In a recent PNAS, Barry Hall has a paper demonstrating
directed mutagenesis (he's calling it selection-induced mutation
now) in yeast. He shows that the mutation in question occurs
more often when adaptive than otherwise, that the mutation rate
increases at that locus only and that the mutation is time-
dependent not replication dependent. He also shows that the
appearance of the mutations cannot be explained by phenotypic
lag (i.e. the mutation happened a while back but the phenotype
hadn't been expressed yet.)

 The importance of the paper lies primarily in the fact
that he has shown this process (directed mutagenesis) occurs
in an organism distantly related to the one he and Cairns
originally showed it in (_E. coli_). This could either be a
case of convergent evolution or evidence that this process
is ancestral (i.e. present in the common ancestor to prokaryotes
(like _E. coli_ ) and eukaryotes (like yeast).

 The mechanism for directed mutagenesis is still unknown.

REFERENCE
Hall, 1992, Selection-induced mutations occur in yeast, PNAS 89:
4300-4303