The underlying conceptual differences in exploiting two- and threefold rotational symmetry in the design of chiral ligands for asymmetric catalysis have been addressed in a comparative study of the catalytic performance of bisoxazoline (BOX) and tris(oxazolinyl)ethanes (trisox) containing copper(II) Lewis acid catalysts. The differences become apparent in constructing new catalysts by systematically "deforming" the stereodirecting ligand by inverting chiral centres or replacing chiral by achiral oxazolines. The catalytic alpha-amination of ethyl 2-methylacetoacetate with dibenzyl azodicaboxylate, which occurs with high enantioselectivity for both Ph(2)-BOX and Ph(3)-trisox copper catalysts, has been employed as the test reaction. In the trisox-copper complex [Cu(II)(iPr(3)-trisox)(kappa(2)-O,O'-MeCOCHCOOEt)](+)[BF(4)](-) (1), which was characterised by X-ray diffraction, two of the oxazoline groups are coordinated to the central copper atom, whilst the third oxazoline unit is dangling with the N-donor pointing away from the metal centre. A similar arrangement is found for the stereochemically "mixed" C(1)-trisox complex [Cu(II){(Ph(3)-trisox(R,S,S)}(kappa(2)-O,O'-MeCOCHCOOEt)(H(2)O)](+)[ClO(4)](-) (2), in which the phenyl substituents adopt a first coordination sphere meso arrangement. The almost identical selectivity of the Ph(3)-trisox(R,R,R)- and Ph(2)-BOX(R,R)-derived catalysts is as expected from the proposed model of the active catalyst based on a partially decoordinated podand. The behaviour of the "desymmetrised" trisox-Cu catalysts may be rationalised in terms of a general steady-state kinetic model for the three possible active bisoxazoline-copper species, which are expected to be in rapid exchange with each other in solution. This applies to both the trisox derivatives with stereochemically inverted and achiral oxazoline rings. The study underscores previous observations of superior performance of the catalysts bearing C(3)-chiral stereodirecting ligands as compared to systems of lower symmetry.