There is no consensus on the inter-ordinal relationships of eumalacostracans, despite the recent synthesis of several morphological matrices with data from four molecular markers. Signals from different molecules confl ict with each other, and all are conspicuously at odds with morphology. Can fossils help to resolve the problem? Here, we utilize palaeontological data in two ways. Firstly we coded a selection of fossil taxa into our morphological matrix, and assessed their impact upon inferred phylogeny relative to that of their living counterparts (fi rst order jackknifing). This revealed that our morphological tree is very sensitive to the precise taxon sample (a problem that must be addressed in future studies), but that our fossil groups were not disproportionately infl uential. Secondly, we asked whether the order in which groups appear in the fossil record provides a means to choose between competing trees. The congruence between morphological and stratigraphic signals was extremely weak and non-signifi cant in most cases, precluding the use of fossil dates in this way. Many trees imply ghost ranges of duration near the theoretical maximum, and worse than for the majority of other animal groups so far investigated. An incomplete fossil record and fragile/weakly-supported trees combine with considerable molecular rate heterogeneity to make the Eumalacostraca extremely poorly suited to molecular clock studies. Future insights into their phylogeny are likely to come from the development of new molecular markers, as well as hard-won data on internal anatomy and ultrastructure.