p300. footnote. Unless one takes the equivalence relation to be numerical equivalence, there is also the problem of proving that the category is abelian (in fact, it isn't). When one takes numerical equivalence, there is the problem of proving that the usual cohomology theories define fibre functors.
p305, 3t. The group at top-left should be $E/\mathfrak{a}$.
p314, 7.2.... is equivalent to this map's being…
p318, 1.1. It is not true that Theorem 1.1 (the Baily-Borel theorem) "can be regarded as a special case of the more general theorem of Nadel and Tsuji (1988)" --- Nadel and Tsuji claim this in their introduction, but, in fact, their results only apply when the boundary has dimension zero.
p340. In the proof of Theorem 5.5, one needs to verify that the cocycle obtained satisfies a certain continuity condition in order for the descent to be effective. See my paper: Milne, J. S. Descent for Shimura varieties. Michigan Math. J. 46 (1999), no. 1, 203--208.
p345. In the statement of Lemma 10.1, one needs to assume that the $E_{K}$ are separated.
p347, 3b. … acts as +1, not -1.
p354. Example 2.5 looks wrong.
p363, 4t. In the case that the weight…
p368. In order for $(\alpha,\beta)$ to be an isogeny, $\alpha$ must be injective with finite cokernel and $\beta$ must be surjective with finite kernel.
p371. The first line of the proof of 2.6 is misstated.
p374, 12b. degenerates.