Transition metal complexes with orbitally degenerate electronic states are geometrically unstable with respect to distortion from idealized symmetry to give structures with non-degenerate electronic states in accord with the Jahn-Teller theorem. This paper examines such distortions for metal complexes of the general type MLn (4 less than or equal to n less than or equal to 9) having open-shell electronic configurations, Such distortions can be of the following two types: (i) a classical Jahn-Teller distortion in which the topology of the polyhedron is maintained but the symmetry is lowered from a point group G to a smaller point group H where G is a subgroup of H; (ii) distortions in which both the topology and symmetry of the polyhedron are changed, typically by diamond-square processes, from a point group G(1) to G(2) where G(2) is not necessarily a subgroup of G(1). Certain octahedral metal complexes ML6 deviating from the favored closed-shell 18-electron rare gas electronic configuration are observed to be geometrically unstable to both types of distortions, e.g., elongation of a hyperelectronic O-h octahedron to a D-4h square bipyramid in octahedral d(9) Cu(II) complexes as compared with trigonal twists of hypoelectronic early transition metal dithiolates (e.g., Mo[S2C2(CF3)(2)](3)) and hexamethyls (e.g., W(CH3)(6) and Zr(CH3)(6)(2-)) to give D-3h or C-3v trigonal prismatic complexes. Open-shell tetrahedral metal complexes may similarly be unstable to both types of distortions, e,g, elongation or compression of opposite edges of a T-d tetrahedral complex ML4 to give a D-2d disphenoidal complex as compared with a 90 degrees twist of a d(8) tetrahedral complex to give a D-4h square planar complex, Similar geometric instabilities of open-shell penta-coordinate metal complexes may be resolved by partial diamond-square distortions to intermediate points along the trigonal bipyramid/square pyramid continuum to remove an of the orbital degeneracies. Analogous diamond-square processes in open-shell superoctahedral complexes MLn (n=7, 8, 9) along the pentagonal bipyramid/capped octahedron, bisdisphenoid/square antiprism, and tricapped trigonal prism/capped antiprism continua, respectively, are not sufficient to remove all degeneracies. However, neither hypoelectronic nor hyperelectronic superoctahedral metal complexes are expected to be stable. (C) 1998 Elsevier Science S.A.