This paper explores some aspects of the crystal chemistry and structural constraints on tourmaline by examining 127 samples from the literature. According to the bond-valence model, the tourmaline structure shows lattice-induced strain at each polyhedron. The overall effect is an expansion of the triangular (BO3) group and compression of the tetrahedron. The X polyhedron can be either compressed or expanded: compression increases with vacancy content, whereas expansion is typical of Ca-rich tourmaline. The Y octahedron changes extensively from compressed through an unstrained to expanded state as a function of increasing Li content. The Z octahedron is almost unstrained in crystals with ∑ZR2+ < 0.40 apfu, whereas it is compressed in crystals with ∑ZR2+ > 0.40 apfu.
The configuration of the six-membered tetrahedral ring is strongly affected by <Y-O>, which is the most important parameter linked to the deviation of the tetrahedral ring from hexagonal symmetry. The whole structure is stable when the channels through the Z octahedron framework are able to accommodate the Y cations. As <Y-O> becomes larger, the less puckered the tetrahedral ring and the more the O7 atom is displaced away from Z. Consequently, the difference between <Y-O> and <Z-O> cannot be too large, otherwise <Z-O> will be too small to be commensurate with shifting of the O7 atom. One possible mechanism to reduce the difference between <Y-O> and <Z-O>, is the disordering reaction YAl + ZR → YR + ZAl, which increases <Z-O> and decreases <Y-O>. In ideal dravite, schorl, and “tsilaisite,” <Y-O> and <Z-O> are incommensurate.