Superlative quantifiers ("at least 3", "at most 3") and comparative quantifiers ("more than 2", "fewer than 4") are traditionally taken to be interdefinable: the received view is that "at least n" and "at most n" are equivalent to "more than n-1" and "fewer than n+1", respectively. Notwithstanding the prima facie plausibility of this claim, Geurts and Nouwen (2007) argue that superlative quantifiers have essentially richer meanings than comparative ones. Geurts and Nouwen's theory makes three kinds of predictions that can be tested by experimental means. First, it predicts that superlative and comparative quantifiers should give rise to different patterns of reasoning. Second, the theory leads us to expect that children will master comparative quantifiers before superlative ones. Third, superlative quantifiers should be harder to process than comparative ones. We present three experiments that confirm these predictions. (Contains 7 footnotes and 8 tables.)