The current status of our understanding of mechanistic details of GaAs growth by metal-organic chemical vapour deposition from various starting materials is reviewed. Despite a high level of recent activity in the study of precursor decomposition and reactivity there are still considerable gaps in our knowledge; for example, a clear differentiation between homogeneous (gas phase) and heterogenous (surface-controlled) processes is not yet possible. The decomposition of trimethylgallium in dihydrogen as carrier gas and the concomitant production of methane may occur by means of a radical process where hydrogen is abstracted by methyl radicals from the trimethylgallium rather than from the carrier gas. Triethylgallium on the other hand also offers the possibility of $\beta $-elimination as a facile pathway and this is reflected in the ratio ethene: ethane (3:1). In the presence of arsine there is a more facile pathway for hydrogen abstraction by alkyl radicals, giving rise in the case of triethylgallium and arsine to more ethane than ethene and in the case of trimethylgallium with arsine there is a difference in the reactivity to that found in studies of its decomposition in dihydrogen alone. Several workers have therefore deduced the participation of an intermediate adduct under growth conditions because arsine and trimethylgallium lower each others' decomposition temperature significantly. However, it should be remembered that arsine is almost certainly a better donor of hydrogen to methyl radicals than is dihydrogen. Phenylarsine has been found to be a potentially useful alternative precursor to arsine for the preparation of epitaxial GaAs. Thin films with excellent electrical and morphological characteristics have been prepared. Phenylarsine decomposes in dihydrogen to yield benzene and arsine and in the presence of trimethylgallium and triethylgallium mechanisms similar to those found with trimethylgallium and triethylgallium with arsine are suggested. These appear not to involve the dihydrogen carrier gas.