The use of molecular genetic techniques is becoming increasingly wides
pread in analyses of hybrid zones. Yet, exactly how many molecular mar
kers are required to provide a given degree of resolution remains an o
pen question. We present statistical models that relate the number of
markers examined to their power to discriminate between pure species,
F1's, and backcross individuals. Our models indicate that only about f
our or five markers are required to provide a coarse classification of
individuals in hybrid zones, whereas upwards of 70 markers are requir
ed to discriminate between pure species and advanced backcrosses. The
models provide hybrid zone researchers a basis upon which to balance t
he time and expense of examining large numbers of markers against the
magnitude of classification errors when smaller numbers of markers are
examined.