Morphological identification of protists remains an expert task, especially for little known and poorly described species. Culture collections normally accept organisms under the name provided by depositors and are not responsible for identification. Uncritical acceptance of these names by molecular phylogeneticists may result in serious errors of interpretation of phylogenetic trees based on DNA sequences, making them appear more incongruent with morphology than they really are. Several cases of misidentification in a major culture collection have recently been reported. Here we provide evidence for misidentifications of two more gymnamoebae. The first concerns "Gephyramoeba sp." ATCC 50654; it is not Gephyramoeba, a leptomyxid with lobose pseudopods, but a hitherto undescribed branching amoeba with fine, filamentous subpseudopods named here Acramoeba dendroida gen. et sp. nov. We also sequenced 18S rRNA of Page's strain of Rhizamoeba saxonica (CCAP 1570/2) and show that it is the most deeply branching leptomyxid and is not phylogenetically close to 'Rhizamoeba saxonica' ATCC 50742, which was misidentified. Correcting these misidentifications improves the congruence between morphological diversity of Amoebozoa and their rRNA-based phylogenies, both for Leptomyxida and for the Acramoeba part of the tree. On morphological grounds we transfer Gephyramoebidae from Varipodida back to Leptomyxida and remove Flamella from Leptomyxida; sequences are needed to confirm these two revisions.