Cell motility plays a critical role in many biological and medical processes, including wound healing, morphogenesis, and cancer metastasis (1). Often, this movement is guided by external chemical cues in the form of chemoattractant gradients. These cues consist of diffusive chemoattractant molecules that bind to surface receptors on the cell membrane and activate intracellular signaling pathways, resulting in a polarized asymmetric cell that chemotaxes in the direction of higher chemoattractant concentrations. How chemotaxis works on a single-cell level has been the subject of many detailed experimental and modeling studies (2, 3). In most physiologically relevant cases, however, cells do not move in isolation but, instead, move in groups. This collective motion is a process that is not yet well understood and may play a critical role in the spreading of cancer (4). In particular, it is not clear whether cells that move within a group communicate with each other and, if so, how this cell-cell communication affects the directionality of the group. In PNAS, Ellison et al. (5) performed experiments that suggest that cell-cell communication plays a critical role in branching morphogenesis of the epithelial tissue in mammary glands. Furthermore, in a companion PNAS study, they present a mathematical model of this communication and derive the fundamental limits of the precision of gradient sensing of this model (6).The experiments by Ellison et al. (5) investigate the collective cellular response of epithelial branches in mammary glands using organoids, 3D in vitro organotypic cultures (7). When placed in a gradient of epidermal growth factor (EGF) Ellison et al. (5) find that the formation and extension of these branches exhibit a significant directional bias toward high EGF concentrations (Fig. 1). Without an EGF gradient, however, branch formation displays no directional bias, implying that the multicellular structure is guided by external EGF cues. Importantly, the EGF gradients are generated in mesoscopic fluidic devices and are stable for several days, allowing the quantification of the branching process over a prolonged period.The simplest possible explanation of the observed collective guidance is that each cell is able to sense the chemoattractant gradient and polarizes, independently of its neighbors. The collective branching motion is then the result of the average motion of individual cells. Things are not that simple, however. Ellison et al. (5) clearly show that single cells, dissociated from the organoid, do not respond to EGF gradients and move around aimlessly (Fig. 1). This result is consistent with other multicell experiments that show that collective chemotaxis is possible in the absence of single-cell chemotaxis. For example, both lymphocytes and neural crest cell clusters have been shown to migrate directionally and to display a much higher chemotactic sensitivity than individual cells (8-10).Another possibility is that a multicellular cluster acts as a large "supercell" with concentration det...