Growth, behavior, and textural sector zoning of biotite porphyroblasts during regional metamorphism and the implications for interpretation of inclusion trails: Insights from the Pequop Mountains and Wood Hills, Nevada, USA

This paper documents the growth mechanisms and behavior of biotite porphyroblasts during progressive Barrovian metamorphism and focuses on how strain controls the three- dimensional development, preservation, and modification of passive inclusions and their patterns. This study builds on work in Australia, the Pyrenees, Japan, the Scottish Highlands, and Korea by Vernon and Flood (1979), Lister et al. (1986), Miyake (1993), Barker (2002), and Kim and Cho (2008), respectively. These authors collectively showed that the growth of biotite in upper greenschist to lower amphibolite facies is a two-stage process. The first stage involves growth by matrix replacement and incorporation of passive inclusions that represent an excess reactant of the porphyroblast-forming reaction or matrix material not involved in the reaction. The second stage involves a syntectonic process known as crack-fill porphyroblastesis (Barker, 2002). The crack-fill growth mechanism is facilitated by fracturing along {001} in porphyroblasts where {001} is at a high angle to foliation. Growth in these porphyroblasts occurs by precipitation of biotite in {001}-parallel extension fractures within the porphyroblasts, as well as by syntaxial precipitation in episodically or continuously dilating pressure shadows where the matrix separates from the {001} faces (e.g., Lister et al., 1986; Barker, 2002). This type of growth produces zones of relatively inclusion-free, clear biotite at the ends of porphyroblasts and in intragrain extension fractures.

This study focuses on the growth of biotite in metapelite in the Pequop Mountains and Wood Hills, Nevada (Fig. 1), and (1) provides an additional example of crack-fill porphyroblastesis, which has not been widely recognized; and (2) demonstrates that this growth mechanism is probably an important process even in the first stage of growth. Furthermore, this paper shows that the growth of biotite involves a natural progression through a constructive phase in which diverse inclusion trail patterns develop, and a subsequent destructive phase that obscures these patterns. The destructive phase is similar to the second growth stage recognized by others, but the early constructive growth phase is different because it involves the crack-fill growth mechanism and results in textural sector zoning of passive inclusions (e.g., Fig. 2F). This paper also shows how the zoning patterns can be used as an indicator of the style of strain accompanying growth and why caution must be used when making strain assessments of biotite, and other Barrovian index minerals, on the basis of inclusion trails.

Porphyroblasts that contain evenly distributed trails of passive inclusions are commonly designated as syntectonic, intertectonic, or post-tectonic (e.g., Figs. 2A–2C). For example, porphyroblasts that contain trails that curve into foliation are commonly interpreted as syntectonic, and porphyroblasts with straight inclusion trails can be interpreted as intertectonic if foliation wraps around the grain margin, or as post-tectonic if foliation is not deflected around the porphyroblast (Figs. 2A–2C; Passchier and Trouw, 2005). However, some porphyroblasts contain inclusions that are not distributed evenly throughout the grain, but rather occur in a geometric pattern that inherently reflects the crystal's structure (cf. Figs. 2A–2C and Figs. 2D–2F). These porphyroblasts exhibit textural sector zoning and are regarded as pretectonic or intertectonic (Rice and Mitchell, 1991; Rice, 1993). Well-known examples of zoning include the characteristic cross or hourglass patterns observed in some cut orientations of chiastolite and chloritoid, as well as the more complicated patterns in garnet (e.g., Andersen, 1984; Burton, 1986; Rice and Mitchell, 1991; Rice, 1993; Rice et al., 2006; see Figs. 2D, 2E for examples). Textural zoning occurs when there is preferential incorporation of inclusions along particular crystallographic faces (growth sectors) or boundaries between faces during growth (e.g., Harker, 1932; Rast, 1965; Spry, 1969; Rice and Mitchell, 1993; see also Barker, 1998; Vernon, 2004, for overviews). Rice and Mitchell (1991) observed that there is a common co-occurrence of cleavage or graphite domes along the margins of texturally zoned porphyroblasts. Because the domes represent material displaced by the porphyroblast during growth in a hydrostatic state of stress (e.g., Ferguson et al., 1981), Rice and Mitchell (1991) and Rice (2001) suggested that the presence of textural sector zoning is an indicator of growth in a hydrostatic state, and hence the porphyroblasts are either pretectonic or intertectonic. Although textural sector zoning in biotite has not been studied heretofore, this paper will show that it is a syntectonic phenomenon.

The Pequop Mountains, Wood Hills, and adjacent Ruby Mountains–East Humboldt Range, Nevada, expose parts of a metamorphosed thrust footwall that formed in the hinterland of the Mesozoic Sevier fold-and-thrust belt; these metamorphic rocks were subsequently exhumed by normal faulting during Mesozoic and Cenozoic extension (Fig. 1; Camilleri and Chamberlain, 1997). The exhumed thrust footwall constitutes a Barrovian metamorphic terrain that ranges from a sillimanite zone to the west and progressively decreases in metamorphic grade and pressure to the east (Fig. 1). The footwall contains Paleozoic carbonate and clastic strata that underwent metamorphism synchronous with extensional collapse in response to tectonic burial. Footwall collapse was accommodated by dominantly coaxial strain that resulted in development of a regional prograde metamorphic foliation (S₁) and as much as 50% attenuation of stratigraphic units (Camilleri, 1998).

