The topic of whether or not porphyroblasts rotate relative to a fixed external reference frame during deformation has engendered much debate, mainly owing to the fact that pre-deformational orientations of porphyroblasts and their inclusion trails can rarely be reconstructed. In 20 years of researching this topic, I have encountered only two examples where pre-deformation orientations can be reconstructed, and both examples show rotation of porphyroblasts relative to one another during crenulation cleavage development (Johnson et al., 2006; Johnson, 2009). Sanislav (2009) has taken issue with my conclusions (Johnson, 2009), and I welcome this opportunity to reply. Sanislav's comments focus on the use of numerical modeling, so I will first reply briefly to those remarks before discussing the microstructural evidence.

The rheological behavior of deforming metamorphic rocks involves complex coupling of chemical and mechanical processes. This is a challenging problem at the forefront of numerical modeling. Sanislav (2009) employed Druker-Prager plasticity to achieve his results, but the deformed mesh in his models suggests that he has also employed a strain- or strain-rate-weakening formulation. Given that I cannot reproduce his results, and that there is no description or justification for the modeling methodology, the modeling is difficult to assess. A shortcoming in Sanislav's modeling is that he did not reproduce the geometry in my Figure 1 (Johnson, 2009). I now move on to this key point.

Sanislav claims that the microstructures described by Bell and Bruce (2007) provide certain evidence that the porphyroblasts in these rocks have not rotated relative to one another. Bell and Bruce (2007) argue for pre- and syn-growth microfolding of matrix S2 to explain the large variation (≥79°, figure 2 of Johnson, 2009) of inclusion-trail orientations in these rocks, but the microstructure shown in Figure 1 cannot be explained in this way. Figure 1 shows an image of the two porphyroblasts discussed by me in a slightly offset serial section that has been processed in order to emphasize the S2 foliation in the matrix and porphyroblasts. I argued that these two porphyroblasts were directly adjacent to one another and overgrew parallel S2 foliation surfaces prior to any visible development of S3 (thin rims on the porphyroblasts appear to have grown syn-S3). During the development of S3, the two porphyroblasts were separated by heterogeneous stretching, leaving a quartz-rich dilatational region between them. After retrodeforming the porphyroblasts so that they are adjacent to one another, the S2 foliation that would have been directly above and below both porphyroblasts was, and still is, oriented subvertically in the image reference frame (Fig. 1A), and this is the orientation that was originally overgrown by both porphyroblasts. Retrodeformation of the two porphyroblasts with and without relative rotation leads to the geometries in Figures 1B and 1C, with Figure 1B representing the favored original configuration.

In response to the final sentence of Sanislav (2009), porphyroblasts can, and probably generally do, rotate relative to one another during crenulation cleavage development and other types of continuous deformation, although the total amount of rotation will be difficult to predict for a variety of reasons (reviewed in Johnson, 2009; and Johnson et al., 2006, 2009). The data and observations presented in Johnson (2009) and in this Reply are difficult to interpret in any other way.