To attempt any physical interpretation of this change is scarcely safe: the conditions are so complex. There are reasons for suspecting, however, that it arises from a vacuolation of the substance deposited on the cell-wall. If rapidly deposited, as it is likely to be along lines where sap is freely supplied, this may, in passing from the state of a soluble colloid to that of an insoluble colloid, so contract as to leave uncovered spaces on the cell-membrane; and this change, originally consequent on a physico-chemical action, may be so maintained and utilized by natural selection, as to result in structures of definite kinds, regularly formed in growing parts in anticipation of functions to be afterwards discharged. But, without alleging any special cause for this metamorphosis, we may reasonably conclude that it is in some way consequent upon the carrying of sap. If we examine tissues such as that in the interior of a growing turnip that has not yet become stringy, we may, in the first place, find bundles of elongated cells not having yet developed in them those fenestrated or reticulated structures by which the ducts are eventually characterized. Along the centres of adjacent bundles we may find incomplete lines of such cells--some that are partially or wholly transformed, with some between them that are not transformed. In other bundles, completed chains of such transformed cells are visible. And then, in still older bundles, there are several complete chains running side by side. All which facts imply a metamorphosis of the elongated cells, indirectly caused by the continued action of the currents carried. § 281. Here, however, presents itself a further problem. Taking it as manifest that there is a typical distribution of supporting tissue adapted to meet the mechanical strains a plant is exposed to by its typical mode of growth, and also that there goes on special adaptation of the supporting tissue to the special strains the individual plant has to bear; and taking it as tolerably evident that the sap-channels are originally determined by the passage of currents along lines of least resistance; there still remains the ultimate question--Through what physical actions are established these general and special adjustments of supporting tissue to the strains borne, and these distributions of nutritive liquid required to make possible such adjustments? Clearly, if the external actions produce internal reactions; and if this play of actions and reactions results in a balancing of the strains by the resistances; we may rationally suspect that the incident forces are directly conducive to the structural changes by which they are met. Let us consider how they must work.