one/multi-dimensional scaling with incomplete dissimilarity matrix

The scaling functions I know about,

isoMDS(MASS)            Kruskal's Non-metric Multidimensional Scaling
cmdscale(mva)           Classical (Metric) Multidimensional Scaling
sammon(MASS)            Sammon's Non-Linear Mapping

all seem to require a complete dissimilarity matrix (ie, no 
dissimilarities missing).  Is there anything available for situations 
where not all dissimilarities are known, ie the dissimilarity matrix 
has (symmetrically) NAs in it?

David