The exposed footwall in the Pequop Mountains consists of a continuous transition of unmetamorphosed to garnet zone rocks that contain the S₁ foliation. A west-northwest–trending mineral lineation is present in many of the metamorphic rocks as well as a macroscopic to megascopic scale pinch-and-swell structure that accommodated as much as ~30% attenuation of stratigraphic units (Camilleri, 1998). Crystallographic preferred orientations of quartz c-axes in quartzite indicate that coaxial flattening accompanied attenuation (Camilleri, 1998). The footwall rocks in the adjacent Wood Hills differ from the Pequop Mountains in that units are more attenuated and they are largely in the kyanite zone. Following peak metamorphism, rocks in both ranges underwent a second deformation that involved the development of the out-of-sequence, small-displacement Independence thrust exposed in the Pequop Mountains as well as associated folds and small-scale thrusts in the hanging wall and footwall of the thrust (Fig. 1; Camilleri and Chamberlain, 1997). In both ranges the S₁ foliation is locally folded, and in a few places, weakly crenulated (S₂) by the subsidiary structures in the hanging wall and foot-wall of the Independence thrust (see Appendix 1 in Supplemental File 1¹ for more information).

This study focuses on biotite porphyroblasts in biotitezone phyllite and schist primarily from Cambrian Dunderberg Shale that is not overprinted by the S₂ crenulation in the Pequop Mountains. In addition, to more fully assess the growth and behavioral characteristics of biotite through the spectrum of metamorphic grades, the same rocks were also observed in the adjacent chlorite zone and in the kyanite zone in the Wood Hills. Figure 1 shows the locations of the samples. These metapelite samples are lithologically diverse and have varying proportions of graphite, quartz, and carbonate minerals. In addition, the metapelites are structurally diverse and range from rocks that contain little tectonic fabric to S to S > L tectonites with biotite porphyroblasts defining the lineation (Figs. 3C–3D).

Characterization of the orientation, size, and shapes of biotite porphyroblasts in three dimensions is important to understanding growth processes and textural zoning patterns. This was accomplished by examination of thin sections cut (1) parallel to foliation, and (2) perpendicular to foliation in both lineation-parallel and lineation-normal sections (all photomicrographs in this paper are shown in plane-polarized light). These observations indicate that most biotite porphyroblasts in the Pequop Mountains are subhedral with undefined {010} and {110} faces and well to moderately defined {001} faces (Fig. 3A). Biotite composes a few percent to as much as ~35% of the volume of the rock and crystals are as large as ~1.5 mm in their longest dimension. Biotite is present as both porphyroblastic and foliation-defining matrix crystals with matrix grains typically having {001} at a low angle or parallel to foliation. The porphyroblastic crystals have a preferred orientation in the sense that their crystallographic c-axes are most commonly oriented at a low angle, or parallel, to foliation; i.e., the {001} cleavage tends to be at a high angle to foliation. Porphyroblasts with {001} parallel or subparallel to foliation are generally sparse, but are more abundant in rocks where biotite composes a large part of the volume of the rock.

An ideal biotite crystal is monoclinic, thin in the direction of the crystallographic c axis, and wider in the directions of the a- and b-axes (Fig. 3A); however, biotite porphyroblasts in the Pequop Mountains differ from ideal in many respects. Their relative crystallographic dimensions appear to be related to strain because crystals are generally elongated parallel to lineation regardless of crystallographic orientation (Fig. 3). This has an effect on the dimensions of biotite in that porphyroblasts having c-axes that are approximately parallel to foliation and lineation tend to be wider parallel to the c axis than those with c-axes subparallel to foliation but normal to lineation (e.g., porphyroblasts M and O, respectively, in Figs. 3A–3B). Furthermore, crystals with c-axes that are normal to foliation appear to be thinnest parallel to the c axis (e.g., porphyroblast N in Figs. 3A, 3B). Porphyroblasts in rock that has no appreciable lineation have more equant shapes (cf. Figs. 3C, 3D).

Although the c-axes of porphyroblasts are dominantly at a low angle to foliation, they have a diversity of orientations relative to lineation, if present. Consequently, in any foliation-normal thin section, porphyroblast cross sections have a diversity of shapes. These include elliptical, circular, rectangular, parallelogrammatic, and parabolic shapes (Figs. 4–8).

To understand the predominance of biotite porphyroblasts with {001} at a high angle to foliation, which is unusual for phyllosilicates, possible porphyroblast nucleation sites were observed in phyllite in the proximity of the biotitein isograd (Fig. 1). Chlorite-zone phyllite near this isograd is characterized by foliation defined by white mica, chlorite, deformed detrital quartz, and carbonate (ankerite?) grains, and sparse extension fractures (veins) that are at a low angle or parallel to foliation and are filled with chlorite, quartz, and plagioclase [see Fig. 9A and inset photomicrograph in Fig. 9B; see also additional discussion about the origin of these veins in Appendix 2 (see footnote 1)]. The same phyllitic unit just across the isograd in the biotite zone contains small biotite grains that may reflect the initial stage of growth of biotite porphyroblasts (Figs. 9C–9F). Figures 9C–9F show an example of a biotite-bearing phyllite with unfoliated laminations (bedding) of quartz, carbonate, and mica that are separated by foliated layers defined by flattened detrital quartz grains, and chlorite and white mica with {001} at a low angle to parallel to foliation (Fig. 9C). Both foliated and unfoliated parts in this sample have biotite, but the biotite tends to be larger in the undeformed layers. Biotite in the undeformed layers has no apparent preferred orientation, whereas biotite in the foliated parts tends to be oriented with {001} either at a low angle or parallel to foliation, or at any angle when in the strain shadow of a quartz grain (Figs. 9 D–9F). Consequently, biotite with {001} at a high angle to foliation may have nucleated in no-strain to low-strain areas. Additional but sparse sites of nucleation include strain shadows on diagenetic(?) pyrite and in veins as a replacement of chlorite (Fig. 9B).

The constructive phase involves growth processes that are responsible for the bulk of growth of, and development of inclusion patterns within, the biotite porphyroblasts. In contrast, the destructive phase involves destruction or modification of constructive phase inclusion patterns by fracturing, rotation, and residual growth around fractured and rotated porphyroblasts. The constructive phase of a biotite porphyroblast is typically demarcated by passive inclusions that define a texturally sector-zoned core that in some cases is surrounded by a poikiloblastic or dendritic overgrowth. The inclusion trails are generally straight to gently curved, and the inclusions are mostly quartz and graphite, although in some rocks a few large porphyroblasts have overgrown smaller biotite grains as whole or partial inclusions.

