There are numerous guidelines recommending that captive primates live in complex environments in which they have the opportunity to make choices and the ability to control aspects of the environment, despite the lack of quantitative evidence to suggest these qualities improve welfare. Complexity, choice and control (the ‘Three Cs’) are inter-related and therefore it is complicated to separate their effects. The main aim of this thesis was to examine how the ‘Three Cs’ affect welfare, using the common marmoset (Callithrix jacchus) as a model. Behavioural measures and preference tests were used to determine the impact and significance of the ‘Three Cs’ on welfare. Experimental manipulations were natural (i.e. access to outside runs), or unnatural (e.g. pressing a button to control additional illumination). In a series of different studies, marmosets were moved to larger and more complex enclosures, were allowed to choose between indoor cages and outdoor complex enclosures and were able to control additional white light or coloured lights in their home enclosures. The results of these studies show that appropriate levels of each of the ‘Three Cs’ had a positive influence on the welfare of the marmosets, especially on youngsters. Although having control over light, and increased illumination itself improved welfare, providing a choice of access to outside runs (which were more complex and allowed the marmosets greater control over their activities) resulted in the greatest welfare improvement for marmosets of all ages. Loss of access, or control, did not appear to have a negative impact. The marmosets were housed in pairs or in family groups, in the different studies. A cross-study comparison shows that the composition of the groups affected the behavioural response of adult marmosets to environmental enrichment. Unexpectedly, it was also found that, when housed in standard laboratory conditions, adult marmosets were more relaxed when housed in pairs than when housed with their offspring. A secondary aim of the thesis was to quantify welfare indicators and activity budgets of common marmosets in a range of different social and physical contexts, and to compare this with the behaviour of wild marmosets, to increase our understanding of what is “normal” in captive situations. It is concluded that it is critical to sub-divide locomotion and inactivity into different levels to interpret these measures accurately. Levels of calm locomotion increased in enriched environments, while levels of relaxed inactivity and scent marking decreased. A number of recommendations for the care and housing of marmosets are made.