During tissue development, cells are able to form different layers and to self organize. At the moment, there are two principal hypotheses that try to explain cell sorting. The so called differential adhesion hypothesis (DAH) posits that cells organize according to the level of expression of cell-cell adhesion proteins so as to reduce free-energy in the system. The second hypothesis explains the formation of different germ layers according to cortical tension among cells of distinct phenotypes. For this work, we wanted to study cell sorting in the mammary glands, which are bilayered branching structures composed of outer myoepithelial (MEP) and inner luminal epithelial cells (LEP). Both sorting hypotheses were tested using bioengineered cell culture substrata. Using arrays of micro-wells that were cast in polydimethylsiloxane (PDMS) coupled with quantitative image analysis, we observed that cell sorting occurred over time between LEP and MEP that were isolated by fluorescence-activated cell sorting from primary human mammary epithelial cells. Inhibitors of specific cadherins were able to prevent cell sorting in accordance with the DAH. However, chemical inhibitors of cytoskeletal proteins, which effectively prevented the cells from modulating their innate elasticity, also prevented cell sorting suggesting cortical tension also plays an important role. Thus, our data suggests a third hypothesis, which is essentially a unification of the original two, such that differences in cortical tension are essentially lineage-specific and are governed wholly or in-part by cell-cell adhesion molecules. Experimental resolution of this hypothesis is being actively pursued