I had thought Alice in Wonderland would have quite a few hapax legomena, even if mostly nonsense words I think (ie 'twas brillig and the slithy toves...), but I found from this academic article [0] that Twain's Tom Sawyer actually has 5% more hapax legomena. The article has some pretty surprising findings about ratios of hapax/vocabulary - though hapaxes fairly consistently make up around 50% of the words in any text, they steadily increase in corpora over 3,000,000 words.