Thi~ paper con~iders the nature of preferences for the preservation of biodiversity, and the extent to which individuab are well-informed about biodiversity. We present evidence that the elicitation of monetary bids to pay for biodiversity preservation, as required for cost-benefit analysis, fails as a measure of welfare changes due to the prevalence of preferences which neoclassical economics defines as lexicographic. That is, a ~ignificant proportion of individuals refuse to make trade-offs which require the substitution of biodiver~ity for other goods. In addition, we show that understanding of the biodiver~ity concept is extremely limited, raising concerns over a reliance on stated preferences, as revealed in contingent valuation studies, for decision-making on this issue. Results from two samples (students and the general public) are described.