Despite direction by the Conservation Act (1987) to give effect to the principles of the Treaty of Waitangi, New Zealand's Department of Conservation has few formal collaborative management arrangements with Maori. Obstacles to establishing agreements that involve Maori in equitable conservation decision-making roles include divergent philosophies (preservation versus conservation for future use), institutional inertia, a lack of concrete models of co-management to evaluate success or otherwise to promote conservation, a lack of resources and opportunities for capacity building and scientific research amongst Maori, opposition and a lack of trust from conservation non-governmental organizations that are predominantly euro-centric in approach and membership, and a fundamental reluctance of some to share power with Maori. Recent examples of work towards co-management emphasize the need for innovative methods to build trust and explore common ground and differences. Meetings on marae (traditional Maori gathering places) have established guiding principles, lengthy dialogue, and a collective symbol as a metaphor for co-management. These were valuable steps towards building trust and understanding required for the restoration of coastal lakes and a river, and the potential joint management of two national parks on the west coast of the North Island. Establishment of a research project to assess the sustainability of a traditional harvest of a sea-bird (Puffinus griseus) by Rakiura Maori was facilitated by drawing up a 'cultural safety' contract. This contract underscored the role of Maori as directors of the research, protected their intellectual property rights to their traditional environmental knowledge, guaranteed continuity of the collaborative research project and regulated how results were to be communicated. The scientific ethics of a university ecological research team were safeguarded by the contract, which ensured that they could publish their inferences without erasure or interference. The New Zealand experience shows that even when legislation signals from the top down that the doorway is open for co-management with indigenous people, this by itself is unlikely to make it happen. Active facilitation by innovative middle-level agreements and the creation of new administrative structures are needed to govern co-management of a broad spectrum of resource issues. Bottom-up initiatives involving single, or very localized, resource uses may also trigger co-management. Models for successful co-management involving indigenous peoples must focus more strongly on issues of equity or power sharing, and therefore may be very different from models directed at a single conservation outcome.