sets of actions and reactions, we have seen implied by the primordial identity of the endoderm and ectoderm in simple animals, and of the skin and mucous membrane in complex animals (§§ 288, 289). Here we have further to observe that as food is the original source of internal differentiations, these may be expected to show themselves first where the influence of the food is greatest; and to appear later in proportion as the parts are more removed from the influence of the food. They do this. In animals of low type, the coats of the alimentary cavity or canal are more differentiated than the tissue which lies between the alimentary canal and the wall of the body. This tissue in the higher _Cœlenterata_, is a feebly-organized parenchyma traversed by canals lined with simple ciliated cells; and in the lower _Mollusca_ the structures bounding the peri-visceral cavity and its ramifying sinuses, are similarly imperfect. Further, it is observable that the differentiation of this peri-visceral sac and its sinuses into a vascular system, proceeds centrifugally from the region where the absorbed nutriment enters the mass of circulating liquid, and where this liquid is qualitatively more unlike the tissues than it is at the remoter parts of the body. Physiological development, then, is initiated by that instability of the homogeneous which we have seen to be everywhere a cause of evolution (_First Principles_, §§ 149–155). That the passage from comparative uniformity of composition and minute structure to comparative multiformity, is set up in organic aggregates, as in all other aggregates, by the necessary unlikenesses of the actions to which the parts are subject, is shown by the universal rise of the primary differentiation into the parts that are universally most contrasted in their circumstances, and by the rise of secondary differentiations obviously related in their order to secondary contrasts of conditions. § 312. How physiological development has all along been aided by the multiplication of effects--how each differentiation has ever tended to become the parent of new differentiations, we have had, incidentally, various illustrations. Let us here review the working of this cause.