Textural sector zoning appears only in porphyroblasts whose inclusion trails are at a high angle to {001}, and hence it developed in crystals that grew with {001} oriented at a high angle to foliation. Figure 4 shows a schematic illustration of the three-dimensional geometry of, and two-dimensional cuts through, an ideal textural sector-zoned porphyroblast with {001} normal to foliation, and the photomicrographs in Figures 5–8 show examples of these cuts. Many of the photomicrographs of zoned porphyroblasts are accompanied by line drawings where the included area is always indicated by a number 1, the areas that lack significant inclusions (i.e., clear zones) by 2, and post-zoning dendritic or poikiloblastic overgrowth by 3. In addition, photomicrographs of texturally zoned porphyroblasts from samples not shown in this paper are presented in Supplemental File 1 (see footnote 1), which includes those that exhibit more subtle manifestations of zoning and those with zoning patterns that vary from ideal.

The two-dimensional zoning patterns viewed in any thin section are diverse and reflect cut effects (Fig. 4). Porphyroblasts that are cut parallel or subparallel to the c axis contain a central hourglass-shaped distribution of passive inclusions (B₁ and B₂ sections in Figs. 4, 6, and 7). The included hourglass shape is bounded by zones of clear biotite that contain a low density of, or no, inclusions. In three dimensions, the clear zones are broadly cone shaped, concave outward, and terminate at {001} crystal faces (Fig. 4). The axes of cones are roughly parallel to the c axis. The shape of the hourglass, however, varies depending on proximity of the c axis to the section cut. For example, a section cut nearly parallel to, and through, the c axis yields an hourglass shape with a thin or no neck (B₁ sections in Figs. 4C, 6B, and 7C), whereas a section cut closer to the margin of the porphyroblast will yield the same pattern but have a much thicker neck and correspondingly smaller clear zones (B₂ sections in Figs. 4C and 7A, 7B, and 7D).

Porphyroblasts that are cut highly oblique to the c axis differ from the B₁ and B₂ sections in that they are either broadly elliptical or parabolic in shape and have different inclusion patterns. Elliptical sections that do not cut the {001} faces display an off-centered, internal clear zone if cut approximately through one of the conical clear zones, or have no apparent clear zones if the cut is through the center of the hourglass shape (C₁ and C₂ sections in Figs. 4C, 8A, and 8B). A parabolic section cuts one of the {001} faces and contains one arcuate clear zone that terminates at the (001) face (D section in Figs. 4C and 8C). The appearance of sections cut approximately perpendicular to the c axis varies depending on where the section is cut. If the cut is through one of the clear zones, then inclusions are distributed in an elliptical pattern along the grain boundary with a central clear zone (A₂ section in Figs. 4B, 5B, and 5C). If the cut is through the center, then the porphyroblast will appear replete with inclusions (A₃ section in Figs. 4B and 5D). In contrast, if the section is cut near the end of the porphyroblast, then it should be mostly clear with sparse inclusions along the margin of the grain (A₁ section in Figs. 4B and 5A).

The origin of the conical clear zones in biotite porphyroblasts is probably, in part, similar to the fluid-induced crack-seal growth mechanism described by Lister et al. (1986) for biotite porphyroblasts in the Pyrenees. They described porphyroblasts with {001} at a high angle to foliation that initially grew over foliation incorporating inclusions followed by growth in a direction parallel to foliation in dilating strain shadows. Growth in the strain shadows occurs when porefluid pressure exceeds the tensile strength of the matrix-{001} interface, and biotite is precipitated in the void created by separation of matrix from the porphyroblast {001} faces. This process produces an included core bounded by paired zones of clear biotite with little or no inclusions at the ends of the porphyroblast (e.g., Figs. 10A–10D). Lister et al. (1986) also indicated that overall growth rate kept pace with matrix separation, but that occasionally, in places, either strain decreased or growth rate increased, resulting in matrix overgrowth yielding sparse inclusions in the clear zones, which is similar to that observed in the Pequop Mountains.

The textural sector-zoned porphyroblasts from the Pequop Mountains differ from those in the Pyrenees primarily in the shapes of the included versus clear areas, and more specifically with regard to the hourglass-shaped included core in the Pequop Mountains (cf. Figs. 10E and 10L). The typical inclusion patterns observed in the Pequop Mountains can be produced if growth (1) occurs in all directions away from the core, rather than being inhibited in directions perpendicular to foliation, and (2) is accompanied by protracted shortening approximately perpendicular to foliation and extension parallel to foliation. Figures 10F–10K show the progressive two-dimensional development of an hourglass-shaped distribution of straight inclusion trails due to growth in all directions. Growth occurs on the {001} faces where matrix separates from the porphyroblast in symmetric dilating strain shadows at the same time as growth in a direction perpendicular to foliation along the porphyroblast margins that parallel foliation. Only growth in directions perpendicular to foliation involves matrix replacement and hence incorporation of inclusions, which ultimately yields the hourglass-shaped included core. In three dimensions this process would produce paired clear zones that have a broadly conical geometry (Fig. 10L). Restricted growth of biotite as shown in Figures 10A–10D could reflect a relatively high strain rate inhibiting growth perpendicular to foliation, or perhaps the redistribution of biotite by solution on grain margins parallel to foliation and reprecipitation in dilating strain shadows.

The textural sector zoning patterns in the Pequop Mountains probably developed on small unzoned cores that represent grains that initially grew in strain-free or low-strain areas where growth of crystals with {001} at a high angle to foliation was favorable. Once these biotite crystals grow larger than the surrounding grains, the separation of matrix from {001} faces becomes permissible and textural sector growth can begin. For example, a biotite crystal with {001} perpendicular to foliation may nucleate in the strain shadow of a detrital quartz grain, but then grow out of the shadow and into the deforming matrix, at which point the biotite will generate its own strain shadows and textural sector growth will ensue. This scenario requires the growth of biotite out of a strain shadow in a direction perpendicular to foliation and parallel to {001}, which is the natural fast-growth direction of biotite, and hence it probably facilitates this process. Furthermore, the alignment of the biotite fast-growth direction perpendicular to foliation, coupled with growth parallel to foliation facilitated by matrix separation along {001} faces, allows the zoned crystals to become large, i.e., porphyroblastic, unlike the smaller unzoned biotite crystals with {001} parallel to foliation that have a slow growth direction perpendicular to foliation.

