1 Correction to: Math. Z. (2013) 275:863–897 https://doi.org/10.1007/s00209-013-1163-8

  • On p. 874: In the statement of Lemma 2.2, which reads “Let V be a real finite dimensional vector space endowed with a non-positive definite and non-degenerate quadratic form”, should read “Let V be a real finite dimensional vector space endowed with a indefinite non-degenerate quadratic form”.

  • On p. 875 at item (2) of Theorem 2.7, the last line should be replaced by:

In particular, we get \(\partial _t\log |\mathcal {J}(A_t(x)v)|\ge \delta (X_t(x)), x\in V, t\ge 0, v\in E_x\) and \(\mathcal {J}(v)>0\); or \(\partial _t\log |\mathcal {J}(A_t(x)v)|\le \delta (X_t(x)), x\in V, t\ge 0, v\in E_x\) and \(\mathcal {J}(v)<0\).

  • On p. 876: although not used anywhere in the article, item (5) of Theorem 2.7 is false. The proof presented in p. 878 has a wrong sign in the calculation. The corrected calculation gives only \(\partial _t\left( \frac{|\mathcal {J}(A_t(x)w)|}{\mathcal {J}(A_t(x)v)}\right) \mid _{t=0}\ge 0\) for the \(\mathcal {J}\)-separated cocycle \(A_t(x)\) and strictly positive if \(A_t(x)\) is a strictly \(\mathcal {J}\)-separated cocycle.

  • On p. 884: items (1) and (2) in the statement of Theorem 2.23 in [1] need to be corrected, since [1, Example 5 with index 2] shows that hyperbolic behavior of the subbundles cannot be expressed as necessary and sufficient conditions using the sign of the function \(\delta \). Item (3) of [1, Thm. 2.7] is correct. We reformulate the statement of items (1) and (2) as follows.

Theorem 0.1

[1, Theorem 2.23] Let \(\Gamma \) be a compact invariant set for \(X_t\) admitting a dominated splitting \(E_\Gamma = F_-\oplus F_+\) for a linear multiplicative cocycle \(A_t(x)\) over \(\Gamma \) with values in E. Let \(\mathcal {J}\) be a \(C^1\) field of indefinite quadratic forms such that \(A_t(x)\) is strictly \(\mathcal {J}\)-separated admitting a function \(\delta :\Gamma \rightarrow {\mathbb R}\) as given in Theorem 2.7. Then

  1. (1)

    If \(\Delta _s^t(x)\xrightarrow [(t-s)\rightarrow +\infty ]{}-\infty \) for all \(x\in \Gamma \), then \(F_-\) is a uniformly contracted subbundle.

  2. (2)

    If \(\Delta _s^t(x)\xrightarrow [(t-s)\rightarrow +\infty ]{} +\infty \) for all \(x\in \Gamma \), then \(F_+\) is a uniformly expanding subbundle.

Since the short proof of these items uses the corrected expressions in item (2) of Theorem 2.7, we present it below.

Proof

If \(\Delta _0^t(x)\xrightarrow [t\rightarrow +\infty ]{}-\infty \), then from item (2) of [1, Thm 2.7] we get \(\frac{\mathcal {J}(A_t(x)v)}{\mathcal {J}(v)}\le e^{\Delta _0^t(x)}\xrightarrow [t\rightarrow +\infty ]{}0 \) for all \(x\in \Gamma \) and \(v\in F_-(x)\). So \(F_-\) in uniformly contracted, by [1, Lemmas 2.18 & 2.24].

If \(\Delta _s^t(x)\xrightarrow [(t-s)\rightarrow +\infty ]{} +\infty \), then analogously \( \frac{\mathcal {J}(A_t(x)v)}{\mathcal {J}(v)}\ge e^{\Delta _0^t(x)}\xrightarrow [t\rightarrow +\infty ]{}+\infty \) for all \(x\in \Gamma \) and \(v\in F_+(x)\). So \(F_+\) in uniformly contracted, again by [1, Lemmas 2.18 & 2.24]. \(\square \)