Anuran amphibians have traditionally been considered suitable model organisms for community ecology studies. In this paper, we review and quantitatively re-analyze the original studies published on dietary relationships among anuran communities worldwide. We re-analyzed 33 independent communities from all continents, including data on prey numbers and/or prey volumes. All datasets underwent null model re-analysis using two randomization algorithms. In addition, logistic regression models were used to identify specific assemblage patterns. We discovered three main patterns: (1) one particular algorithm (RA2, which replaces every cell in the matrix with one randomly chosen, but retains the zero structure of the matrix) performed much better at uncovering community structure; (2) nonrandom structure was more likely to occur in tropical than in nontropical regions; (3) prey volume was a better descriptor of dietary relationships than prey number when attempting to detect nonrandom structure. We showed that: pattern (1) depended on both the overall generalist nature of the anurans in general and on the mathematical properties of RA2; pattern (2) was consistent with theoretical attributes of tropicality; and pattern (3) was due to the intrinsic ecological properties of generalist organisms, which forage more or less opportunistically on prey of very different sizes; hence, when prey number is used as a diet descriptor, there is a risk of approximating different prey taxa that make different relative volume contributions to the overall diet as being of similar relevance .