T.J. Crow (1987) argued that psychosis is caused by a high rate of mutations occurring specifically in the course of gametogenisis of the male

.
The impression is that the fathers were relatively old too, their mean age
being between 31 and 36 years. Allen et al. (1971) and Gillberg (1982)
applied tentative reference groups, pointing at patient's fathers being 3-4
years older than the control.
When controlling for sex, social group, time of birth, and place of
birth we found that the mean paternal age (see Table V) in the three diagnostic
groups was raised 0.94-2.05 years, but only statistically significant
(p = .02) for the borderline childhood psychosis diagnostic group. For
mothers (see Table IV) the respective figures varied between -0.06 and
0.83 years and did not reach statistical significance. It is possible that earlier
findings of greater differences for mothers were influenced by lack of con
trol of relevant confounders or was partly a consequence of raised paternal
age (Hare & Moran, 1979).
Our findings are open to numerous interpretations. Even though
there is a paucity of scientific data there may be psychological effects on
the child of having an elderly father. The results are in accordance with
investigations of the age of parents in other psychiatric disorders. Hare and
Moran (1979) found father's age raised more than mother's, particularly
in schizophrenia. They concluded that the most likely explanation was that
the "constitutional characteristics" of the fathers of individuals who develop
psychiatric disorders predispose to late parenthood. Crow (1987) argued
that psychosis is caused by a high rate of mutations occurring specifically
in the course of gametogenesis in the male. The increased likelihood of
such mutations occurring with age was suggested to account for the association
of psychosis with increased paternal age.