In this paper, we study adaptive evaluation techniques for querying XML- based electronic catalogs, and show, by means of experiments performed on real-world catalogs, that our approach can be used to integrate them with virtually zero effort at start-up time, and a small constant factor needed to perform the adaptive evaluation for all subsequent queries. We reach the conclusion that, from a strictly technical perspective, the classic role of the global catalog can be assumed in an ad-hoc manner by any catalog that forms part of a collaborative federation of XML-based catalogs, and implements our adaptive query algorithms, independently of the storage model used to access its contents.