One increase for the number of diagnosed cases of autism is just that: more cases are being diagnosed. That is, more children who once were simply labeled "slow" or "difficult" or some other adjective are now diagnosed with autism, even though very little is known about autism itself. My wife teaches special ed at an elementary school, so I hear about this all the time. "Autism" has become a sort of catch-all for kids who have strange issues that do not respond to conventional treatements. That's not to say that there are not some genuine cases of autism, but many things that once were not labeled autism now are. That is one big reason that we see a rise in cases--not because more kids have it than used to, but because we are more liable to diagnose something as "autism" today than we were 20 years ago.
The conservative commentator Thomas Sowell has an interesting book on children who start talking late; he had such a child himself.
An increase in the number of diagnoses vs. the number of cases that exist, is not really the point. We are really more concerned about why some kids are autistic--however many it may be.
There are so many things in our environment that have changed since the 70's, including the levels of estrogens, hormones, food colorings, and preservatives in the diet; the levels of use of electronics and filtered air in the home; the types of chemicals that are used to manufacture building materials, paints and carpets; chemicals that have been removed from the environment like lead in paint and pollution; personal health habits like the use of Advil and acetaminophen rather than aspirin; the increased use of birth control pills, latex condoms, latex gloves, factory-produced vitamins, insecticides; and there is an increase in the routine use of sonograms during pregnancy.
Any one of these things or something completely unrelated could be the culprit.
Keeping in mind that autism affects three times as many boys as girls, it would probably behoove researchers to explore the testosterone connection.
Not very long ago, autism was associated with low IQ test performance but more and more you see very bright kids who do perfectly fine on standardized tests including IQ tests being labeled as "autistic" just because they are a little bit square pegs being pushed into round holes. Poor grades shouldn't count for any kind of diagnosis because if a kid does well on standardized tests, he's learning --- even if his grades are very poor. Behavior problems or distraction, mind-wandering to something more interesting can cause bad grades.