First graph is from the Autism Research Institute (though named "Institute for Child Behavior Research" during most of this graph). Second is my re-working to show the changing ratio of age of onset. How easily is this data compatible with no increase?