In the scenario proposed in [63, 17] the "small" 
is used to fit the atmospheric neutrino data, while the "large"
is used to explain the LSND data. Then the survival
probability for solar neutrinos is predicted to be energy-independent.
This is not in good agreement with solar neutrino data, even taking
the 
neutrino flux as a free parameter [64]. In the approximation
of averaged atmospheric neutrino oscillation, and stretching the
data to their 90-95% C.L. limits, one can fit all the data with
two in the range 
and 1-2 
, and with a specific
mixing matrix [63].