Most of the work on treebank-based statistical parsing exclusively uses the WallStreet-Journal part of the Penn treebank for evaluation purposes. Due to the presence of this quasi-standard, the question of to which degree parsing results depend on the properties of treebanks was often ignored. In this paper, we use two similar German treebanks, T¨ Ba-D/Z and NeGra, u and investigate the role that different annotation decisions play for parsing.