Another general observation which undoubtedly is connected in some way with the problem of coordinated streaming is the following. In externally unstimulated amebas, the new pseudopods are almost without exception directed 60° or less from the direction in which the parent pseudopods are moving. It is a matter of common observation that an ameba may throw out a pseudopod in any direction whatsoever when stimulated. The ameba may reverse its direction of movement completely, or it may move in scores of different directions at one time for awhile, if properly stimulated. There is no restraint or limit imposed upon the ameba insofar as the direction of movement is concerned. Why then should a great majority of new pseudopods in an unstimulated ameba be projected at an angle of approximately 60° to the parent pseudopod? It might seem at first sight as if the merely physical aspect of the streaming would be a sufficient explanation, in that less resistance would be met with in sending a stream off at a small angle than at a large. But it is probable that inertia plays no part in maintaining the direction of streaming (see p. 123, footnote, for further discussion). It requires perhaps more energy for a pseudopod to flow off from the main stream at an angle of 120° than at an angle of 30°. But it is plain that as many pseudopods are withdrawn as are thrown out, and they are withdrawn at an angle against the main stream of endoplasm in the ameba that is the complement of the angle at which they were projected. Whatever energy might be saved therefore in the projection of a new pseudopod at a small angle with the main stream is lost in withdrawing the pseudopod against the stream at a correspondingly large angle. It is clear therefore that the physics of moving viscous fluids cannot solve the problem. It is probable that the mechanism which controls the direction of locomotion as exemplified in the wavy path of the ameba (see p. 109) is also involved in the direction in which pseudopods are projected.