The research reported in this paper is part of a larger project on the classification of Web pages by genre. Such classification is a potentially powerful tool in filtering the results of online searches. In this paper, we describe two sets of experiments investigating the automatic classification of Web pages by their genres. In these experiments, our approach is to represent the Web pages by profiles that are composed of fixed-length byte n-grams. The first set of experiments in this study examines the effect of three feature selection measures on the accuracy of Web page classification. The second set of experiments in this study compares the classification accuracy of three classification methods, each using n-gram representations of the Web pages. The classification methods which are compared are a distance function approach, the k-nearest neighbors method, and the support vector machine approach. We also examine a range of n-gram lengths and a range of Web page profile sizes to determine what combination(s) of n-gram length and profile size give the best classification accuracy. Each set of experiments is run on two well-known data sets, 7-Genre and KI-04, for which published results are available.