Wilbur and Ague (2006) presented a detailed Monte Carlo model for chemical zonation in garnets, using texturally sector-zoned porphyroblasts. Two aspects of their article left me puzzled.
1) It is unclear why Wilbur and Ague do not regard textural sector zoning as ‘true’ sector zoning (even if it was not in the original definition). Sector zoning requires that adjacent sectors (or growth pyramids, the bases of which are the crystal faces) develop to some extent ‘independently’ of each other. In chemical sector zoning, this enables sectors of different crystallographic form to have different chemical compositions (e.g., staurolite; Hollister, 1970) while, in twin sector zoning, adjacent pyramids of the same form are twinned (e.g., grandite garnets; Jamtveit, 1991). In textural sector zoning, matrix-derived inclusions (type 1 inclusions) lie along and define sector boundaries (Andersen, 1984). In contrast, elongate quartz rods that form during garnet growth (type 2 intergrowths; Fig. 1B; Burton, 1986) lie perpendicular to the crystal face forming the base of a particular growth pyramid (sector) and indicate the lineage growth direction. These attest to the ‘independence’ of the growth direction of each sector. In crystals of homogenous form, such as rhombdodecahedral and icositetrahedral garnets, no chemical sector zoning can be present, even if twin and/or textural sector zoning is developed.
2) In the model proposed by Wilbur and Ague, garnet growth is concentrated at crystal edges, leading to hopper and dendritic habits. In 2-D sections, this gives ‘re-entrants’ in crystal faces (here termed ‘face re-entrants’ for simplicity) (Fig.1A; Wilbur and Ague's Fig. 3, hereafter called WAFig. 3). Wilbur and Ague correlate these to re-entrants between growth branches seen in photographs and chemical maps of their porphyroblasts (WAFigs. 1 and 2).
In textural sector zoning, sector patterns are defined by the type 1 inclusions and type 2 intergrowths (as described above); observed patterns vary in complexity, depending on the crystal habit and section orientation through the crystal (see http://homepage.univie.ac.at/alexander.hugh.rice/). In WAFig. 1A, sector boundaries are particularly well defined by type 1 inclusions, and in WAFig. 1E, they are defined by type 2 intergrowths.
Texturally sector-zoned crystals may develop re-entrants at the boundaries of crystal faces (‘edge re-entrants;’ Fig. 1B). Thus, by definition, these occur at places where type 1 inclusions intersect an edge of the crystal (or an edge of a particular growth-zone, since subsequent growth can fill re-entrants). Edge re-entrants form when sectors on opposite sides of a sector boundary grow too quickly perpendicular to their faces for growth parallel to the faces to fill the incipient re-entrant (cf. Andersen, 1984); that is, through slow-edge growth.
Although all the garnets in WAFig. 1A–1E have re-entrants, these occur at points where type 1 inclusions reach the rim of the central growth zone (organic matter-free zone) and are thus edge re-entrants, typical of textural sector zoning, rather than the face re-entrants of dendritic growth. Comparison of the textural and chemical development of the porphyroblasts (Fig. 2A–F; WAFigs. 1 and 2) shows that the chemical branches mapped lie between these edge re-entrants and therefore reflect enhanced growth on the crystal faces, rather than at the crystal edges.
Thus the modeled chemistry, which has enhanced crystal edge growth, resulting in face re-entrants, differs entirely from the observed chemistry with enhanced crystal face growth and edge re-entrants. The former, as noted by Wilbur and Ague, is consistent with dendritic growth, which would seem to lie at the opposite end of a textural spectrum from textural sector zoning (Fig. 1), the one characterized by rapid edge growth and the other by inhibited edge growth.
I thank Jay Ague for details concerning the relative orientations of the textural figures and chemical maps.