Although the growth model in Figure 10 shows development of a porphyroblast with an hourglass-shaped included core under conditions where growth rate is the same in directions parallel and perpendicular to foliation, many porphyroblasts show evidence that growth rate varied in different directions, resulting in shapes that deviate from that shown in the model. The spectrum of crystal and hourglass shapes in B₁ and B₂ sections that can result from different relative growth rates in directions perpendicular and parallel to foliation are shown in Figures 11A–11D. In the Pequop Mountains, the end-member crystal and hourglass shapes in Figures 11A and 11C tend to form when the c axis is aligned parallel to lineation, whereas the shapes shown in Figure 11B form when the c axis is aligned normal to this direction (cf. the M and O porphyroblasts in Fig. 3B with the shapes of B₂ sections in Figs. 11A and 11B and the porphyroblasts in Figs. 7C and 7D with the shapes of B_I sections in Figs. 11C and 11B, respectively). These observations imply that growth rate in the direction of the c axis is (1) generally enhanced when it is aligned parallel to the direction of maximum extension (lineation), resulting in development of a crystal that is generally long parallel to the c axis, and (2) slower when the c axis is aligned parallel to the minimum extension direction, resulting in crystals that are shorter parallel to the c axis. The dependence of crystal and hourglass shapes on the orientation of the c axis relative to lineation suggests at least a partial control of strain on growth rates. In addition, in some porphyroblasts the extremities of clear zone–hourglass boundaries are arcuate (cf. Fig. 11D and Figs. 12A–12B) rather than straight, as depicted in Figures 11A–11C, which indicates a general slowing of growth rate in a direction perpendicular to foliation during the latter stages of growth (additional data on variations of hourglass shapes are in Appendix 3 in Supplemental File 1; see footnote 1).

Development of the ideal zoning patterns shown in Figures 10K and 11A–11C does not involve rotation of the porphyroblast and results in production of straight inclusion trails. Although this is apparent in the zoned parts of many porphyroblasts (e.g., Figs. 7C–7D), it is also common to find a diversity of curved inclusion trail patterns, which yields information on the behavior of biotite and the matrix during growth. Some curved inclusion trails are gently convergent toward the center of the porphyroblast and some are parallel to foliation on the margin, but have been rotated in the center. Furthermore, some curved inclusion trails appear to represent growth over foliation deflected by a neighboring porphyroblast.

Porphyroblasts with straight to convergent inclusion trails exhibit no apparent rotation during textural sector growth (Figs. 11E–11G). Straight trails are typical of porphyroblasts with quartz inclusions, whereas broadly convergent trails are more characteristic of those with graphite trails (e.g., Figs. 7C and 7D, and 12, respectively). This may reflect different matrix rheologies with rocks that contain a significant graphite component having greater ductility (and possibly strain rate), resulting in convergence of foliation in strain shadows during textural sector growth. In addition, some porphyroblasts exhibit straight to weakly convergent trails in the center with more strongly convergent trails and graphite concentration on the margins, which may reflect an overall decrease in growth rate relative to strain rate (Fig. 11G).

In thin sections that have porphyroblasts with straight to convergent trails, it is not uncommon to find some with rotated cores that reflect dextral and sinistral senses of rotation during the latter stages of textural sector growth (e.g., Figs. 13A–13D,^13021303). This is apparent in both lineation-parallel and lineation-normal thin sections. Inclusion trails in the rotated core generally form an acute angle to {001}, with the sense of rotation commonly predictable by observing the restored (unrotated) position of {001} relative to foliation, which was probably the approximate long dimension of the crystal prior to rotation. For example, when the cores are restored to their pre-rotation positions, those with a dextral sense of rotation had {001} canted to the right and those with a sinistral sense were canted to the left (Figs. 13C, 13D). The observed opposing senses of rotation are consistent with strain that involves a significant component of shortening perpendicular to foliation. Such opposing senses of rotation can develop when the long axes of some porphyroblasts are inclined in opposite directions and are not parallel to the instantaneous stretching axes, one of which is presumably parallel or at a low angle to foliation (e.g., Ghosh and Ramberg, 1976). It is also plausible that porphyroblasts may rotate, regardless of orientation, due to strain interference with neighboring porphyroblasts. Furthermore, rotation also has the effect of inducing cessation of textural sector growth as the {001} faces rotate toward foliation and away from the optimum angle for tensile separation of the matrix from the {001} faces. When this happens, growth continues in all directions with incorporation of inclusions (e.g., see rotated B₂ section with clear zones surrounded by included biotite in Fig. 13D). Some curved inclusion trails may be unrelated to rotation and are probably a result of growth of a porphyroblast over foliation that has been deflected around a neighboring porphyroblast (e.g., Figs. 13E, 13F). All of the foregoing types of inclusion trails serve to illustrate that trail shapes reflect matrix rheology, flow perturbations, and growth over matrix deflected by a neighboring porphyroblast.

