Three notable references to the use of MULTAN for the solution of the partial structure
are by Navia et al [87] and Wilson [118] where the isomorphous differences between
native and derivative
structures 
were utilised, and Mukherjee et al [86] where
Bijvoet differences 
were used. In the latter reference, heavy atom positions
were determined for three structures: Cytochrome-
containing two iron atoms per asymmetric unit,
Pea Lectin with two manganese atoms in the asymmetric unit and a mercury derivative of 
-amylase
with one mercury atom per asymmetric unit. In these cases Bijvoet differences were used for the calculation
of normalised structure factors.
In each of the cases studied, the phase set with the highest combined figure of merit corresponded to the correct solution when compared with MIR results. For two of the structures however the heavy atom positions were displaced by 2.5 Å and 5.0 Å away from the refined MIR positions.
Mukherjee et al showed that the largest Bijvoet differences measured provide a reasonable estimate
of the magnitude of the imaginary part of the heavy atom structure factor, since (from Eq. 
)
and for the largest 
's where 
and 
are orthogonal we see that
This is convenient since MULTAN uses a set of largest normalised structure factors, 
's, in the
calculation of triple product relations. Thus the measured may be used directly as input
to MULTAN in a manner similar to using 
for a small molecule.