommodated solely by shortening across the selvages and extension in the fibre direction. Textural evidence suggests that the sandstones had little porosity at the start of SMT deformation. All of the space between grains is presently occupied by selvages or directed fibre overgrowth. Dissolution along selvage surfaces would quickly remove an initial porosity. Transient porosity might have existed along the incoherent surfaces that separated the fibre overgrowths from their host grains. However, the porosity along this surface would have been small given that displacement-controlled fibre overgrowths only form when crack apertures are small, on the scale of microns or less (Urai et al., 1991; Fisher and Brantley, 1992). We can think of no other textural features that might indicate significant porosity during SMT deformation. We suggest that mechanical compaction had already removed much of the primary porosity before the onset of SMT deformation. This result would be expected for a poorly sorted sediment where grains of different sizes could be compacted into a tightly packed aggregate. Our observations indicate that fibre overgrowth was the sole mechanism of precipitation during SMT deformation. Alternative possibilities include: (1) mass transfer associated with metamorphic recrystallization of the original detrital grains, (2) the precipitation of syntaxial overgrowths on existing detrital grains (e.g., the overgrowth of quartz on detrital quartz grains), and (3) the formation of a fine￾grained “matrix” around the grains. The following evidence indicates that the contribution of these other processes was minor. (1) The original grain boundaries of the detrital grains are well preserved and the grains themselves show little evidence of internal recrystallization. (2) There is evidence of alkali mass transfer, such as albitization of plagioclase, but this exchange appears to have occurred by in-situ transfer. For instance, albitized plagioclase grains retain their detrial shapes, which no sign of recrystallization. (3) The fibre overgrowths appear compositionally uniform within a thin section, which suggests that they grew from a common