Understanding the mutational process is essential before relationships between observed variation and genetic distance or population substructure can be inferred. Most of the current estimators of these relationships were developed based on interpretation of allozyme variation. In allozymes, it was assumed that most new mutations gave rise to new distinguishable alleles. In addition, the electrophoretic mobility, upon which alleles were assigned, could not be used to assess the mutational relationship amongst the alleles. This process was modelled by an Infinite Allele Model (IAM) where every new mutation is assumed to give rise to a new electrophoretically distinguishable allele (or electromorph). It is obvious that some homoplasy (i.e. an electromorph which comprises two or more alleles) exists in allozyme markers, however, the IAM has proved very successful in explaining the observed allozyme variation (Nei 1987). The situation is very different in microsatellite markers. First, since most of the mutations seem to involve the gain or loss of a single repeat unit (Weber and Wong 1993), it is clear that there exists a high amount of homoplasy. This is important as homoplasy leads to the underestimation of the total amount of variation and genetic distance, and to the overestimation of the similarities among populations. The Stepwise Mutation Model (SMM) has been used to simulate this situation. In this model alleles can only mutate by the gain and loss of 1 repeat unit. These two models, the IAM and SMM, represent the extremes of the situation. In the IAM no homoplasy exists, while in the SMM, a large amount is present.
Microsatellites have been estimated to mutate at rate between 
and 
mutations per gamete
(Edwards et al. 1992;
Bowcock et al. 1994;
Forbes et al. 1995). However,
the mechanisms by which microsatellites mutate are poorly understood.
Two main mechanisms have been proposed, which may act in concert; 1)
unequal crossing over in meiosis and 2) strand-slippage replication
(Levinson and Gutman, 1987a). Of these,
strand-slippage replication appears to be the predominant mode at
microsatellites (Wolff et al. 1989). Strand-slippage is speculated to occur primarily during lagging
strand synthesis (Schlotterer and Tautz, 1992). For example, it may involve the slippage of the newly
synthesised DNA strand upon dissociation of a polymerase complex. This
slippage creates a transient bulge which upon DNA repair would be
either removed or lead to the elongation of the repeat
(Schlotterer and Tautz, 1992).
Alternatively, the formation of a transient bulge in the template
strand may lead to the shortening of the repeat.
