From further observations Rhumbler (’05, ’10) came to modify his earlier views as stated above. The rapid advances in the study of the chemistry of colloids doubtless suggested to Rhumbler that the change from endoplasm to ectoplasm resembled the change from a sol to a gel state, and that in this process of gelation lay the source of energy manifested in ameboid movement. In thus calling attention to, and emphasizing the colloidal nature of, the conversion of endoplasm into ectoplasm and vice versa, the problem of ameboid movement came to be discussed from an entirely new angle. Certain phases of Rhumbler’s theory are developed and elaborated by Hyman (’17) who agrees in general with Rhumbler’s conclusions. In a series of papers on feeding and other reactions of ameba, Schaeffer (’12, ’16, ’17) concluded that Rhumbler’s general statement, wherein he says that changes in behavior are directly deducible from the action of stimuli in effecting liquefaction or gelation of the ectoplasm, does not hold in many cases of feeding, and that the mechanism controlling locomotion and feeding is not external, as maintained by Rhumbler, but internal. CHAPTER III _The General Features of Endoplasmic Streaming_ The streaming of the endoplasm is the most conspicuous feature of ameboid movement. It is even more noticeable than the movement of the pseudopods themselves, because of its greater speed and because it occurs in all parts of the ameba. Its importance in movement is essential, for no continued locomotion can be observed unless accompanied by streaming. It may be profitable therefore to enquire into the general features of streaming, and to observe some of the necessary consequences streaming imposes upon such an animal as the ameba. Let us take as an example an _Amoeba proteus_ (Pallas, ’66, emend. Leidy, ’79, emend. Schaeffer, ’16) in characteristic movement (see Figure 11, p. 37). The main streams of endoplasm are in the same direction as that in which the ameba moves. In the withdrawing pseudopods the current is, of course, toward the main mass of the ameba. The endoplasmic stream is continuous from the posterior end to the tips of the advancing pseudopods. The retracting pseudopods flow into the main stream as tributaries. If, as often happens, the ameba is without pseudopods, there is then a single stream arising in the posterior end and flowing to the anterior end. In such a case it is readily observed how absolutely dependent locomotion is upon endoplasmic streaming.