In comparison to other texturally zoned mineral species that require a hydrostatic state of stress and no distortional strain to produce textural sector zoning (e.g., Rice and Mitchell, 1991; Rice et al., 2006), biotite is unique in that it requires strain to develop. Therefore, zoning in biotite indicates syntectonic growth, which contrasts with the pretectonic or intertectonic origin of other zoned minerals. The presence of zoning can also be used to imply the style of strain during growth. As shown in the growth model in Figure 10 (F–L), the typical zoning patterns observed in biotite can form if growth is accompanied by progressive shortening perpendicular to foliation and extension parallel to foliation, which implies that each increment of strain accumulates coaxially. The model therefore illustrates the development of zoning under progressive end-member coaxial strain, i.e., no rotation of finite stretching axes. However, textural sector growth is probably permissible with a small component of noncoaxial strain, but prolonged textural sector growth and full development of the patterns in Figures 10 and 11 are probably not favored in a regime of general shear with a large component of noncoaxial strain. This is because a strong noncoaxial component would likely result in biotite crystals with {001} at a high angle to foliation being rotated toward foliation and hence away from the optimum angle for tensile separation of the matrix from the {001} faces, resulting in cessation of precipitation of clear biotite on these faces. Consequently, prolonged textural sector growth is likely only operative in regimes where the coaxial component of strain is significant and persistent. Therefore, textural sector zoning patterns in biotite (e.g., Figs. 10K and 11) are inferred to require a strong coaxial component of strain during growth, and hence the presence of such patterns may be used as an indicator of this type of strain. Furthermore, the presence of zoning in porphyroblasts that have c-axes nearly parallel to foliation but normal to lineation signifies extension perpendicular to lineation, and can be used to suggest that flattening strain accompanied growth. In the Pequop Mountains, porphyroblasts of this orientation are zoned, indicating flattening strain. This is consistent with patterns of crystallographic preferred orientation of quartz c-axes in quartzite in the adjacent garnet zone that are indicative of coaxial flattening (Camilleri, 1998).

Porphyroblasts without textural sector zoning occur in two settings: (1) in rocks where porphyroblasts with {001} at a low angle or parallel to foliation are unzoned and porphyroblasts with {001} at a high angle to foliation are zoned, and (2) in rocks where porphyroblasts appear to lack zoning regardless of crystallographic orientation. In the first setting, the apparent lack of zoning in porphyroblasts with {001} parallel to foliation may be a function of alignment of the biotite fast growth direction with the extension direction, thereby enhancing the probability of growth rate outpacing retreat of the matrix in strain shadows. This would generally preclude development of a void in which to precipitate clear biotite along the foliation-normal, non-{001} faces, and hence growth would proceed by matrix replacement. In samples where all biotite appears unzoned, porphyroblasts either do not have ample inclusions to indicate zoning, or have inclusions but no overt evidence of zoning (e.g., Figs. 14A, 14B). There are several possible reasons for the lack of zoning. First, the rock may not be of appropriate composition to develop inclusions, e.g., insufficient graphite or quartz. Second, the relative rates of strain versus growth may not be appropriate for textural sector growth, i.e., there may be low or negligible strain resulting in apparent growth under static conditions (e.g., Figs. 14A, 14B). On a local to regional scale, the absence of textural sector zoning, although sparse, in part reflects the partitioning of strain. Figures 14C–14E show an example of this, where schist with a graphite-poor layer containing crudely zoned porphyroblasts is overlain by a graphite-rich layer with unzoned, highly included porphyroblasts that did not undergo the textural sector growth process.

For many rocks the last vestige of biotite's constructive phase is marked by cessation of textural sector growth and overgrowth of matrix in all directions. The overgrowth appears to reflect a strain regime similar to that that produced the zoning, but with growth rate exceeding strain rate. In B section porphyroblasts the overgrowth appears as a poikiloblastic, or less commonly, dendritic rim framing the included hourglass shape and clear zones (Figs. 6A, 15A, and 15C). However, thin sections that contain porphyroblasts with overgrowth rims also generally contain a few B sections that are entirely dendritic and lack any evidence of textural sector zoning (e.g., Fig. 15B). These porphyroblasts probably represent edge cuts that only penetrate the dendritic rim (cf. Figs. 15A, 15B). The degree of overgrowth is usually consistent on a thin-section scale, but can vary from negligible to extreme in the same outcrop area, and therefore the driving forces for overgrowth are probably localized (Figs. 7A and 12C show extreme, and Figs. 12A and 12B show negligible, examples of overgrowth from the same area).

The overgrowth rim could be interpreted as indicating a change to either growth rate exceeding strain rate, hence precluding matrix separation at the {001} faces, or simply growth under static conditions. However, observation of inclusion trails in the overgrowth rims suggests that some strain accompanied overgrowth. This is suggested by the persistence of the convergence of inclusion trails from the zoned parts of the crystals through the overgrowth rim coupled with a geometrically similar convergence of foliation surrounding the porphyroblasts (e.g., Figs. 7A and 12C). This implies that the general strain regime that produced textural sector growth continued through overgrowth.

The inception of the destructive phase is generally marked by (1) the cessation of porphyroblastic growth in all directions and (2) the beginning of intragrain strain coupled with restricted residual growth of biotite in strain shadows and extension fractures. These processes occurred as the growth of biotite diminished and the fill of dilating strain shadows transitioned from precipitation of biotite during textural sector growth to predominantly quartz. Overall, the destructive phase appears to have taken place during a broad continuation of the strain regime that characterized textural sector growth, but it involved straining more than growing of the porphyroblasts. The destructive phase involved fracturing, rotation, dissolution, and minor kinking and subgrain development on the corners of grains. These processes collectively obscured or modified the textural sector zoning and inclusion trail patterns developed during the constructive phase.

Following development of poikiloblastic or dendritic rims, fracturing in biotite appears to mark a fundamental transition from growth to strain of porphyroblasts and a shift from strain focused in the matrix that was accommodating textural sector growth (i.e., as in Fig. 10) to transference of some strain to the porphyroblast. Fractures are mostly present in porphyroblasts with {001} at a high angle to foliation. The fractures do not penetrate the matrix, suggesting a rheologic contrast between a ductile matrix and brittle porphyroblast at the time of fracture. Fractures can be divided into four categories based on apparent slip or separation viewed in lineation-parallel, foliation-normal thin sections. In order of decreasing abundance, they are (1) extension fractures along {001}, (2) shear fractures along {001}, (3) extension fractures oblique to {001}, and (4) shear fractures oblique to {001} (Figs. 16,¹⁶⁰² and 17).

