A correlation between the degree of specialization for food hoarding and the volume of the hippocampal formation in passerine birds has been accepted for over a decade. The relationship was first demonstrated in family-level comparisons, and subsequently in species comparisons within two families containing a large number of hoarding species, the Corvidae and the Paridae. Recently, this approach has been criticized as invalid and excessively adaptationist. A recent test of the predicted trends with data pooled from previous studies found no evidence for such a correlation in either of these two families. This result has been interpreted as support for the critique. Here we reanalyse the original dataset and also include additional new data on several parid species. Our results show a surprising difference between continents, with North American species possessing significantly smaller hippocampi than Eurasian ones. Controlling for the continent effect makes the hoarding capacity/hippocampal formation correlation clearly significant in both families. We discuss possible reasons for the continent effect.