This article explores constraints encountered when using PRA on an ODA-funded natural resources project in a tribal area of Western India. It was particularly evident that women's participation in the PRAs was minimal. The reasons for this were practical (women were not available collectively for long periods of time & there were few women fieldworkers as the project had just begun), social (PRA activities tended to take place in public places where women felt awkward) and methodological (women respond to PRA activities in different ways, sometimes feeling bored and "communicating by singing instead"). The author argues that an organised PRA "gives privilege to certain kinds of knowledge and representation and suppresses others" : the emphasis given to formal knowledge and activities tends to "reinforce the invisibility of women's roles". However, once the formal and public nature of PRA is perceived as a problem, it can become a means by which "women's knowledge and activities.. can be transferred from the informal to the formal arena of project planning", thereby increasing women's profile. Suggestions for encouraging women's participation in PRA include: making non-public contexts (since women are more used to the "private sphere"), using women's knowledge and ways of communicating (songs, sayings, proverbs). There are constraints: the "production of observable outputs (maps, diagrams of PRA) have more status for fieldworkers" than scribbled songs or informal interview notes and women's expressed needs (eg a flour mill) "don't fit easily into established categories of natural resource development".