It is pointed out that the difficulty in boson models in explaining the lar
ge B(E2) experimental data is not due to the lack of collectivity in boson
model wavefunction. This long standing problem of reduction in collectivity
of wavefunction in boson models can be solved by choosing an appropriate E
2 transition operator in the SU(3) limit of the sdg-interacting boson model
. The ratio B(E2, L + 2 --> L)/B(E2, 2 --> 0) can be almost any large numbe
r.