Numerous hypotheses have been developed to explain how a brood parasite selects a host nest into which it lays its eggs. Most hypotheses address various aspects of nest placement. We used an extensive dataset to tease apart the relative strength of various hypotheses associated with nest placement and timing. These data were from North American tallgrass prairie and included nearly 2,000 nests of 17 host species known to accept eggs of the brood parasitic Brown-headed cowbird Molothrus ater. Regression tree analyses, with host species as a categorical covariate built around successive logistic regressions, implied that the ''edge effect'' and ''perch proximity'' hypotheses, the latter assessed as distance to woody vegetation, received the strongest support. Hypotheses concerning nest height, livestock proximity, habitat density, nest exposure, laying date, and host clutch size received weaker or subsidiary support, the latter meaning that the hypotheses received significant support only after edge effects or distance to woody vegetation were accounted for first. Host species was associated significantly with parasitism rate, but host species was itself correlated with various vegetation and landscape variables that we assessed. Brood parasitism rate and nest height were associated non-linearly. In addition to a clear hierarchical pattern among factors associated with rates of parasitism, several key explanatory variables had marked interactions, such as prairie edge and extent of woody vegetation or nest height and nest exposure. Such interactions, including between host species and certain landscape and vegetation variables, such as nest height and distance from woody vegetation, suggest caution is warranted when assessing the various competing hypotheses, which are not mutually exclusive. Keywords Brood parasitism Á Edge effects Á Molothrus ater Á Nest placement Á Regression trees Zusammenfassung Es gibt zahlreiche Hypothesen zur Erklärung, wie ein Brutparasit ein Nest für die eigene Eiablage aussucht, wobei sich die meisten dieser Hypothesen an der spezifischen Lage der Nester ausrichten. Wir benutzten eine umfangreiche Datensammlung, um die relative Stärke derjenigen Hypothesen zu entwirren, die sich auf Nestlage und Eiablage-Zeitpunkt beziehen. Die Daten stammten aus der nordamerikanischen Tallgrass-Prärie, von fast 2000 Nestern 17 unterschiedlicher Wirtsarten, von denen bekannt ist, daß sie Eier des Brutparasiten Braunkopf-Kuhstärling (Molothrus ater) in ihren Nestern akzeptieren. Regressionsbaumanalysen mit den Wirtsspezies als kategorialen Kovariaten, um sukzessive logistische Regressionen herum gerechnet, unterstützten am stärksten die Hypothesen zum,, Rand-Effekt'' und zur ,,Sitzplätze-Nähe'' (Entfernung zu Gebüsch-Vegetation). Andere Hypothesen, die die Nesthöhe, die Nähe zu Vieh, die Habitatdichte, die Nestlage, oder den Zeitpunkt der Eiablage und die Gelegegröße des Wirtes heranziehen, wurden schwächer oder nur ,,subsidiär'' unterstützt -wobei letzteres bedeutet, dass diese Hypothesen nur dann signif...