We used principal component analysis and logistic regression to evaluate the effect of 11 pond water quality variables on the presence and absence of Great Crested Newts (Triturus cristatus) in a cluster of 29 ponds in south-central Sweden. Variables of importance for the patterns observed were comprised into four principal components. Using logistic regression analysis and Akaike's Information Criteria (AIC) we found that the best model explaining the distribution of Great Crested Newts included three of the principal components. Temperature and nutrient levels (nitrogen and phosphorus) were important in distinguishing between ponds with and without Great Crested Newts, whereas other physical variables were less important. Ponds with newts had higher temperatures and nutrient levels than ponds where the species was absent. Our results also suggest that the Great Crested Newt selects ponds with low nutrient levels for breeding, whereas they may be present in ponds with higher nutrient levels. Although this study was performed in a single area with a limited sample the results raise several issues of general importance for the management and conservation of Great Crested Newts in pond landscapes.