Extension fractures that developed parallel to {001} constitute nearly all of the fractures in biotite porphyroblasts and they tend to have a predictable spacing and geometry. To understand the distribution of these fractures, it is important to view B sections so that most or all of the porphyroblast can be seen. From these observations it appears that porphyroblasts typically contain one to three fractures, which tend to divide porphyroblasts into segments of approximately equal width, with one commonly being in the center (Figs. 16A–16D). Precipitation of biotite, quartz, and sparse opaque and epidote minerals by intergranular fluids followed fracturing (Figs. 16 and 17). Growth of biotite in fractures was either syntaxial or nonsyntaxial. Syntaxial growth is easily recognized by uniform extinction of the porphyroblast and the fill. In fractures that are mostly filled with syntaxial biotite, quartz appears as elongated blebs parallel to {001} and can resemble inclusions (Figs. 7C and 16C; additional data are in Appendix 4; see footnote 1). Nonsyntaxial growth is more common and is indicated by various sheaths of biotite that are generally only slightly misoriented with respect to the host. Nonsyntaxial growth may have resulted from crystal defects altering growth patterns or syntaxial growth on fragments of the host that sustained minor rotation during fracturing (e.g., Fig. 17B). Some porphyroblasts appear to have a combination of syntaxial and nonsyntaxial growth in fractures, but in one extension fracture observed, half of the fracture was filled by syntaxial growth on the host and the other half by growth of a neighboring porphyroblast into the void (Fig. 17A). Post-zoning growth of biotite in extension fractures results in the production of a clear zone that can be distinguished from those produced during textural sector growth, specifically because they are rectangular rather than conical in shape and they transect zoning patterns (e.g., Fig. 17).

Several observations suggest that the extension fractures developed by the fiber-loading or stress-transfer mechanism (Lloyd et al., 1982) as a consequence of stresses imparted to the rigid porphyroblasts by the ductiley flowing matrix. It has been shown that the development of extension fractures in fibers in composite materials occurs when stress is transferred from the flowing matrix to the fiber. In these cases tensile stress is greatest at the mid-point of the fiber (or mineral), and when tensile strength is exceeded, a fracture develops in the center (e.g., Boullier, 1980; Watts and Williams, 1980; White et al., 1980; Lloyd et al., 1982; Ji and Zhao, 1993). Fractured segments will then continue to fracture at their midpoints until the length of a segment is below some critical length where stress cannot exceed the tensile strength of the material (e.g., Lloyd et al., 1982). The stress-transfer mechanism appears to be a plausible explanation of the consistency of the geometry and number of fractures observed in many biotite porphyroblasts. This is supported by the observations that the fractures formed at a time when there was a rheologic contrast between the matrix and porphyroblasts, and that most fractures are equally spaced with at least one at or near the center, albeit the order of fracturing can't be ascertained (additional information about porphyroblasts with off-centered fractures is in Appendix 4 in Supplemental File 1; see footnote 1). Moreover, stress transfer from the matrix and resulting fracture was probably triggered by attainment of a critical size coupled with cessation of growth, which may facilitate stress transfer when the porphyroblast ceases to replace matrix (i.e., grow) in a direction perpendicular to foliation.

Porphyroblasts with shear fractures are sparse but are present in some thin sections with abundant extension fractures. These fractures cannot be explained by the stress-transfer mechanism; however, most shear fractures are parallel to {001}, occur in porphyroblasts with rotated inclusion trails (Fig. 16E), and may have formed when {001} was rotated toward foliation to an optimum angle for shear fracture according to distribution of principle stresses for the whole rock (i.e., they formed when {001} was rotated to a moderate angle from σ₁). Alternatively, some may be extension fractures with negligible separation that were rotated and reactivated as shear fractures.

In any given foliation-normal thin section it is evident that some whole and fractured segments of porphyroblasts rotated toward foliation following textural sector growth (Figs. 18,¹⁸⁰² and 19). This rotation, coupled in some cases with minor growth (Fig. 18F), yielded B sections that have inclined hourglass shapes with inclusion trails that have a diversity of apparent dip angles and dip directions (Fig. 19). Rotation of porphyroblasts in opposing directions is generally evident in both thin sections cut parallel and perpendicular to lineation (e.g., Figs. 16A–16B, 18, and 19). Direction of rotation for most porphyroblasts appears to be controlled by the orientation of {001} prior to rotation whereby those with a dextral sense of rotation had {001} canted to the right and those with a sinistral sense were canted to the left (e.g., Figs. 16A, 16B, 18F, and 18G). Such multidirectional rotation is likely a product of strain that is dominantly coaxial. However, in some samples the majority of porphyroblasts indicate a top-toward-the-west sense of shear or rotation. These samples probably record a small component of noncoaxial strain significant enough to produce a preferred sense of rotation (a discussion of the regional tectonic implications of such a noncoaxial component is in Appendix 5 in Supplemental File 1; see footnote 1).

The observed multidirectional rotation of biotite porphyroblasts toward foliation in the Pequop Mountains is similar to that reported by Miyake (1993) in Japan. Miyake (1993) described unzoned biotite porphyroblasts with straight inclusion trails exhibiting multidirectional senses and amounts of rotation, with some exhibiting no apparent rotation. Rotation sense and rate were interpreted by Miyake (1993) to be controlled by their shape and the orientation of {001}. Variable rotation was attributed to strain with a pure shear (coaxial) flattening component perpendicular to foliation and a minor simple shear (noncoaxial) component subparallel to foliation.

