Such differentiations, marked in various degrees, are to be seen among _Mollusca_. In the _Pteropoda_ the only modification which appears to facilitate respiration, is the minute vascularity of one part of the skin. Higher types possess special skin-developments. The _Doris_ has appendages developed into elaborately-branched forms--small trees of blood-vessels covered by slightly-changed dermal tissues. And these arborescent branchiæ are gathered together into a single cluster. Thus there is evidence that large external respiratory organs have arisen by degrees from simple skin: as, indeed, they do arise during the development of each individual having them. Just as gradually as in the embryo a simple bud on the integument, with its contained vascular loop, passes by secondary buddings into a tree-like growth penetrated everywhere by dividing and subdividing blood-vessels; so gradually has there probably proceeded the differentiation which has turned part of the outer surface into an organ for excreting carbonic acid and absorbing oxygen. Certain inferior vertebrate animals present us with a like metamorphosis of tissues. These are the _Amphibia_. The branchiæ here developed from the skin, are covered with cellular epidermis, not much thinner than that covering the rest of the body. Like it they have their surfaces speckled with pigment-cells; and are not even conspicuous by their extra vascularity--where they are temporary at least. They facilitate the exchange of gases in scarcely any other way than by affording a larger area of contact with the water, and interposing a rather thinner layer of tissue between the water and the blood-vessels. Those very simple branchiæ of the larval _Amphibia_ that have them but for a short time, graduate into the more complex ones of those that have them for a long time or permanently; showing, as before, the small stages by which this heterogeneity of surface accompanying heterogeneity of function may arise. In what way are such differentiations established? Mainly, no doubt, by natural selection; but also to some degree, I think, by the inheritance of direct adaptations. That a portion of the integument at which aëration is favoured by local conditions, should thereby be led to grow into a larger surface of aëration, appears improbable. Survival of those individuals which happen to have this portion of the integument somewhat more-developed, seems here the only likely cause. § 293. Among the conspicuous modifications by which the originally-uniform outer layer is rendered multiform, are the protective structures. Let us look first at the few cases in which the formation of these is ascribable mainly to direct equilibration.