Spatial ability is the ability to mentally rotate a figure in two or three dimensions. It is reflected in visual-spatial and visual-motor tasks.
Autism spectrum disorders often go with poor performance on tasks requiring spatial ability, or with an aptitude profile in which spatial ability is low compared to verbal and numerical ability. Impaired or relatively low spatial ability shows up as (relatively) low Performance scores on a test like the W.A.I.S., and in the spatial factor of such a test, when factor scores are computed.
In individuals with Asperger's disorder, Performance scores are typically lower than Verbal scores. In individuals with classic autism, Verbal scores tend to be lower than Performance scores because of the additional language impairment, but Performance scores are low still.
Although motor clumsiness (which is an expression of relatively low spatial ability) is not required by the current diagnostic criteria for Asperger's or Autistic disorder, it was mentioned by Hans Asperger in his original description of the disorder, and is also included in the the ICD-10 criteria for Asperger syndrome published by the World Health Organization.
Hans Asperger, who discovered the disorder named after him, formulated the idea that autism is an extreme form of masculinity. Currently, this theory is being researched by Simon Baron-Cohen and others. Baron-Cohen gives impressive evidence for the theory, and names genetic differences and prenatal testosterone levels as possible causes of the extreme male brain[2, 3]. High testosterone levels give higher risk of autism, Baron-Cohen thinks. Proof of this is still being sought.
It is known that males on average have higher spatial ability than females. This is in fact, of all cognitive abilities, the one that shows the largest sex difference. In the light of the extreme male brain theory of autism, this leaves us with a paradox; if autism is an extreme form of masculinity, why do many autistic persons then have low spatial ability? One would expect the opposite. The solution to this paradox may lie in the relation between testosterone and spatial ability.
The relation found between the estrogen/testosterone balance and spatial ability is such that there is an optimal testosterone level for spatial ability within each sex, above and below which spatial ability is lower (inverted U within sex). The optimal testosterone levels lie below the male average and above the female average for testosterone, so that in fact the highest (within sex) spatial ability is found in females with (for females) well above average testosterone, and in males with (for males) below average testosterone.
Combined with the prenatal testosterone theory of autism (higher testosterone gives greater risk of autism), this predicts that females with autism will have (relatively) high spatial ability, while males with autism will have (relatively) low spatial ability. One would expect this to be expressed in Performance scores on I.Q. tests, performance on various visual-spatial and visual-motor tasks and possibly also in choice of occupation and hobbies.
If findings regarding the relation between spatial ability and testosterone are combined with the prenatal testosterone theory of autism, we get a possible explanation for the paradox that spatial ability is low in many autistic persons, in spite of the extreme male brain theory of autism on basis of which we would expect high spatial ability in autistic persons. This explanation predicts that females with autism will have relatively high spatial ability, while males with autism will have relatively low spatial ability.
 Schopler, Mesibov, Kunce (1998): Asperger Syndrome or High-Functioning Autism? ; Search index for I.Q. studies, verbal-performance discrepancy.
 Baron-Cohen (2002): The extreme male brain theory of autism.
 Lutchmaya, Baron-Cohen, Raggatt (2002): Foetal testosterone and vocabulary size in 18- and 24-month-old infants.
 Jensen (1998): The g factor; Chapter 13, "Sex differences in g".