A cytochemical method has been used to test claims that the differentiation of the mouse morulae can be accounted for by the position of the blastomeres: the central becoming inner mass cells and the peripheral trophoblast cells. The cytochemical method is an adaptation of Gomori-Takamatsu's demonstration of alkaline phosphatase activity, which stains the inner cell mass and not the trophoblast. The significance of this procedure is dicussed. A simple physical model shows that by packing spheres of similar size, at least 17 are needed to have one completely enclosed. On these basis, if the position is the decisive factor, alkaline phosphatase activity should be detected in embryos with more than 17 cells. Table 1 shows that it is a correct assumption. But since blastomeres that have gone through more cell cycles are smaller and have a better chance of reaching a central position, it became necessary to dissociate number of cells from number of cell cycles. This we did by detecting the appearance of the enzyme activity in halved and fused embryos developedin vitro. Table 2 shows that the activity is detected in halved embryos from the 11-cell stage onwards. Table 3 shows that it is detected in fusion embryos from the 22-cell stage onwards. The results suggest that a temporal coordinate, the number of cell cycles or another temporal change, might interact with the spatial coordinate to produce differentiation.