The current study is pan of a two-wave normative data collection study with the Rorschach Comprehensive System (CS; Exner, 2003) in The Netherlands. The first data wave yielded 108 valid Rorschach protocols for which descriptive data on CS variables were calculated. Interrater reliability between two raters for 10 response segments on a subset of 24 protocols was good to excellent. The findings show that mean values for a number of important CS variables, such as EA, Egocentricity index, and WDA% in the Dutch sample are highly similar to those found in a recent U.S. study by Shaffer and colleagues (2007/this issue). On the other hand, the Dutch sample contains a higher percentage of subjects with positive scores on the DEPI and CDI, compared with the Shaffer et al. data, but to an even greater extent in comparison with Exner's (2001) nonpatient data. We argue, on the basis of findings from psychiatric epidemiology, that the exclusion of subjects with a history of mental disorder or mental health treatment from Rorschach normative data bases yields a dataset that is nonrepresentative of the general population.