Recent studies have reported that clinical relevant numbers of cardiac stem cells (CSCs) with cardiomyogenic potential can be obtained from small heart tissue biopsies, by an intrinsic ability of CSCs to form beating cardiospheres (CSs) during ex vivo culture. Such data have provided optimism that injuried heart tissue may be repaired by stem cell therapy using autologous CS derived cells, and pre-clinical studies have already been described in literature. Herein, we established CSs from neonatal rats, and by immunofluorescence, qRT-PCR, and microscopic examination we demonstrated that spontaneously beating CSs likely consisted of contaminating myocardial tissue remnants, and that inclusion of an intervening filtration step to the original CS protocol resulted in CSs that were devoid of any cardiomyogenic potential. Similar results were obtained with CSs derived from neonatal mice as wells as adult rats/mice. Additionally, we demonstrated by in vitro culture, FACS cell sorting, and immunofluorescence that CSs were generated by aggregation of Gata4+/collagen I+/αSMA+/CD45- cells, whereas previously proposed CS forming cells, “phase bright cells”, were Gata4-/collagen I-/αSMA-/CD45+ and unable to form CSs by themselves. Phenotypically, CS cells largely resembled fibroblasts, and they lacked cardiomyogenic as well as endothelial differentiation potential. Our data imply that at least the murine cardiosphere model seems unsuitable for enrichment of cardiac stem cells with cardiomyogenic potential, and although the present study may remain controversial, we suggest that human CSs should be thoroughly characterized with respect to phenotype and differentiation potential before initiating human trials.