Dissolution of biotite porphyroblasts and sparse development of quartz strain shadows accompanied and outlasted fracturing and rotation. Dissolution of biotite and quartz in the matrix along porphyroblast margins parallel to foliation resulted in mild to moderate development of strain caps and hourglass shapes with missing components (e.g., Fig. 18C), and it facilitated development of quartz strain shadows accompanied by minor syntaxial growth of biotite (e.g., Fig. 7A). The quartz-filled strain shadows have a diversity of shapes (e.g., Figs. 13C, 14D, 14E, 18A–18D, 19A–19D, and 20) and can range from mildly asymmetric to symmetric in the same thin section, but the most unusual and informative types of strain shadows are those that have an unusual bicuspate geometry that is present on one or both {001} faces (Fig. 20). The formation of bicuspate strain shadows appears to reflect coaxial strain resulting in divergent separation of the matrix away from the corners of the porphyroblast (Fig. 20). Many bicuspate shadows appear to have initiated following fracturing of cusped corners of porphyroblasts or a {001} boundary between a thin overgrowth rim and the zoned part of the crystal. Development of the shadows ensued when the fractured fragments separated and rotated away from the host (Fig. 20). Bicuspate strain shadows are generally only present on porphyroblasts whose inclusion trails exhibit little or no apparent rotation and are absent on rotated porphyroblasts in the same thin section, attesting to the heterogeneity of strain on the thin-section scale.

To assess whether processes operative in the biotite zone and their resultant microstructure persisted to higher grades of metamorphism, correlative kyanite zone schist from the adjacent Wood Hills (Fig. 1) was examined. Many of the schists in the Wood Hills contain biotite as the sole porphyroblast species, but some also contain staurolite, kyanite, and garnet (Fig. 21,²¹⁰²). In comparison to the Pequop Mountains, biotite porphyroblasts in the Wood Hills are, in places, slightly larger (as much ~2 mm in the longest dimensions), matrix grains are coarser, chlorite is absent as a prograde phase, and they lack characteristic low-grade pressure solution microstructure such as strain caps. Despite higher metamorphic grade and growth of other porphyroblast species, some relict textural sector zoning patterns and poikiloblastic rims of the constructive phase as well as relicts of extension fractures and evidence of rotation from the destructive phase are present in the Wood Hills. Figures 21A–21E show examples of relicts of these features present in a schist sample that contains sparse staurolite and kyanite. This sample is similar to the Pequop Mountains sample shown in Figure 19 in that inclusion trails are variable in apparent dip direction and angle. Samples with abundant kyanite, staurolite, and garnet (Fig. 21G) lack strong relicts of zoning probably because reactions associated with these minerals involved consumption of old zoned biotite and growth of new biotite. For example, biotite porphyroblasts in these rocks tend to have an included center surrounded by a clear rim that obscures what may be relict zoning patterns, which contrasts with overtly zoned relicts with poikiloblastic rims in samples that lack or have minor amounts of kyanite, staurolite, and garnet (cf. Figs. 21A–21E with 21H). The growth of the clear rim may be a product of the kyaniteproducing reaction in this rock (e.g., quartz + muscovite + staurolite = biotite + garnet + kyanite; Camilleri and Chamberlain, 1997). Consequently, if searching for evidence of relict textural sector zoning in high-grade rocks, it is best to look in schist that has biotite as the sole porphyroblast species.

In summary, because biotite porphyroblasts in the Wood Hills contain relicts of the same features as those in the Pequop Mountains, and overall are not much larger, it is reasonable to assume that most nearly reached their peak size, and underwent modification in the destructive phase, in the biotite zone. Therefore, it appears that textural sector growth was only operative in the biotite zone.

The growth processes observed in the Pequop Mountains may characterize low-grade Barrovian metamorphism of pelite in parts of orogenic belts undergoing burial, heating, and collapse. In essence, in the biotite zone, biotite should undergo a natural progression through the constructive and destructive phases as depicted in Figure 22. The constructive phase involves strain that is accommodated by both the textural sector growth mechanism and the shear and reactions in the matrix that facilitate it. This phase ceases as porphyroblasts mature in size and growth rate diminishes as nutrients become depleted, and may be marked in some areas by dendritic overgrowth as crystals extend growth prongs to reach nutrients. The transition to the destructive phase is triggered when porphyroblasts cease to replace matrix perpendicular to foliation, coupled with markedly diminished accommodation of strain by textural sector growth. When this happens some strain is transferred to the porphyroblasts, resulting in rotation and tensile fracturing (via the stress transfer mechanism).

Progression through the constructive and destructive phases inevitably results in porphyroblasts with a diversity of inclusion trail patterns that record variable strain and growth histories at all scales. The variability observed on a small scale in a thin section is a function of where the porphyroblast is cut relative to its growth center (e.g., Fig. 15) and the heterogeneity and partitioning of strain. Partitioning and heterogeneity results in: (1) development of opposing senses of shear due to rotation of porphyroblasts during textural sector growth (e.g., Figs. 13A–13D), (2) perturbations of matrix around growing and impinging porphyroblasts during textural sector growth that leads to overgrowth of the perturbed matrix (e.g., Figs. 13E, 13F), and (3) production of variable senses and amounts of rotation of fractured and unfractured porphyroblasts during the destructive phase (e.g., Figs. 18, 16A, and 16B). Variability in strain history on a mountain-range scale is probably also related to strain partitioning and is reflected in the development of textural sector zoning in biotite in most rocks, but not all, during the constructive phase and in the variability in the style and intensity of the destructive phase (e.g., compare the lack versus abundance of destructive phase microstructure in Figs. 15 and 21F, respectively).

