Autism is a strongly genetic disorder where risk-conferring variation in a number of genes contributes to the phenotype. Twin studies show incomplete concordance for autism in monozygotic sibling pairs, suggesting that environmental and/or epigenetic effects also contribute to the disorder. This thesis investigates association of clock gene variants with autistic disorder. Significant indirect positive genetic association for autistic disorder was found for two single nucleotide polymorphisms in the clock gene PER1 and for two single nucleotide polymorphisms in the clock gene NPAS2. Bioinformatics analysis of these single nucleotide polymorphisms showed: SNP rs885747 disrupts a splicing enhancer/suppressor element, SNP rs34705978 is within a differentially methylated control element, SNP rs6416892 is four nucleotides from the tissue specific binding site of sterol regulatory element binding transcription factor 2 and SNP rs1811399 alters the structure of a candidate microRNA. Investigation within the most significant haplotype in NPAS2 highlighted a conserved circadian regulatory element (RRE) whilst that of PER1 contained alternative and essential splice site SNPs. Analysis of genes containing conserved circadian regulatory elements, the E-box, D-box and RRE, showed that some of the strongest candidate genes for autism, schizophrenia and bipolar disorder are likely to be circadian clock controlled genes. Synchronization of high frequency oscillations between different brain regions is a correlate of normal brain function that is altered in autism; a role for clock genes in regulating the dendritic architecture of high frequency neural oscillators is discussed. Interplay between the molecular processes of the circadian clock, the sex determination pathway and alternative splicing is highlighted as the basis of a hypothesis suggesting how clock gene mutations might also determine short period oscillator phenotypes.