Major issues involved in the classification of the amphiboles are examined: (1) the role of (OH), Li and Fe3+, (2) the formal definition of a root name, (3) irreducible charge-arrangements and distinct species, (4) the use of prefixes, (5) the principal chemical variables used in a classification procedure, and (6) the use of the dominant-constituent principle. The current IMA-approved classification scheme is based on the A, B and T groups of cations in the amphibole formula: AB(2)C(5)T(8)O(22)W(2). We argue here that classification should be based on the A, B and C groups of cations as (i) it is in these groups of cations that the maximum variation in chemical composition occurs, and (ii) as a result of (i), the scheme is more in accord with the IMA-sanctioned dominant-constituent principle, which governs the recognition (and approval) of distinct mineral species. Two new classifications are presented here: one is based on the A, B and C groups of cations, and another on the dominant-constituent principle. These two schemes were produced to illustrate (i) the problems inherent in the classification of a group of minerals as complicated as the amphiboles, and (ii) the sometimes disparate needs of crystallographer, mineralogist, petrologist and geochemist. Scheme 1 conserves current formulae and names as much as possible, whereas scheme 2 minimizes the number of formulae and names as much as possible. The differences between the current classification and the two schemes presented here are discussed, and we highlight the problems associated with each scheme.
