R. W. FIELD, G. B. PARK, P. B. CHANGALA, J. H. BARABAN, Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139, USA; J. F. STANTON, Institute for Theoretical Chemistry, Departments of Chemistry and Biochemistry, The University of Texas at Austin, Austin, Texas 78712; A. J. MERER, Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan. Department of Chemistry, University of British Columbia, Vancouver, B.C., Canada V6T 1Z1.

Spectroscopy - it is all about patterns. Some patterns look so indescribably complicated that, unlike pornography, you do not know one when you see one. It is tempting to say that, at high vibrational excitation, interactions among normal mode basis states are so strong and widespread that all patterns are obliterated. But this is not true. When normal mode frequencies are in near integer multiple ratios, polyads emerge. A polyad is a robust pattern often comprising many vibrational eigenstates. Each such pattern might span many hundreds of cm-1, and it is inevitable that several unrelated polyad patterns overlap. When polyads overlap, it might seem impossible to disentangle them. However, the key to disentanglement is that polyads come in families in which successive generations are related by harmonic oscillator matrix element selection and scaling rules . Families of polyads are described by families of scaling-based effective Hamiltonian matrices, \mathbfH^\mathrmeff. No matter how complex and overlapped, the polyad \mathbfH^\mathrmeff serves as a magic decoder for picking out the polyad pattern. Sometimes the polyad patterns are systematically broken (a meta-pattern), owing to proximity to an isomerization barrier, as occurs in highly excited bending levels of the S1 state of HCCH, which encode the trans-cis minimum energy isomerization path. Quantum Chemists often dismiss \mathbfH^\mathrmeff models, precisely because they are models that do not express the full dimensionality of the complete Hamiltonian. But an \mathbfH^\mathrmeff explains rather than describes. Shunning \mathbfH^\mathrmeffs is like throwing out the baby with the bath water. Don't do it!