Demographics are always tricky to decipher, and it is easy to make too much of too little, but at some point all sorts of families began having autistic kids, and that ominous pattern may well be visible in the first 100 cases.
The latest research (not from 1938) that I've seen on the causes of autism seems to show at least one link to older fathers.
Nothing is terribly clear in complex genetics. For example a predisposition to autism that doesn't manifest as autism may tend to make people very intelligent.
There may be an environmental window for an outside factor to occur to activate autism; if the child makes it past that window, they are no longer susceptible.
I agree that parental age may be a factor, but is it age, or is it something else the parents "picked up" along the way, like smoking, drinking, diet, etc., that had a cumulative effect.
And yet, in practical terms, intelligent people may well be statistically more likely to have autistic children.
But as with another recent discovery, Fragile X Syndrome, it might be very obvious or very subtle.