On the basis of chiasma distributions along bivalents in human male me
iosis, genetic were counted for several chromosomal segments. These es
timates appeared to be lower than the lengths corresponding genetic le
ngths produced in the recombination analysis. When al so cytological d
istances and numbers of markers used in multilocus mapping were taken
into account, the regression estimates of genetic lengths were shown t
o satisfactorily fit the observed values. This indicates that mitotic
crossingover, genetic conversion, mutation and errors in allel's ident
ification in addition to meiotic crossingover, appear to contribute to
the observed genetic maps. It is suggested that these namely events,
rather than the typing errors suggested by Morton (1991), seem observe
d over those predicted on the basis of chiasma counts.