A new classification and nomenclature scheme for the amphibole-supergroup minerals is described, based on the general formula AB₂C₅T₈O₂₂W₂, where A = □, Na, K, Ca, Pb, Li; B = Na, Ca, Mn²⁺, Fe²⁺, Mg, Li; C = Mg, Fe²⁺, Mn²⁺, Al, Fe³⁺, Mn³⁺, Ti⁴⁺, Li; T = Si, Al, Ti⁴⁺, Be; W = (OH), F, Cl, O²⁻. Distinct arrangements of formal charges at the sites (or groups of sites) in the amphibole structure warrant distinct root names, and are, by implication, distinct species; for a specific root name, different homovalent cations (e.g., Mg vs. Fe²⁺) or anions (e.g., OH vs. F) are indicated by prefixes (e.g., ferro-, fluoro-). The classification is based on the A, B, and C groups of cations and the W group of anions, as these groups show the maximum compositional variability in the amphibole structure. The amphibole supergroup is divided into two groups according to the dominant W species: ^W(OH,F,Cl)-dominant amphiboles and ^WO-dominant amphiboles (oxo-amphiboles). Amphiboles with (OH, F, Cl) dominant at W are divided into eight subgroups according to the dominant charge-arrangements and type of B-group cations: magnesium-iron-manganese amphiboles, calcium amphiboles, sodium-calcium amphiboles, sodium amphiboles, lithium amphiboles, sodium-(magnesium-iron-manganese) amphiboles, lithium-(magnesium-iron-manganese) amphiboles and lithium-calcium amphiboles. Within each of these subgroups, the A- and C-group cations are used to assign specific names to specific compositional ranges and root compositions. Root names are assigned to distinct arrangements of formal charges at the sites, and prefixes are assigned to describe homovalent variation in the dominant ion of the root composition. For amphiboles with O dominant at W, distinct root-compositions are currently known for four (calcium and sodium) amphiboles, and homovalent variation in the dominant cation is handled as for the ^W(OH,F,Cl)-dominant amphiboles. With this classification, we attempt to recognize the concerns of each constituent community interested in amphiboles and incorporate these into this classification scheme. Where such concerns conflict, we have attempted to act in accord with the more important concerns of each community.

The International Mineralogical Association's approved amphibole nomenclature has been revised in order to simplify it, make it more consistent with divisions generally at 50%, define prefixes and modifiers more precisely and include new amphibole species discovered and named since 1978, when the previous scheme was approved. The same reference axes form the basis of the new scheme and most names are little changed but compound species names like tremolitic hornblende (now magnesiohornblende) are abolished and also crossite (now glaucophane or ferroglaucophane or magnesioriebeckite or riebeckite), tirodite (now manganocummingtonite) and dannemorite (now manganogrunerite). The 50% rule has been broken only to retain tremolite and actinolite as in the 1978 scheme so the sodic calcic amphibole range has therefore been expanded. Alkali amphiboles are now sodic amphiboles. The use of hyphens is defined. New amphibole names approved since 1978 include nyböite, leakeite, kornite, ungarettiite, sadanagaite and cannilloite. All abandoned names are listed. The formulae and source of the amphibole end member names are listed and procedures outlined to calculate Fe³⁺and Fe²⁺when not determined by analysis.

A nomenclature for tourmaline-supergroup minerals is based on chemical systematics using the generalized tourmaline structural formula: XY₃Z₆(T₆O₁₈)(BO₃)₃V₃W, where the most common ions (or vacancy) at each site are X = Na¹⁺, Ca²⁺, K¹⁺, and vacancy; Y = Fe²⁺, Mg²⁺, Mn²⁺, Al³⁺, Li¹⁺, Fe³⁺, and Cr³⁺; Z = Al³⁺, Fe³⁺, Mg²⁺, and Cr³⁺; T = Si⁴⁺, Al³⁺, and B³⁺; B = B³⁺; V = OH¹⁻ and O²⁻; and W = OH¹⁻, F¹⁻, and O²⁻. Most compositional variability occurs at the X, Y, Z, W, and V sites. Tourmaline species are defined in accordance with the dominant-valency rule such that in a relevant site the dominant ion of the dominant valence state is used for the basis of nomenclature. Tourmaline can be divided into several groups and subgroups. The primary groups are based on occupancy of the X site, which yields alkali, calcic, or X-vacant groups. Because each of these groups involves cations (or vacancy) with a different charge, coupled substitutions are required to relate the compositions of the groups. Within each group, there are several subgroups related by heterovalent coupled substitutions. If there is more than one tourmaline species within a subgroup, they are related by homovalent substitutions. Additionally, the following considerations are made. (1) In tourmaline-supergroup minerals dominated by either OH¹⁻ or F¹⁻ at the W site, the OH¹⁻-dominant species is considered the reference root composition for that root name: e.g., dravite. (2) For a tourmaline composition that has most of the chemical characteristics of a root composition, but is dominated by other cations or anions at one or more sites, the mineral species is designated by the root name plus prefix modifiers, e.g., fluor-dravite. (3) If there are multiple prefixes, they should be arranged in the order occurring in the structural formula, e.g., “potassium-fluor-dravite.”

Published two-body bond-valence parameters for cation–oxygen bonds have been evaluatedviathe root mean-square deviation (RMSD) from the valence-sum rule for 128 cations, using 180 194 filtered bond lengths from 31 489 coordination polyhedra. Values of the RMSD range from 0.033–2.451 v.u. (1.1–40.9% per unit of charge) with a weighted mean of 0.174 v.u. (7.34% per unit of charge). The set of best published parameters has been determined for 128 ions and used as a benchmark for the determination of new bond-valence parameters in this paper. Two common methods for the derivation of bond-valence parameters have been evaluated: (1) fixingBand solving forR_o; (2) the graphical method. On a subset of 90 ions observed in more than one coordination, fixingBat 0.37 Å leads to a mean weighted-RMSD of 0.139 v.u. (6.7% per unit of charge), while graphical derivation gives 0.161 v.u. (8.0% per unit of charge). The advantages and disadvantages of these (and other) methods of derivation have been considered, leading to the conclusion that current methods of derivation of bond-valence parameters are not satisfactory. A new method of derivation is introduced, the GRG (generalized reduced gradient) method, which leads to a mean weighted-RMSD of 0.128 v.u. (6.1% per unit of charge) over the same sample of 90 multiple-coordination ions. The evaluation of 19 two-parameter equations and 7 three-parameter equations to model the bond-valence–bond-length relation indicates that: (1) many equations can adequately describe the relation; (2) a plateau has been reached in the fit for two-parameter equations; (3) the equation of Brown & Altermatt (1985) is sufficiently good that use of any of the other equations tested is not warranted. Improved bond-valence parameters have been derived for 135 ions for the equation of Brown & Altermatt (1985) in terms of both the cation and anion bond-valence sums using the GRG method and our complete data set.