a considerable amount of experiment and observation on endoplasmic streaming in plants, Ewart (’03) has come to the conclusion that there are differences in electrical potential between the protoplasm-vacuole boundary and the protoplasm-cell wall boundary, and that as a consequence electrical currents are passing between these points, traversing the protoplasmic stream. If now it is assumed that the particles in the endoplasm, which are electrically polarized, have the surface tension of their corresponding ends decreased when electric currents traverse the endoplasmic stream, the particles and, of necessity, the whole stream of endoplasm would move in the direction of lowered surface tension (Figure 31, p. 96). Continuous chemical actions would be necessary to maintain the conditions as outlined. This theory accords with the facts so far as it goes, but it does not explain the streaming in threads across the vacuole in the plant cell, thus necessitating two theories for the explanation of streaming within a single cell at the same moment. Moreover a central vacuole of cell sap seems always to be required to fulfill the conditions of this theory, and this, as is readily seen, makes it impossible to apply it to streaming in amebas, myxomycetes, foraminifera and ciliates. The fundamental cause of streaming is therefore still to be discovered, for neither the theories of streaming as applied to ameba, nor those described above which refer especially to plant cells, are satisfactory. But a significant point in these theories is that with increasing information, they come more and more to demand a colloidal structure in the protoplasm. It is the surface energy in the interfaces in the colloidal system which comes to be regarded as the primary source of the energy. But all attempts thus far to explain exactly how this energy is utilized have been unsuccessful. Gaidukov’s (’10) observation is of some interest, however in this connection. He found that the occasional stopping of streaming in cells of _Vallisneria_ is accompanied by a cessation of Brownian movement, which indicates a change from a sol to a gel state. This proves therefore that colloidal changes are possible in streaming protoplasm, and that the general search for an explanation of streaming along this line is proceeding in the right direction. The researches of Bancroft (’13, ’14) and especially of Clowes (’13, ’16) on the nature of the change of phase in emulsions are very instructive in this connection; and it is undoubtedly true that as rapid progress is now being made by the investigation of colloidal solutions as by the direct study of protoplasm, in solving the problem of streaming. The problem of the _control_ of the streaming process, which is of course much the most important feature of streaming, will probably be solved, at least in part, when the mechanics of streaming is understood. CHAPTER XI THE SURFACE LAYER AS A LOCOMOTOR ORGAN