Based on a bound constituent gluon picture and the gluon anomaly in QC
D, I argue that the decay branching ratios of the 2(++) glueball candi
date xi(2230) may satisfy B(xi --> eta'eta') greater than or equal to
0.6B(xi --> eta eta'), B(xi --> eta eta') greater than or equal to 5.6
B(xi --> eta eta) and B(xi --> eta eta) much greater than B(xi --> pi
pi). These features might provide a useful test for xi(2230).