A review of studies of part-whole valuation effects in contingent valuation highlights the difficulty of distinguishing part-whole bias from the effect of substitution among goods. A contingent valuation of instream flow preservation indicates that respondents with more information about substitutes were more sensitive in their valuations to the number of rivers protected than were respondents with less information. These results, in combination with those of other studies of part-whole valuation effects, suggest that contingent valuation researchers must design guidelines for deciding what information about substitutes should be presented to respondents and how that information should be presented. ¿ American Geophysical Union 1995