Glueball candidateι(1440), anomalous Ward identities, and two-photon decays

Anomalous Ward identities are given for the U(1) problem, showing that some recent papers have neglected the large topological susceptibility coming from the pure Yang-Mills sector of QCD. A reanalysis of the Ward identities is given, including the pseudoscalar glueball candidate $\iota$ (1440) with the pseudoscalar nonet. It is shown that positivity of the topological susceptibility together with other constraints is sufficient to narrow down the permitted range of pseudoscalar axial couplings. In particular the $\iota$ (1440) couplings are consistent with those expected for a glueball with the decay $\iota \to \gamma \gamma$ probably immeasurably small. Contrary to a recent claim, the results are not sensitive to the branching ratio for $\iota \to K\overline{K}\pi$, which may be as large as 100%.