Because of the intense interest in the stem cell field, even reports of failure to replicate previous findings have appeared in prominent journals. For example, two recent reports question whether adult bone marrow-derived cells contribute to central nervous system (CNS) neurons because the authors failed to see markers of such cells in brains (“Failure of bone marrow cells to transdifferentiate into neural cells in vivo,” RF Castro et al., Brevia, 23 Aug., p. 1299;“Little evidence for developmental plasticity of adult hematopoietic stem cells,” AJ Wagers et al., Reports, 27 Sept., p. 2256; published online 5 Sept.; 10.1126/science. 1074807. These conclusions are in marked contrast with previous reports by us and others (1–3) that found that bone marrow-derived cells transit to the CNS in adult mice, express proteins typical of neurons in the olfactory bulb, and contribute to well-defined subsets of neurons such as Purkinje cells in the cerebellum. Efforts to replicate discoveries are critical to the scientific process, and convincing failures to do so are important contributions to the literature. The fundamental issue is defining what makes a study convincing and, therefore, what should be the criteria for overturning previously published findings. The findings by Castro et al. and Wagers et al. underscore the need to establish criteria for publishing negative reports.

First, a prerequisite for proving a null finding is the clear ability to detect a positive control. A case in point is Castro et al., who fail to detect not just neurons but also bone marrow-derived microglial cells within the CNS. At least 20 reports over the past 15 years have shown that bone marrow transplantation results in readily detectable replacement of a large proportion of microglial cells in the brain (4–8). Moreover, following a stab wound, the presence of such cells in the brain would be impossible to miss, as they are localized in great abundance at the site of the wound (9). Thus, the lack of detection of microglia by Castro et al. suggests that their system was unable to detect marrow-derived cells that should have been present in the brain.