Understanding the origin of inclusion trails in Barrovian index minerals (e.g., biotite, garnet, and staurolite) is important because the trails are used to assess strain during progressive metamorphism and to aid geothermobarometric studies of these minerals (e.g., Johnson, 1999). Because progression through the biotite constructive and destructive phases results in production of inclusion trails with a heterogeneous array of apparent dip angles, dip directions, and trail patterns (Fig. 22), caution must be used when inferring strain histories from inclusion trails in biotite. Similarly, although it is common practice to designate porphyroblasts as intertectonic, syntectonic, or post-tectonic on the basis of the nature of inclusion trails relative to the external foliation (e.g., Fig. 2A; Passchier and Trouw, 2005, p. 196), such designations should be made circumspectly because of the diversity of inclusion patterns. For example, the inclusion trail pattern developed during the constructive phase in the porphyroblasts shown in Figures 5D and 8A could be interpreted as inter-tectonic, that in Figure 12C syntectonic, and that in Figure 15B post-tectonic. These porphyroblasts are probably all syntectonic, but the microstructure reflects variations in rheology, strain and growth rate, and cut effects. Furthermore, growing evidence suggests that some porphyroblast species that chemically replace others may inherit their inclusion trails (e.g., Rubenach and Bell, 1988; Barker, 2002; Passchier and Trouw, 2005; Kim and Cho, 2008), which could lead to erroneous interpretations, especially if replacement of biotite is involved. This may be an issue for garnet because the garnet-forming reaction may involve replacement of biotite, and hence garnet may inherit inclusion trails and other microstructures that represent the sum of the influence of the constructive and destructive phases of biotite growth in the biotite zone rather than conditions in the garnet zone. Garnet that inherited biotite inclusion trails was documented in the Scottish Highlands by Barker (2002) and in the Imjingang belt in Korea by Kim and Cho (2008). In summary, the foregoing observations indicate that caution should be used when interpreting inclusion trails in biotite in general and specifically in higher grade index porphyroblasts that may have replaced other porphyroblast species.

The growth of biotite porphyroblasts in the Pequop Mountains–Wood Hills area reflects a two-stage process similar to that described in other, generally higher grade, regional metamorphic terrains in the Pyrenees, Japan, Scottish Highlands, and Korea (Lister et al., 1986; Miyake, 1993; Barker, 2002; Kim and Cho, 2008, respectively), but differs with regard to the first stage. The first growth stage in the other terrains is characterized as matrix replacement, whereas in the Pequop Mountains it is characterized by textural sector zoning that involves both matrix replacement and the crack-fill growth mechanism, although the crack-fill mechanism in this stage is restricted to separation of the matrix from {001} faces that are at a high angle to foliation (i.e., no intragrain extension fracturing). Nonetheless, the first stages in all these terrains, whether they involve textural sector zoning, can be considered the constructive phase, and the second phase, which similarly involves the crack-fill growth mechanism in all terrains, is the destructive phase. In addition, textural sector zoning in biotite may be an underrecognized process for two reasons. First, the textural sector growth process would not be apparent if a pelitic rock lacks graphite or excess quartz, because these are needed to produce inclusions and hence textural zoning. Second, because of the progression to the destructive phase within the biotite zone, zoning patterns become naturally obscured and are further obfuscated during progression through higher grades of metamorphism, which can make zoning difficult to recognize.

If the growth of biotite in the biotite zone is accompanied by a strong component of coaxial strain, biotite porphyroblasts will undergo a constructive growth phase characterized by textural sector zoning followed by a destructive phase involving fracturing, rotation, and residual growth. Relicts of constructive and destructive phase microstructures become obscured at higher metamorphic grades, but are still recognizable in kyanite zone rocks. The implications of this study, and important points about the progression from the constructive to the destructive phases, are the following:

• (1) Following nucleation of biotite, the growth of crystals with {001} at a high angle to foliation will become enhanced when grains attain a size large enough to create their own strain shadows and can sustain a growth rate high enough to overgrow matrix in a direction perpendicular to foliation. At this point, the constructive growth phase ensues and results in textural sector zoning of passive inclusions, which ultimately yields an hourglass-shaped distribution of inclusions in cuts viewed parallel or subparallel to the porphyroblast c axis. Zoning requires uninhibited growth in all directions and a delicate balance between growth and strain rate such that the growth of biotite in the direction of the c axis broadly keeps pace with retreat of the matrix from the {001} faces in dilating strain shadows.

• (2) The shapes of included hourglasses and porphyroblasts are variable and, in part, are controlled by strain as a function of whether the c axis was aligned with the maximum or minimum extension directions (i.e., parallel or perpendicular to lineation, respectively).

• (3) Inclusion trail geometry developed during textural sector growth can be variable on a thin-section scale. Inclusion trails may be straight or convergent reflecting nonrotation or curved reflecting multidirectional rotation of grains that grew with {001} inclined at a high angle to, but canted toward, foliation. Inclusion trails also may reflect growth over foliation wrapped around or perturbed by neighboring growing porphyroblasts.

• (4) The transition to the destructive phase occurs when porphyroblasts mature in size and cease to grow in a direction perpendicular to foliation. At this point, the textural sector growth mechanism no longer partially accommodates strain and some strain is transferred to the porphyroblasts, resulting in fracturing, rotation, and obscuring of zoning and inclusion trail patterns developed in the constructive phase. The destructive phase is also characterized by minor growth of biotite in strain shadows and extension fractures, and by the filling of strain shadows with mostly quartz instead of biotite.

• (5) The partitioning of strain and the constructive and destructive phases ultimately produce biotite porphyroblasts with inclusion trails that have a diversity of apparent dip angles and dip directions. Thus, caution must be used when inferring strain histories or designating porphyroblasts as syntectonic, intertectonic, or posttectonic on the basis of inclusions. Furthermore, caution should be used when interpreting inclusion trails in other mineral species that may chemically replace zoned biotite because the minerals may inherit the diverse array of inclusion trails of biotite.

• (6) Development of textural sector zoning in biotite during regional metamorphism differs from textural zoning reported in other mineral species because it is strain induced, i.e., it is syntectonic rather than pretectonic or intertectonic. Recognition of relict textural sector zoning in biotite porphyroblasts may be used to signify growth during either coaxial strain or general shear with a strong coaxial component. Furthermore, the presence of zoning in porphyroblasts that have c-axes perpendicular to lineation can be used to infer flattening strain.

This study was completed while I was on Faculty Development Leave at Austin Peay State University. I thank Geosphere reviewers Michael Wells and Colin Shaw and Associate Editor Michael Williams for providing comments and suggestions that helped to greatly improve this manuscript.