Different aspects of expanded polyglutamine tracts and of their pathogenetic role are taken into consideration here. (i) The (CAG)n length of wild-type alleles of the Huntington disease gene was analysed in instability-prone tumour tissue from colon cancer patients to test whether the process leading to the elongation of alleles towards the expansion range involves single-unit stepwise mutations or larger jumps. The analysis showed that length changes of a single unit had a relatively low frequency. (ii) The observation of an expanded spinocerebellar ataxia (SCA)1 allele with an unusual pattern of multiple CAT interruptions showed that cryptic sequence variations are critical not only for sequence length stability but also for the expression of the disease phenotype. (iii) Small expansions of the (CAG)n sequence at the CACNA1A gene have been reported as causing SCA6. The analysis of families with SCA6 and episodic ataxia type 2 showed that these phenotypes are, in fact, expressions of the same disorder caused either by point mutations or by small (CAG)n expansions. A gain of function has been hypothesized for all proteins containing an expanded polyglutamine stretch, including the alpha 1A subunit of the voltage-gated calcium channel type P/Q coded by the CACNA1A gene. Because point mutations at the same gene with similar phenotypic consequences are highly unlikely to have this effect, an alternative common pathogenetic mechanism for all these mutations, including small expansions, can be hypothesized.